Ampakines Stimulate Diaphragm Activity after Spinal Cord Injury

Abstract

Respiratory compromise after cervical spinal cord injury (SCI) is a leading cause of mortality and morbidity. Most SCIs are incomplete, and spinal respiratory motoneurons as well as proprio- and bulbospinal synaptic pathways provide a neurological substrate to enhance respiratory output. Ampakines are allosteric modulators of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, which are prevalent on respiratory neurons. We hypothesized that low dose ampakine treatment could safely and effectively increase diaphragm electromyography (EMG) activity that has been impaired as a result of acute- or sub-acute cervical SCI. Diaphragm EMG was recorded using chronic indwelling electrodes in unanesthetized, freely moving rats. A spinal hemi-lesion was induced at C2 (C2Hx), and rats were studied at 4 and 14 days post-injury during room air breathing and acute respiratory challenge accomplished by inspiring a 10% O₂, 7% CO₂ gas mixture. Once a stable baseline recording was established, one of two different ampakines (CX717 or CX1739, 5 mg/kg, intravenous) or a vehicle (2-hydroxypropyl-beta-cyclodextrin [HPCD]) was delivered. At 4 days post-injury, both ampakines increased diaphragm EMG output ipsilateral to C2Hx during both baseline breathing and acute respiratory challenge. Only CX1739 treatment also led to a sustained (15 min) increase in ipsilateral EMG output. At 14 days post-injury, both ampakines produced sustained increases in ipsilateral diaphragm EMG output and enabled increased output during the respiratory challenge. We conclude that low dose ampakine treatment can increase diaphragm EMG activity after cervical SCI, and therefore may provide a pharmacological strategy that could be useful in the context of respiratory rehabilitation.

Introduction

Up to 500,000 new cases of spinal cord injury (SCI) are diagnosed globally every year.¹ The majority of SCIs are to the cervical region,² and such injuries impair respiratory muscle activation. In severe cases, this necessitates the use of mechanical ventilation or direct diaphragm pacing to sustain ventilation. However, even if independent breathing is maintained, the ability to increase respiratory muscle activation is typically impaired.³ Inability to cough or sigh, or to increase breathing during physical exertion, can negatively impact health and quality of life after SCI.

Ampakines are positive allosteric modulators of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors.⁴ AMPA receptors are widely expressed in brainstem pre-motor circuits⁵ and spinal phrenic motoneurons,⁶ thereby providing a molecular substrate for stimulating respiratory motor output by using ampakines.^7,8 Recent studies in anesthetized rats indicate that ampakines can enhance expression of respiratory neuroplasticity,⁹ and stimulate phrenic nerve discharge after cervical SCI.¹⁰ In the latter study, intravenous (i.v.) ampakine delivery increased phrenic burst amplitude in anesthetized rats with cervical SCI. These proof-of-concept studies demonstrated the potential for ampakine treatment to stimulate neural drive to the diaphragm. However, anesthesia can dramatically alter diaphragm activation after SCI,¹¹ making it difficult to interpret the potential clinical significance of using ampakines to increase diaphragm activation after cervical SCI.

We hypothesized that low dose ampakine therapy could effectively stimulate diaphragm activity in freely moving unanesthetized rats following a unilateral cervical SCI. The effect of two different ampakines, CX717 and CX1739, was studied on diaphragm muscle activation and breathing. Based on studies of AMPA receptor kinetics, CX717 and CX1739 are considered to be “low impact,”⁴ and have a greater safety profile than other ampakines.

The current work utilized a C2 spinal hemisection model (C2Hx) that disrupts the excitatory bulbospinal glutamatergic drive to the ipsilateral phrenic motor neuron pool.^12,13 Diaphragm electromyographic (EMG) activity was recorded using indwelling electrodes to enable longitudinal, within-animal, quantitative measurements. The impact of ampakine treatment on diaphragm activation was assessed during quiet breathing and respiratory behaviors requiring substantially greater diaphragm activation. The CX717 and CX1739 compounds were administered at a dose (5 mg/kg) well below what has previously been given to spinally intact humans.¹⁴ A secondary aim was to determine whether combining low dose ampakine with brief hypoxia exposure could enhance the ampakine impact. That question was inspired by recent work indicating that intracellular signaling pathways activated by hypoxia converge with pathways activated by ampakines in a manner that enhances the expression of spinal respiratory neuroplasticity.⁹

Methods

Experimental animals

A total of 25 adult (12 weeks of age) male and female Sprague–Dawley rats were obtained from Envigo (Hsd:Sprague Dawley^® SD, Indianapolis, IN) and randomly assigned to 2-hydroxypropyl-beta-cyclodextrin (HPCD), CX717, or CX1739 groups. An additional seven adult male and female Sprague–Dawley rats were assigned to the intact group. All procedures were approved by the Institutional Animal Care and Use Committee at the University of Florida (Protocol # 201807438) and are in accordance with National Institutes of Health Guidelines. Animals were housed individually in cages under a 12-h light/dark cycle with ad libitum access to food and water. Animals were anesthetized with ketamine (90 mg/kg) and xylazine (10 mg/kg) via intraperitoneal injection for all surgical procedures.

Assessment of diaphragm muscle activity

Chronic diaphragm muscle EMG activity was assessed in all animals using a pair of inwelling electrodes that were implanted in the mid-costal region of the diaphragm^15,16 and connected to a head stage. Briefly, the thoraco-abdominal and skull region of the rat were shaved with clippers and cleaned with three alternating scrubs of 2% chlorohexidine followed by 70% ethanol. Rats were then placed in a stereotaxic frame over a heating pad and the skull surface was exposed under sterile conditions. Six bone anchor screws were placed in the skull (1 in the left and 1 in the right frontal bone, 2 in the left parietal bone, and 2 in the right parietal bone). A ground wire was wrapped around three head screws before securing the head stage in place with dental cement. The animal was removed from the stereotaxic frame. Subsequently, EMG electrodes connected to the head stage were tunneled subcutaneously to the abdominal region. The diaphragm muscle was exposed following a midline laparotomy. Electrodes made from insulated stainless steel fine wire (AS631, Cooner Wire Inc., Chatsworth, CA) with a ∼3mm exposed region were placed into each hemi-diaphragm muscle with an inter-electrode distance of ∼3 mm and secured with a surgical knot. Abdominal muscles and skin (buried suture) were closed with a 4-0 Webcryl suture. Animals received buprenorphine (0.03 mg/kg, b.i.d.), meloxicam (10 mg/kg), Baytril^® (5 mg/kg, q.d.), and lactated Ringer's solution (10 mL/day, q.d.) for the initial 48 h post-injury. Animals were allowed to recover for 7 days prior to baseline data collection. Animals were excluded from the study if there were malfunctioning EMG electrodes or if they developed an abscess around the electrodes tunneled subcutaneously.

Baseline diaphragm EMG activity was recorded 2 days prior to SCI, and subsequently at 4 and 14 days post-C2Hx. EMG data was recorded using a differential amplifier (A-M systems 1700), and digitized using a 16-channel data acquisition unit (1401-Power3) and custom Spike2 acquisition software. EMG data were sampled at 5,000 Hz, analog band-pass filtered between 300 and 10,000 Hz, and saved to disk for offline filtering and analysis (see Data analysis section).

SCI

All animals in this study received a C2Hx, which has previously been described in detail.^17,18 Animals were prepped for survival surgical procedure, as described previously for the EMG implant procedure. Before beginning the procedure, depth of anesthesia was checked using a toe pinch resulting in lack of a change in heart rate, whisker twitch, and hindlimb withdrawal. An appropriate level of anesthesia was also continuously verified during the surgical procedure and the animal was re-dosed with a one-quarter dose of the initial ketamine/xylazine dose if needed. Body temperature was maintained with a water recirculating heating pad. Under sterile conditions, a dorsal incision was made from the base of the skull to the ∼ C6 region of the spine. Dorsal paravertebral muscles between C1 and C6 were incised and retracted. The posterior portion of cervical vertebrae was exposed at C2, and a laminectomy was performed at the level, while preserving the facet joints and leaving the dura intact. The left side of the spinal cord was sectioned using a micro-knife (Cat # 10055-12, Fine Science Tools) starting from the midline to the lateral edge of the spinal cord. This process was repeated three times to ensure completeness of hemisection. The overlying muscles were sutured with sterile 4–0 Webcryl and the overlaying skin was closed using 9 mm wound clips. Animals were maintained on a heating pad until alert and awake. Animals were monitored on a daily basis for signs of distress, dehydration, and weight loss, with appropriate veterinary care given as needed. Animals received buprenorphine (0.03 mg/kg, b.i.d.), meloxicam (1 mg/kg, q.d.), Baytril (5 mg/kg, q.d.), and lactated Ringer's solution (10 mL/day, q.d.) for the initial 48 h post-injury. In order to verify completeness of C2Hx, absence of ipsilateral diaphragm EMG activity was confirmed under anesthesia at the time of surgery. Animals were maintained on a heating pad until alert and awake. Animals were monitored on a daily basis for signs of distress, dehydration, and weight loss, with appropriate veterinary care given as needed. Animals were excluded from the study if they displayed forelimb autotomy or weight loss >20% of a pre-injury time point.

Plethysmography measurements

Unrestrained and unanesthetized animals were studied in a flow through whole body plethysmograph to measure ventilation. Expired air outflow from the plethysmograph chamber was split through a CO₂ sensor to quantify metabolic activity (VCO₂) based on the Fick's principle. The plethysmograph chamber was modified to include a commutator for connecting cables to the head stage of animals placed in the chamber (Fig. 1). This setup allowed concurrent measurements of ventilation and EMG on the same animal. Awake rats were acclimated to the plexiglas recording chamber 2 days in a row, for 1 h each day. Air flow was maintained at 6 L/min through the plethysmography chamber. On the 3rd day, baseline data were collected for a 40-min period under normoxic air (21% O₂, 79% N₂), followed by a 2-min ampakine or HPCD infusion period and a 30-min post-infusion period. Subsequently, rats were challenged to a 7 min hypoxic (10.5% O₂ bal N₂) period, followed by a 45-min post-hypoxia recording period under normoxic air, ending with a 7 min maximal chemoreceptor challenge (10.5% O₂, 7% CO₂ bal N₂). For pre-injury baseline data points, an infusion of sterile saline was used in all animals as a control for ampakine or HPCD infusion. Ventilation was measured by monitoring pressure swings within the plexiglas chamber caused by gas expansion and contraction caused by the heating and humidification of air as the animal breathed. Tidal volume was measured using the Drorbaugh and Fenn equations.¹⁹ VCO₂ was measured using Fick equation VCO₂ = (FECO₂- FICO₂) × V/[1- (FECO₂ × (1- (1/RQ)))] with an assumed RQ of 0.85.²⁰

FIG. 1.

Overview of experimental design and methods. (A) Schematic of the recording setup allowing simultaneous collection of whole-body plethysmography data and electromyographic (EMG) activity data from the diaphragm muscle. (B) Experimental timeline of study. Animals underwent survival surgery for EMG electrode implantation 8 days prior to injury. Acclimation to the plethysmography chamber occurred 3 and 2 days prior to injury, followed by baseline data collection 1 day prior to injury. After C2 spinal hemisection (C2Hx) injury, data were collected at 4 and 14 days post-injury (dpi). (C) During experiments, animals were exposed to normoxic air (21% O₂, balance N₂) for 40 min, followed by infusion of drug or vehicle over 2 min. Post-infusion data were collected under normoxic air for 30 min, followed by an acute hypoxic challenge (10.5% O₂ and balance N₂). Post-hypoxia data were collected under normoxic air for 45 min, and the protocol concluded with a brief maximal chemoreceptor (maxchemo) challenge (10.5% O₂, 7% CO₂, balance N₂). (D) Schematic of bilateral intramuscular EMG electrode placement in the mid-costal region of the diaphragm. Color image is available online.

Pharmacological agents

Ampakine CX717 and CX1739 was provided by RespireRX and was mixed in 10% HPCD solution (Sigma) at soluble concentrations of (5 mg/mL). Aliquots were stored at −80°C for use up to 6 months. Aliquots were thawed to room temperature prior to use on the day of each experiment. On the day of recording, animals were anesthetized under 2% isofluorane, and a 22G catheter was placed on the distal end of the tail vein. The animal was then placed in the plethysmograph chamber, and the i.v. line was externalized through a specially designed port to maintain a pressure seal in the chamber. Based on pilot studies from our laboratory, CX717 and CX1739 were used at low dose concentration of 5 mg/kg. All drugs were administered over the course of 2 min via the tail vein catheter.

Immunohistochemistry

Following the 14-day post-C2HX recording, animals were euthanized by exsanguination and transcardially perfused with 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS) (pH 7.4). The spinal cord was resected from C1 to C4, post-fixed in 4% paraformaldehyde for 24 h and transferred to 30% sucrose in 0.1 M PBS (pH 7.4) for 3 days at 4°C. Spinal cords were subsequently embedded in cryomolds (VWR, Radnor, PA). The spinal cord was cut in five 3-mm segments centered around the injury epicenter. The block was sectioned serially at 20 μm thickness in the transverse plane into 10 serial sets and frozen on Superfrost Plus slides (Thermo Fisher Scientific, Waltham, MA). Sections were then re-hydrated and stained with 0.1% Cresyl violet in glacial acetic acid. After 3 min, slides were rinsed in water and dehydrated through graded alcohol steps, followed by xylene.

Data analysis

EMG activity from the ipsilateral and contralateral diaphragm muscle was recorded throughout the course of the recording paradigm detailed in the plethysmography section. Data for EMG and plethysmography traces were averaged over 5 min of quiet breathing during the baseline period, over 2 min during the infusion period, over 5 min immediately post-infusion and during min 15–20 post-infusion. During the hypoxia and max-chemo challenge, data were averaged both minute by minute during min 2–7 of the challenge, and during the last 3 min of the challenge, during which the diaphragm muscle is expected to elicit maximal response for that behavior. Post-hypoxia response was averaged over the course of 5 min at three time points post- challenge: 5, 20, and 35 min.

Assessment of plethysmography traces and EMG activity was conducted using a custom MATLAB software (MathWorks). Plethysmography tidal volume traces were filtered 0.5–6 Hz to reduce electrocardiogram (ECG) artifact. The EMG traces were “blanked” using pan_tompkin to detect QRS complex in the EMG signal, then filtered 300–1000 Hz to further remove ECG artifacts. The EMG signal was then rectified, moving-median filtered using a 50 ms time constant, and finally smoothed using a 50 ms time constant.

Statistical analysis

All statistical evaluations were performed using JMP statistical software (version 14.0, SAS Institute Inc., Cary, NC). Power calculations for each cohort of animals in the study were determined prior to initiation of study and based on pilot experiments not included in this article. The study statistical design was powered to consider the expected standard deviation in diaphragm EMG activity at 4 days post-C2Hx (21%) and in order to detect a 50% difference at a power of 0.9 and an α of 0.05. Using the mixed linear model, statistical significance was established at the 0.05 level and adjusted for any violation of the assumption of sphericity in repeated measures using the Greenhouse–Geisser correction. Treatment (CX717, CX1739, and HPCD) and time point (baseline, infusion, post-infusion and 15 min post-infusion ) or behavior (hypoxia, maxchemo) were included as model variables, using animals as a random effect. When appropriate, post-hoc analyses were conducted using Tukey–Kramer honestly significant difference (HSD). Normality of the distribution was assessed using the Shapiro–Wilk test within each animal. Outliers were identified for each animal using an outlier box plot. No data were excluded from the analysis to highlight the responders versus the non-responders in each treatment group.

Results

Diaphragm EMG activity: Impact of SCI

Representative histological images of the injury epicenter from three different animals are provided in Figure 2. The lesions removed the ipsilateral ventrolateral funiculus which contains the majority of bulbospinal pathways to phrenic motor neurons.^21–23 Representative EMG traces from the left diaphragm (i.e., ipsilateral to the lesion) are shown in Figure 3; statistical analyses are presented in Table 1. C2Hx injury decreased ipsilateral EMG_peak amplitude at both the 4- and 14-day post-C2Hx time points, with no impact on contralateral EMG_peak. Respiratory rate was elevated above pre-injury values at 14 days but not at 4 days.

FIG. 2.

Histological assessment of C2 spinal hemisection (C2Hx) injury epicenter. Serial, transverse sections (20 μm thickness) spanning ∼1 cm around the lesion epicenter (C1-C3) were stained with Cresyl violet in order to identify the extent of injury. A representative image from two different animals is displayed in panel A–B. Original scale bar, 1 mm. Color image is available online.

FIG. 3.

Impairment of diaphragm electromyographic (EMG) activity following C2 spinal hemisection (C2Hx) injury. (A) Representative EMG traces (black) and corresponding processed traces (red; rectified, smoothed, filtered) from the diaphragm ipsilateral to C2Hx (or left side in spinal-intact animal). ata are shown pre-injury, and 4 and 14 days post-injury (or corresponding time point in spinal-intact animals) during eupneic breathing. The diaphragm EMG recordings from the chronic indwelling electrodes were stable over the course of the study in all animals. (B) Mean ipsilateral and contralateral diaphragm EMG activity (peak amplitude) and respiratory rate. Ipsilateral EMG activity is reduced by ∼50–60% post-C2Hx, with minimal impact on contralateral EMG activity. Data are represented as mean with corresponding confidence interval, and individual data (spinal intact, gray dots, n = 7; C2Hx, black dots, n = 10). Statistical summary for the effect of C2Hx injury on diaphragm EMG activity is shown in Table 1 (a.u., arbitrary units). Color image is available online.

Table 1.

Statistical Summary for the Effect of C2Hx Injury on Diaphragm EMG Activity Using a Mixed Linear Model with Animal as Random Effect

Outcome	Effect	df	dfDen	F ratio	p value
Ipsilateral EMG	DPI	2	32.92	20.7329	<0.0001*
	Group	1	18.3	68.9396	<0.0001*
	DPI × Group	2	32.92	37.9002	<0.0001*
Contralateral EMG	DPI	2	32.92	7.0181	0.0029*
	Group	1	18.44	0.0075	0.9319
	DPI × Group	2	32.92	0.7772	0.468
Respiratory rate	DPI	2	32	5.9251	0.0065*
	Group	1	16	3.8168	0.0685
	DPI × Group	2	32	6.4745	0.0044*

Fixed effects include DPI (pre-injury, 4 DHx, or 14 DHx) and group (intact or C2Hx). df (between-groups degrees of freedom) and dfDen (within-groups degrees of freedom). Summary data presented in Figure 2.

^* denotes p < 0.05.

C2Hx, C2 spinal hemisection model; EMG, electromyography; DPI, days post-injury.

Diaphragm EMG activity: impact of ampakines after acute injury (4 days)

Prior to ampakine administration, deficits in ipsilateral EMG activity were similar across the three experimental groups (Table S2; one-way analysis of variance [ANOVA], p = 0.69). Delivery of the HPCD vehicle did not impact ipsilateral or contralateral diaphragm EMG_peak activity during baseline breathing (Fig. 4, statistical evaluations in Table 2). In contrast, CX717 infusion immediately increased ipsilateral (∼ 120% from baseline EMG_peak) and contralateral (∼ 60%) EMG_peak activity, which returned to pre-infusion values by 15 min. Infusion of CX1739 also caused an increase in ipsilateral EMG_peak (∼ 90%), which was sustained at 15 min post-infusion (∼ n60%). The contralateral recording showed a comparatively lower increase in EMG_peak after CX1739 infusion and at 15 min (∼ 30%).

FIG. 4.

Impact of ampakines on peak diaphragm electromyograpic (EMG) amplitude after acute injury (4 days post-C2 spinal hemisection [C2Hx] injury). The plots depict the peak ipsilateral (left columns) and contralateral (right columns) EMG (EMG_peak) response to intravenous delivery of CX717, CX1739, or vehicle (2-hydroxypropyl-beta-cyclodextrin [HPCD]). Data are expressed as mean (diamond) ± standard error, with individual data points as well: CX717 (blue, n = 9), CX1739 (red, n = 8), and HPCD (black, n = 8) groups. (A) Impact of ampakine or vehicle infusion on diaphragm EMG activity during infusion, at 0–5 min post-infusion and at 15–20 min post-infusion. (B) Impact of ampakine or vehicle infusion on diaphragm EMG activity during respiratory challenge: hypoxia (10.5% O₂) or maxchemo (7% CO₂, 10.5% O₂). (C) Impact of ampakine or vehicle following acute hypoxia on diaphragm EMG activity was evaluated at 5–10 min, 20–25 min, and 35–40 min post-hypoxia. Raw values of EMG amplitudes are reported in Table S2. Statistical results are reported in Table 2. BL, baseline. Color image is available online.

Table 2.

Statistical Summary for the Impact of Ampakines on Diaphragm EMG at 4 Days post-C2Hx Using Mixed Linear Model with Animal as Random Effect

Outcome	Segment	Interaction	df	dfDen	F ratio	p value
Ipsilateral EMG	Infusion	Behavior	3	66	20.3562	<0.0001*
		Treatment	2	22	5.4146	0.0122*
		Behavior × Treatment	6	66	6.2882	<0.0001*
Contralateral EMG	Infusion	Behavior	3	66	18.1481	<0.0001*
		Treatment	2	22	1.8718	0.1775
		Behavior × Treatment	6	66	4.065	0.0016*
Ipsilateral EMG	Challenges	Behavior	2	44	15.313	<0.0001*
		Treatment	2	22	3.8917	0.0357*
		Behavior × Treatment	4	44	3.2219	0.0210*
Contralateral EMG	Challenges	Behavior	2	44	98.1896	<0.0001*
		Treatment	2	22	0.7828	0.4695
		Behavior × Treatment	4	44	0.6127	0.6557
Ipsilateral EMG	Post-hypoxia	Behavior	3	66	8.3951	<0.0001*
		Treatment	2	22	7.6224	0.0031*
		Behavior × Treatment	6	66	0.6447	0.6941
Contralateral EMG	Post-hypoxia	Behavior	3	66	71.2945	<0.0001*
		Treatment	2	22	1.2034	0.3192
		Behavior × Treatment	6	66	0.9035	0.4979

Fixed effects include (infusion: baseline, infusion, 0–5 min post-infusion, 15–20 min post-infusion; challenges: baseline, hypoxia, maxchemo; post-hypoxia: hypoxia, 5–10 min post- hypoxia, 20-25 mins post hypoxia, 35-40 mins post hypoxia) and treatment (HPCD, CX717 or CX1739). df (between-groups degrees of freedom) and dfDen (within-groups degrees of freedom). Summary data presented in Figure 3.

^* denotes p < 0.05.

EMG, electromyography; C2Hx, C2 spinal hemisection model; HPCD, 2-hydroxypropyl-beta-cyclodextrin.

Acute respiratory challenges require a higher level of diaphragm activation. Rats treated with HPCD were unable to increase ipsilateral diaphragm EMG_peak during hypoxia, or during the maxchemo challenge (Fig. 4B). In contrast, rats treated with CX717 were able to increase ipsilateral EMG_peak during the hypoxia (∼ 25%) and maxchemo challenges (∼80%). Rats treated with CX1739 also showed a robust increase in ipsilateral EMG_peak during both the hypoxia (∼ 90%) and maxchemo (∼ 80%) challenges. Contralateral diaphragm EMG_peak increased during both respiratory challenges, but responses were similar among the three experimental groups. Therefore, the impact of ampakines during respiratory challenge was restricted to the ipsilateral EMG output.

We also evaluated if the combination of low dose ampakine and a brief hypoxia exposure could evoke a sustained increase in diaphragm activity. In rats treated with HPCD or CX717, ipsilateral diaphragm EMG_peak rapidly returned to baseline levels after hypoxia (Fig. 4C). In contrast, CX1739-treated rats showed an increase in ipsilateral diaphragm EMG_peak activity that remained elevated up to 40 min after the hypoxia exposure.

Diaphragm EMG activity: Impact of ampakines after sub-acute injury (14 days)

HPCD vehicle had no discernable impact on ipsilateral or contralateral EMG_peak at 14 days post-C2Hx (Fig. 5A, statistical evaluations in Table 3). In contrast, CX717 infusion caused an immediate increase in ipsilateral (∼ 45%) and contralateral (∼ 25%) EMG_peak output, which was sustained for up to 15 min. Infusion of CX1739 also evoked an increase in ipsilateral (∼ 25%) and contralateral (∼ 20%) EMG_peak activity. This increase was sustained on the ipsilateral side for 15 min (∼25%), but the contralateral output rapidly returned to baseline.

FIG. 5.

Impact of ampakines on peak diaphragm electromyographic (EMG) amplitude at a sub-acute time point post-injury (14 days post-C2 spinal hemisection [C2Hx] injury). Plots depict the peak ipsilateral (left columns) and contralateral (right columns) EMG (EMG_peak) response to intravenous delivery of CX717, CX1739 or vehicle (2-hydroxypropyl-beta-cyclodextrin [HPCD]). Data are expressed as mean ± standard error (diamond), with individual data points as well: CX717 (blue, n = 9), CX1739 (red, n = 8), and HPCD (black, n = 8) groups. (A) Impact of ampakine or vehicle infusion on diaphragm EMG activity during infusion, at 0–5 min post-infusion and at 15–20 min post infusion. (B) Impact of ampakine or vehicle infusion on diaphragm EMG activity during respiratory challenge: hypoxia (10.5% O₂) or maxchemo (7% CO₂, 10.5% O₂). (C) Impact of ampakine or vehicle following acute hypoxia on diaphragm EMG activity was evaluated at 5–10 min, 20–25 min, and 35–40 min post-hypoxia. Raw values of EMG amplitudes are reported in Table S2. Statistical results are reported in Table 3. BL, baseline. Color image is available online.

Table 3.

Statistical Summary for the Impact of Ampakines on Diaphragm EMG at 14 Days post-C2Hx Using Mixed Linear Model with Animal as Random Effect.

Outcome	Segment	Interaction	df	dfDen	F ratio	p value
Ipsilateral EMG	Infusion	Behavior	3	66	7.3561	0.0003*
		Treatment	2	22	7.2327	0.0039*
		Behavior × Treatment	6	66	2.1984	0.054
Contralateral EMG	Infusion	Behavior	3	66	7.1632	0.0003*
		Treatment	2	22	5.0665	0.0155*
		Behavior × Treatment	6	66	2.8123	0.0169*
Ipsilateral EMG	Challenges	Behavior	2	44	10.5907	0.0002*
		Treatment	2	22	5.2683	0.0135*
		Behavior × Treatment	4	44	3.9354	0.0081*
Contralateral EMG	Challenges	Behavior	2	44	141.028	<0.0001*
		Treatment	2	22	1.4694	0.2518
		Behavior × Treatment	4	44	0.8043	0.5291
Ipsilateral EMG	Post-hypoxia	Behavior	3	66	12.2737	<0.0001*
		Treatment	2	22	10.1305	0.0008*
		Behavior × Treatment	6	66	0.9863	0.4418
Contralateral EMG	Post-hypoxia	Behavior	3	66	33.7499	<0.0001*
		Treatment	2	22	2.8593	0.0787
		Behavior × Treatment	6	66	0.9045	0.4972

Fixed effects include behavior (infusion: baseline, infusion, 0–5 min post-infusion, 15–20 min post-infusion; challenges: baseline, hypoxia, maxchemo; post-hypoxia: hypoxia, 5 × 10 min post-hypoxia, 20–25 mins post-hypoxia, 35–40 min post-hypoxia) and treatment (HPCD, CX717 or CX1739). df (between-groups degrees of freedom) and dfDen (within-groups degrees of freedom). Summary data presented in Figure 4.

^* denotes p < 0.05.

EMG, electromyography; C2Hx, C2 spinal hemisection model; HPCD, 2-hydroxypropyl-beta-cyclodextrin.

Rats treated with the HPCD vehicle did not increase ipsilateral EMG_peak during either of the acute respiratory challenges (Fig. 5B). In contrast, rats treated with CX717 or CX1739 showed robust increases in EMG_peak. Responses were similar in CX717 versus CX1739 treated rats (∼ 50% increase from baseline during both challenges). There was no difference in the contralateral diaphragm EMG_peak response to respiratory challenge across the three groups.

Evaluation of diaphragm activity following the acute hypoxia exposure revealed a difference between HPCD- and ampakine-treated rats (Fig. 5C). In rats treated with the HPCD vehicle, diaphragm EMG_peak dropped below the pre-hypoxia baseline values in both the ipsilateral (reduced by 20–25%) and contralateral (reduced by 10–15%) recordings. In contrast, both CX717 and CX1739 groups showed a short-term potentiation²⁴ of the ipsilateral EMG_peak, as shown by the elevated output at 5 min post-hypoxia. Further, the ipsilateral EMG_peak in ampakine-treated rats remained greater than values in the HPCD group up to 40 min post-hypoxia. Following hypoxia, the output of the contralateral diaphragm showed a clear tendency to be elevated in the ampakine-treated groups, with several outliers showing particularly robust increases.

Ventilation assessed using plethysmography: Impact of ampakines after acute injury (4 days)

Representative airflow traces recorded using whole-body plethysmography are shown in Figure 6A. Consistent with prior work,^25,26 the C2Hx injury reduced the inspiratory tidal volume at both the 4- and 14-day time points (Fig. 6B). Minute ventilation (mL/min/kg), respiratory rate (breath/min), and tidal volume (mL/breath/kg) data are presented in Table S3; statistical evaluations are summarized in Table 4.

FIG. 6.

Impact of C2 spinal hemisection (C2Hx) injury on tidal volume. (A) Representative breathing traces recorded using whole-body plethysmography from unanesthetized animals. Data are shown from spinal-intact and C2Hx animals at pre-injury and 4 and 14 days post-injury (or post-baseline for intact group) during eupneic breathing. Inspiration is indicated by an upward deflection in the trace. (B) Tidal volume and minute ventilation expressed as a % change from pre-injury values. Tidal volume is reduced by ∼20% and ∼15% in the C2Hx group at 4 and 14 days post-injury. Data are represented as mean ± standard error of animals for each time point (diamond) and individual data for each animal in the intact (gray dots, n = 7) and C2Hx groups (black dots, n = 10). Statistical summary for the effect of C2Hx ventilation is compiled in Table 4.

Table 4.

Statistical Summary for Eupneic Breathing Patterns after C2Hx Using Mixed Linear Model with Animal as Random Effect

Outcome	Effect	df	dfDen	F ratio	p value
Tidal volume	DPI	2	32	6.4541	0.0044*
	Group	1	16	13.6738	0.0020*
	DPI × Group	2	32	5.9713	0.0063*
Minute ventilation	DPI	2	32	5.9251	0.0065*
	Group	1	16	3.8168	0.0685
	DPI × Group	2	32	6.4745	0.0044*

Fixed effects include DPI (pre-injury, 4 DHx or 14 DHx) and group (intact or C2Hx). df (between-groups degrees of freedom) and dfDen (within-groups degrees of freedom). Summary data presented in Figure 5.

^* denotes p < 0.05.

C2Hx, C2 spinal hemisection model; DPI, days post-injury.

When assessing the impact of treatment on ventilation, HPCD infusion had no impact on tidal volume, whereas ampakines caused tidal volume to increase (Fig. 7). Specifically, CX717 triggered an acute increase (∼ 25%) that returned to baseline by 15 min. The CX1739 infusion increased tidal volume by ∼15%, and the values remained elevated for 15 min. Both the vehicle and ampakine solutions were associated with an increase in respiratory rate, suggesting a small non-specific response to tail vein infusion. The ampakines, however, produced an additional stimulation of respiratory rate, because the increase was considerably more pronounced in the CX717 (120%) and CX1739 (100%) groups versus the HPCD (60%) group.

FIG. 7.

Impact of ampakines on respiratory parameters after acute injury (4 days post-C2 spinal hemisection [C2Hx] injury). Mean tidal volume (left), respiratory rate (middle), and minute ventilation (right) measured at 4 days post-injury. Data following infusion of CX717, CX1739, or 2-hydroxypropyl-beta-cyclodextrin (HPCD) are expressed as mean ± standard error (diamond), with individual data points depicting each animal (CX717: blue, n = 9), (CX1739: red, n = 8), (HPCD: black, n = 8) groups. (A) The impact of ampakine or HPCD infusion during eupneic breathing was evaluated during infusion, at 0–5 min post-infusion, and at 15–20 min post-infusion. (B) The impact of ampakine or HPCD infusion during acute respiratory challenge: hypoxia (10.5% O₂) or maxchemo (7% CO₂, 10.5% O₂). (C) The impact of ampakine or HPCD infusion in conjunction with a single bout of hypoxia was evaluated at 5–10 min, 20–25 min, and 35-40 min post-hypoxia. Raw values of respiratory parameters are reported in Table S3. Statistical results are reported in Table 5. BL, baseline. Color image is available online.

We observed that rats treated with HPCD or CX717 showed similar increases in breathing when exposed to hypoxia or maxchemo (Fig. 7B, statistical evaluations in Table 5). However, rats in the groups treated with CX1739 showed a more robust response during both of the respiratory challenges. The increased ventilatory response in CX1739-treated rats was driven primarily by a greater increase in tidal volume during the challenges. Following hypoxia exposure, tidal volume and minute ventilation remained elevated for up to 40 min in rats treated with CX1739.

Table 5.

Statistical Summary for the Impact of Ampakines on Ventilation at 4 Days post-C2Hx Using Mixed Linear Model with Animal as Random Effect

Outcome	Segment	Interaction	df	dfDen	F ratio	p value
Tidal volume	Infusion	Behavior	3	66	50.2327	<0.0001*
		Treatment	2	22	3.2853	0.0564
		Behavior × Treatment	6	66	2.8442	0.0159*
Respiratory rate	Infusion	Behavior	3	66	13.45	<0.0001*
		Treatment	2	22	2.8993	0.0763
		Behavior × Treatment	6	66	2.6201	0.0244*
Minute ventilation	Infusion	Behavior	3	66	62.2818	<0.0001*
		Treatment	2	22	5.6427	0.0105*
		Behavior × Treatment	6	66	5.4203	0.0001*
Tidal volume	Challenges	Behavior	2	44	90.9378	<0.0001*
		Treatment	2	22	1.8815	0.1761
		Behavior × Treatment	4	44	1.2357	0.3097
Respiratory rate	Challenges	Behavior	2	44	234.531	<0.0001*
		Treatment	2	22	4.8859	0.0175*
		Behavior × Treatment	4	44	3.2605	0.0200*
Minute ventilation	Challenges	Behavior	2	44	150.647	<0.0001*
		Treatment	2	22	4.1583	0.0294*
		Behavior × Treatment	4	44	3.3102	0.0187*
Tidal volume	Post-hypoxia	Behavior	3	66	53.7255	<0.0001*
		Treatment	2	22	2.0668	0.1505
		Behavior × Treatment	6	66	0.7638	0.601
Respiratory rate	Post-hypoxia	Behavior	3	66	149.309	<0.0001*
		Treatment	2	22	2.737	0.0868
		Behavior × Treatment	6	66	0.5607	0.76
Minute ventilation	Post-hypoxia	Behavior	3	66	117.554	<0.0001*
		Treatment	2	22	3.0626	0.0671
		Behavior × Treatment	6	66	1.3057	0.267

Fixed effects include behavior (infusion: baseline, infusion, 0–5 min post infusion, 15–20 min post infusion; challenges: baseline, hypoxia, maxchemo; post-hypoxia: hypoxia, 5–10 min post-hypoxia, 20–25 min post hypoxia, 35–40 min post hypoxia) and treatment (HPCD, CX717, or CX1739). df (between-groups degrees of freedom) and dfDen (within-groups degrees of freedom). Summary data presented in Figure 6.

^* denotes p < 0.05.

C2Hx, C2 spinal hemisection model; HPCD, 2-hydroxypropyl-beta-cyclodextrin.

Ventilation assessed using plethysmography: Impact of ampakines after sub-acute injury (14 days)

We observed little to no acute impact of HPCD or CX1739 infusion on tidal volume (Fig 8, statistical evaluation in Table 6.). In contrast, infusion of CX717 caused an increase in tidal volume that remained elevated for 15 min (10–15%). Respiratory rate and minute ventilation increased transiently in response to infusion in all groups, but the response was markedly greater in CX717-treated rats.

Table 6.

Statistical Summary for the Impact of Ampakines on Ventilation at 14 Days post-C2Hx Using Mixed Linear Model with Animal as Random Effect

Outcome	Segment	Interaction	df	dfDen	F ratio	p value
Tidal volume	Infusion	Behavior	3	66	25.0161	<0.0001*
		Treatment	2	22	0.9938	0.3862
		Behavior × Treatment	6	66	2.1674	0.0572
Respiratory rate	Infusion	Behavior	3	66	5.8148	0.0014*
		Treatment	2	22	4.1338	0.0299*
		Behavior × Treatment	6	66	2.096	0.0654
Minute ventilation	Infusion	Behavior	3	66	25.0161	<0.0001*
		Treatment	2	22	0.9938	0.3862
		Behavior × Treatment	6	66	2.1674	0.0572
Tidal volume	Challenges	Behavior	2	44	94.7852	<0.0001*
		Treatment	2	22	2.2633	0.1277
		Behavior × Treatment	4	44	1.2378	0.3088
Respiratory rate	Challenges	Behavior	2	44	140.877	<0.0001*
		Treatment	2	22	0.4005	0.6748
		Behavior × Treatment	4	44	0.3652	0.8321
Minute ventilation	Challenges	Behavior	2	44	121.464	<0.0001*
		Treatment	2	22	1.0918	0.3531
		Behavior × Treatment	4	44	0.834	0.5109
Tidal volume	Post-hypoxia	Behavior	2	44	121.464	<0.0001*
		Treatment	2	22	1.0918	0.3531
		Behavior × Treatment	4	44	0.834	0.5109
Respiratory rate	Post-hypoxia	Behavior	3	66	48.9214	<0.0001*
		Treatment	2	22	2.6688	0.0917
		Behavior × Treatment	6	66	0.7721	0.5946
Minute ventilation	Post-hypoxia	Behavior	3	66	48.3479	<0.0001*
		Treatment	2	22	2.8687	0.0781
		Behavior × Treatment	6	66	0.994	0.4368

Fixed effects include behavior (infusion: baseline, infusion, 0–5 min post-infusion, 15–20 min post–infusion; challenges: baseline, hypoxia, maxchemo; post–hypoxia: hypoxia, 5–10 min post-hypoxia, 20–25 mins post-hypoxia, 35–40 min post-hypoxia) and treatment (HPCD, CX717 or CX1739). df (between-groups degrees of freedom) and dfDen (within-groups degrees of freedom). Summary data presented in Figure 7.

^* denotes p < 0.05.

C2Hx, C2 spinal hemisection model; HPCD, 2-hydroxypropyl-beta-cyclodextrin.

We did not detect any differences in the tidal volume response across the HPCD-, CX717-, and CX1739-treated rats (Fig. 8B) during respiratory challenges. However, both CX717 and CX1739 were associated with increased respiratory rate as compared with the HPCD group. As a result, minute ventilation was also higher in the CX717 and CX1739 groups.

FIG. 8.

Impact of ampakines on respiratory parameters after sub-acute injury (14 days post- C2 spinal hemisection [C2Hx] injury). Mean tidal volume (left), respiratory rate (middle), and minute ventilation (right) at 14 days post-hemisection following infusion of CX717, CX1739, or 2-hydroxypropyl-beta-cyclodextrin (HPCD). Data are expressed as mean ± standard error for each time point (diamond), with individual data points as well: (CX717: blue, n = 9), (CX1739: red, n = 8), and (black: HPCD, n = 8). (A) The impact of ampakine or HPCD infusion on ventilation during room air exposure was evaluated up to 20 min post-infusion. (B) The impact of ampakine or HPCD infusion on ventilation was evaluated during acute respiratory challenge: hypoxia (10.5% O₂) or maxchemo (7% CO₂, 10.5% O₂). (C) The impact of ampakine or vehicle pre-treatment followed by a 5 min bout of hypoxic exposure was evaluated at 5–10 min, 20–25 min, and 35–40 min post-hypoxia. Raw values of respiratory parameters are reported in Table S3. Statistical results are reported in Table 6. BL, baseline. Color image is available online.

Following hypoxia, a few rats treated with CX717 showed a persistent elevation of ventilation, but on average, values for minute ventilation, tidal volume, and breathing frequency returned to baseline values in CX717-treated rats (Fig. 8C). At 5-min post-hypoxia, CX1739- treated rats showed a short-term potentiation of minute ventilation, similar to the CX717 group. However, by 20 min post-hypoxia, values were depressed compared with baseline recordings in both the HPCD and CX1739 groups.

Discussion

Respiratory complications are a leading cause of morbidity and mortality after cervical SCI.³ In this study, we show that treatment with a positive allosteric AMPA receptor modulator effectively increases diaphragm EMG activity after acute or sub-acute cervical SCI. The therapeutic impact of ampakines was not limited to “quiet breathing,” and was also evident during respiratory challenges that require considerably greater diaphragm activation. Collectively, our data indicate that ampakine doses at or below levels previously used in clinical studies¹⁴ can augment diaphragm muscle activation after cervical SCI, and therefore may be a viable pharmacological strategy, either alone or as an adjunctive therapeutic approach for respiratory rehabilitation after SCI.

Diaphragm EMG activity during “eupnea” and the impact of ampakines

During baseline conditions of eupneic breathing, rhythmic diaphragm EMG activity was present in both hemi-diaphragms post-C2Hx. Albeit, EMG_peak activity on the ipsilesional side was ∼40% lower than the pre-injury value, and this deficit allowed us to test the hypothesis that ampakines could stimulate impaired neuromuscular respiratory output after SCI. Consistent with the hypothesis, both ampakines (CX717 and CX1739) caused immediate increases in ipsilateral diaphragm EMG at both 4 and 14 days post-injury. Prior work shows that ampakines can potentiate motor neuron currents, as well as altering respiratory rate by targeting supra-spinal rhythm generating centers.^10,27 In the current study, we observe both a robust acute change in diaphragm EMG amplitude and some degree of increased respiratory rate following ampakine infusion. Although, mechanistically, the extent of impact of ampakines on spinal versus supra-spinal respiratory centers is unclear, it appears that the acute response to ampakines is a result of potentiation at both centers. Further, the sustained effect of ampakines is largely restricted to respiratory centers with a depressed neural drive,²⁸ such as spinal phrenic motor neurons that have considerable disruption of excitatory activation following C2Hx. This mechanism of ampakine-mediated potentiation of “sub-threshold” excitatory inputs is also likely contributing to the differential efficacy of ampakines at the ipsilateral versus contralateral diaphragm. In addition to descending supra-spinal excitatory inputs to phrenic motor neurons, there also exists a complex spinal network of excitatory inputs to the phrenic pool that ampakines could act upon. For example, numerous studies have shown the presence of excitatory interneuronal projections from the upper cervical segments to the phrenic pool that modulate motor neuron excitability.^2,13,29–31 Further, there also exists an excitatory propriospinal network in the spinal cord, originating from thoracic segments of the spinal cord and receiving inputs from intercostal muscle proprioceptors.^32,33 Although the current study design cannot tease apart the relative contribution of each source of excitatory input in supporting the action of ampakines, they each represent candidate populations as a site of action for ampakines.

Diaphragm EMG activity during respiratory challenge

The ability to generate higher trans-diaphragmatic forces for airway clearance is almost always impaired after cervical SCI.³⁴ In turn, this has devastating health consequences, including increased incidence of pneumonia.³⁵ We tested diaphragm activation during higher force behaviors using a brief maxchemo ventilatory challenge. At both the acute and sub-acute post-injury time points, rats treated with CX717 and CX1739 had improved ability to increase diaphragm EMG during the respiratory challenge. Mechanistically, the ampakine treatment may enable activation of sub-threshold AMPA receptors on phrenic motor neurons, allowing a larger proportion of diaphragm motor units to be recruited. This effect was not observed in the contralateral diaphragm, which is already recruiting the full spectrum of motor units because of the largely intact spinal circuitry contralateral to C2Hx.

Effect of ampakine on tidal volume

The diaphragm EMG recordings allowed us to conclusively test the impact of ampakines on the main target muscle, the diaphragm. We also indirectly evaluated ventilation using the method of whole-body barometric plethysmography. Humidification and warming of inspired air as it enters the lung produces changes in plethysmography chamber pressure, which enables an estimation of tidal volume and ventilation.³⁶ Plethysmography data indicated that CX1739 was able to stimulate tidal volume at the acute time point post-injury, both during eupneic and challenged breathing. CX717 seemed to be more effective at the sub-acute time point, where it also increased tidal volume post infusion. Therefore, it is clear that ampakine treatment was able to increase the tidal volume measurements, but that these measures did not always parallel the observed changes in diaphragm EMG. There are physiological and methodological reasons that this may not occur. First, activation of muscles other than the diaphragm contributes to plethysmography pressure waveforms.³⁷ Further, there is considerable “compensatory” reorganization in the intact respiratory muscles following injury or disease.^38,39 This aspect is highlighted in the acute impact of C2Hx on eupneic tidal volume, which starts to recover by 14 days post-C2Hx. It is likely that other muscles (such as intercostals) are now being recruited to accomplish pre-injury levels of inspiratory effort.⁴⁰

It is also interesting to note that CX1739 and CX717 were more effective in stimulating ventilation, assessed with plethysmography, in the acute versus the sub-acute stages, respectively. Similar trends were observed in the EMG data. This is likely because other respiratory muscles are now better integrated into the respiratory system and respond better to increases in neural drive during challenged breathing in this sub-acute phase of injury, masking any ampakine-mediated gains in motor neuron firing at this stage. Nevertheless, the data support our conclusion that ampakines can stimulate overall ventilation, and this lays the groundwork for investigating other combinatorial strategies to amplify respiratory output following injury.

Combined effect of ampakine pre-treatment and hypoxia

Carefully controlled brief exposures to hypoxia can have therapeutic benefits after SCI, including sustained increases in volitional muscle strength and EMG activity.^41,42 Several studies have evidence that an interaction of ampakine treatment with hypoxic exposures causes a sustained increase in respiratory motor output.^9,43 Although such a motor facilitation is generally expressed by exposing the animals to multiple bouts of hypoxia, ampakine pre-treatment can reduce the required number of hypoxia bouts to just one. We also recently showed that in animals with C2Hx, this phenomenon could be replicated on the ipsilateral phrenic nerve, wherein animals treated with ampakines and exposed to a single bout of hypoxia showed enhanced phrenic facilitation.¹⁰ The current study presents evidence in support of short-term facilitation in diaphragm EMG output, selectively in ampakine-treated animals exposed to a single bout of hypoxia. These results are particularly important because they raise the possibility of hypoxia in conjunction with ampakine treatment as a rehabilitation modality in SCI patients.

Conclusion

Ampakines are metabolically stable and cross the blood–brain barrier.^44–47 CX717 has been given to humans with minimal side effects and no adverse events.^14,44,47 Clinical studies aimed at enhancing cognitive function^14,44 or alleviating opioid-induced respiratory depression⁴⁷ used CX717 doses of 13–20 mg/kg. Our study provides evidence that even lower ampakine doses can effectively stimulate breathing and diaphragm muscle activity in a pre-clinical model of cervical SCI, with no evidence of adverse effects. Further, ampakine-treated animals are capable of increasing respiratory motor drive to a larger degree when challenged, a response often dampened in SCI patients. Lastly, the current data suggest that pairing low-dose and low impact ampakines with even a single brief hypoxia exposure may have value in the context of neurorehabilitation paradigms.⁴¹

Author Contributions

S.R., J.J.G., and D.D.F. were responsible for the conception or design of the work; S.R. was responsible for the acquisition of data; S.R., M.D.S., and D.D.F. were responsible for the analysis or interpretation of data for the work; S.R. and D.D.F. were responsible for drafting the work or revising it critically for important intellectual content; S.R., M.D.S., J.J.G., and D.D.F. were responsible for the final approval of the version to be published; and S.R., M.D.S., J.J.G., and D.D.F. agreed to be accountable for all aspects of the work.

Funding Information

This work was supported by the National Institutes of Health (NIH) 1R01HL139708-01A1 (D.D.F.) and the Craig H. Neilsen Foundation (S.R.).

Author Disclosure Statement

Dr. Greer holds a patent related to using ampakines for treating respiratory depression in neuromuscular disease (patent number US8039468B2). The other authors declare no competing financial interests.

Supplementary Material

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Table S4