Phrenicotomy alters phrenic long-term facilitation following intermittent hypoxia in anesthetized rats

Abstract

Intermittent hypoxia (IH) can induce a persistent increase in neural drive to the respiratory muscles known as long-term facilitation (LTF). LTF of phrenic inspiratory activity is often studied in anesthetized animals after phrenicotomy (PhrX), with subsequent recordings being made from the proximal stump of the phrenic nerve. However, severing afferent and efferent axons in the phrenic nerve has the potential to alter the excitability of phrenic motoneurons, which has been hypothesized to be an important determinant of phrenic LTF. Here we test the hypothesis that acute PhrX influences immediate and long-term phrenic motor responses to hypoxia. Phrenic neurograms were recorded in anesthetized, ventilated, and vagotomized adult male rats with intact phrenic nerves or bilateral PhrX. Data were obtained before (i.e., baseline), during, and after three 5-min bouts of isocapnic hypoxia. Inspiratory burst amplitude during hypoxia (%baseline) was greater in PhrX than in phrenic nerve-intact rats (P < 0.001). Similarly, burst amplitude 55 min after IH was greater in PhrX than in phrenic nerve-intact rats (175 ± 9 vs. 126 ± 8% baseline, P < 0.001). In separate experiments, phrenic bursting was recorded before and after PhrX in the same animal. Afferent bursting that was clearly observable in phase with lung deflation was immediately abolished by PhrX. The PhrX procedure also induced a form of facilitation as inspiratory burst amplitude was increased at 30 min post-PhrX (P = 0.01 vs. pre-PhrX). We conclude that, after PhrX, axotomy of phrenic motoneurons and, possibly, removal of phrenic afferents result in increased phrenic motoneuron excitability and enhanced LTF following IH.

exposure to intermittent hypoxia (IH) over short periods (e.g., minutes to hours) can evoke a persistent increase in respiratory motor output, termed long-term facilitation (LTF) (35, 48, 50). LTF has been observed in awake (26, 35) and sleeping humans (53) and in a wide range of animal species (22, 46, 48). The mechanisms underlying LTF have been studied extensively over the last 10–15 years, primarily in anesthetized animals (reviewed in Refs. 42, 49, 50). In anesthetized animals, LTF is typically manifest as an increase in the inspiratory burst amplitude of phrenic (19) and/or hypoglossal (XII) extracellular nerve recordings (20). Recordings of in vivo (19) and in vitro (10) respiratory motor LTF are typically made from cut respiratory muscle nerves. More specifically, the phrenic and/or XII nerves are cut, and subsequent extracellular recordings are made from the central end of the nerve. This procedure can provide stability to the neurophysiological recording procedures, particularly when a dorsal surgical approach is used, but also has the potential to alter respiratory motor output and the subsequent expression of plasticity. For example, although LTF of phrenic burst amplitude is typically robust in anesthetized rats with cut phrenic nerves [i.e., phrenicotomy (PhrX)] (7, 22), LTF of diaphragm electromyogram (EMG) activity is absent in anesthetized, spontaneously breathing rats with intact phrenic nerves (30). This difference probably does not reflect vagally mediated inhibition of LTF during spontaneous breathing, because Golder and Martinez (24) demonstrated that, under otherwise similar conditions, vagotomized rats have substantially diminished phrenic LTF compared with vagal-intact rats. Similarly, LTF of inspiratory volume [which correlates with phrenic burst amplitude (16)] is reduced in unanesthetized, spontaneously breathing animals (52) compared with phrenic LTF in anesthetized PhrX rats (7, 22). Consistent with these observations, studies in humans show ventilatory LTF that is substantially less than is typically seen in PhrX, anesthetized rats (35). Collectively, these observations are consistent with the notion that PhrX may create preconditions that enhance the subsequent expression of LTF of phrenic burst amplitude following IH.

At least two potential mechanisms support the hypothesis that PhrX alters the expression of LTF in the phrenic nerve. 1) Axotomy can alter neuronal excitability. Indeed, a substantial body of evidence shows that chronic axotomy increases neuronal excitability, as indicated by a decrease in the amount of electrical current needed to evoke an action potential (i.e., decreased rheobase current) (reviewed in Ref. 69). Studies of acute axotomy reveal relatively rapid changes in the morphology of mammalian neurons (13) and increases in intracellular Ca²⁺ concentration in invertebrate (59) and vertebrate (45) axons. Accordingly, phrenic motoneuron excitability may be increased after PhrX. 2) Removal of afferent signals, which normally travel in the phrenic nerve, could alter the neural control of phrenic motoneurons (17, 56). Indeed, afferent fibers make up 40–45% of the phrenic nerve (33, 34), and these afferents project to spinal (25, 61) and supraspinal (12, 62) structures. Although a direct role of phrenic afferents in modulating the excitability of phrenic motoneurons has not been definitively shown, there is indirect evidence suggesting that afferents can reflexly affect the phrenic motor drive (15, 23, 29, 55, 63).

Our purpose was to test the hypotheses that 1) IH-induced LTF of phrenic inspiratory burst amplitude is greater in rats with cut than intact phrenic nerves and 2) acute PhrX influences efferent phrenic bursting during “eupneic” (baseline) conditions.

METHODS

Animals

All procedures were approved by the University of Florida Institutional Animal Care and Use Committee. Adult male Sprague-Dawley rats (Harlan, Indianapolis, IN) were divided into four groups: 1) bilateral PhrX with IH (PhrX-LTF, n = 8), 2) phrenic nerves intact with IH (PhrI-LTF, n = 9), 3) bilateral PhrX with sham hypoxia (i.e., a “time control”; PhrX-Sham, n = 7), and 4) phrenic nerves intact with sham hypoxia (PhrI-Sham, n = 8). Six additional rats were used to investigate the effect of acute PhrX on phrenic output.

Experimental Preparation

The general procedures have been described recently (14, 36). Rats were anesthetized with isoflurane (5% in 100% O₂) in a closed chamber and then transferred to a nose cone (2–3% isoflurane in 50% O₂-balance N₂). The trachea was cannulated with polyethylene (PE-240) tubing, and rats were mechanically ventilated for the remainder of the experiment. The lungs were briefly hyperinflated (2–3 s) approximately once per hour to minimize atelectasis. The tracheal pressure was monitored with a pressure transducer (Statham P-10EZ pressure transducer and CP122 AC/DC strain gauge amplifier, Grass Instruments, West Warwick, RI) connected to the tracheal cannula. A catheter (PE-50) was inserted into the femoral vein, and the anesthesia was switched from isoflurane to urethane (1.6 g/kg iv, 0.12 g/ml distilled water). During this period, the limb withdrawal response to toe pinch was monitored to ensure the adequacy of anesthesia, and supplemental urethane was given if indicated (0.3 g/kg iv). A femoral arterial catheter (PE-50) was inserted to measure blood pressure (Statham P-10EZ pressure transducer and CP122 AC/DC strain gauge amplifier, Grass Instruments) and to periodically withdraw blood samples (see Experimental Protocols).

Rats were bilaterally vagotomized to prevent entrainment of phrenic motor output with the ventilator and paralyzed with pancuronium bromide (2.5 mg/kg iv) to eliminate spontaneous breathing efforts. After paralysis, blood pressure and phrenic nerve response to toe pinch were monitored to ensure the adequacy of anesthesia. A slow infusion (1.5 ml/h) of 3:1 lactated Ringer solution-sodium bicarbonate was maintained to promote acid-base balance (4, 40). Arterial Po₂ (Pa_O2) and Pco₂ (Pa_CO2), as well as pH, were determined from 0.2-ml arterial blood samples using an i-Stat blood gas analyzer (Heska, Fort Collins, CO). Blood gas and pH values were corrected to the rectal temperature measured at the time the blood sample was obtained. Throughout the protocol, the end-tidal Pco₂ (Pet_CO2) was measured using a rapidly responding mainstream CO₂ analyzer positioned a few centimeters from the tracheostomy tube on the expired line of the ventilator circuit (Capnogard CO₂ monitor, Novametrix Medical Systems, Wallingford, CT). Rectal temperature was maintained within 37 ± 1°C (see results) using a rectal thermistor connected to a servo-controlled heating pad (model TC-1000, CWE, Ardmore, PA).

Both phrenic nerves were isolated within the caudal neck region medial to the brachial plexus via a ventral surgical approach. Electrical activity was recorded using silver wire electrodes with a monopolar configuration, amplified (1,000×), and filtered (band pass = 300–10,000 Hz, notch = 60 Hz) using a differential AC amplifier (model 1700, A-M Systems, Carlsborg, WA). The amplified signal was full-wave rectified and moving averaged (time constant = 100 ms; model MA-1000, CWE). Data were digitized using a Power 1401 data acquisition interface (Cambridge Electronic Design, Cambridge, UK) and recorded on a personal computer using Spike2 software (Cambridge Electronic Design).

Experimental Protocols

LTF.

The LTF protocol was similar to that described in prior studies (3, 22, 31). After an adequate plane of anesthesia was established, Pet_CO2 was maintained at 40 ± 2 Torr for 5–10 min; the inspired O₂ fraction (Fi_O2) was held at 0.5. For determination of the Pet_CO2 apneic threshold for inspiratory phrenic activity, the ventilator pump rate gradually increased until inspiratory bursting ceased in both phrenic nerves. Apnea was maintained for 1 min, and the ventilator rate was then gradually decreased until inspiratory bursting resumed in the phrenic nerves. The Pet_CO2 associated with the onset of inspiratory bursting was noted, and the ventilator rate was adjusted to maintain Pet_CO2 at 2 Torr above this value throughout the experiment. Pet_CO2 measurements, however, were merely a guide to help maintain isocapnia, and conclusions regarding isocapnia were determined exclusively by arterial blood analyses. After a 10- to 20-min baseline period, an arterial blood sample was drawn. Rats were then exposed to either three 5-min episodes of hypoxia (Fi_O2 = 0.14–0.16) separated by 5 min of hyperoxic recovery or sham hypoxia (Fi_O2 same as baseline). Blood samples were obtained during the first episode of hypoxia and 25 and 55 min after hypoxia. At the conclusion of the protocol, the animals were exposed to another 5-min episode of hypoxia, in an effort to confirm the integrity of the nerve-electrode contact. A >10% decrease in phrenic burst amplitude relative to the initial hypoxic response was taken as an indicator that the nerve-electrode contract was unstable. On the basis of this criterion, data from a single rat were excluded from the final analyses.

Influence of PhrX on phrenic bursting.

The intact phrenic nerves were placed on electrodes, and the nerve-electrode contact area was covered with a silicone elastomer (Kwik-Sil, World Precision Instruments, Sarasota, FL). This procedure ensured that the contact was preserved during the subsequent PhrX. Once a stable phrenic recording was obtained, PhrX was performed while electrical activity was being recorded from the phrenic nerves. In four of six rats, arterial blood samples were taken before and after the PhrX procedure. These blood samples were drawn at the end of the baseline period, and Pet_{CO2 was maintained at pre-PhrX level for 60 min. Another blood sample was drawn at the end of the recording to confirm isocapnia relative to pre-PhrX.}

Data Analysis

Phrenic neurograms were analyzed as described in our recent publication (36). Peak integrated phrenic amplitude (∫Phr) and burst frequency, averaged over 1 min for each recorded data point, are expressed in millivolts [i.e., arbitrary units (AU)] and bursts per minute, respectively, and are also normalized to values recorded during baseline. Statistical tests were done using SigmaStat version 2.03 software. Differences in phrenic LTF and hypoxic responses and in arterial blood gases between groups (e.g., PhrX-LTF vs. PhrI-LTF and PhrX-LTF vs. PhrX-Sham) were determined using a two-way repeated-measures ANOVA. The Student-Newman-Keuls test was used for post hoc analyses. The effects of acute PhrX on phrenic output were examined using a one-way repeated-measures ANOVA. Body weight was compared between groups using one-way ANOVA. Differences were considered statistically significant when P ≤ 0.05. All data are presented as means ± SE.

RESULTS

Body weight was similar between the five experimental groups: 410 ± 9, 419 ± 7, 415 ± 10, 411 ± 13, and 407 ± 11 g in PhrX-LTF, PhrI-LTF, PhrX-Sham, PhrI-Sham, and acute-PhrX, respectively (P = 0.91). Pa_O2 decreased during hypoxia in the LTF groups, as expected (P < 0.001). Rectal temperature did not change by >0.6°C over the course of the experimental protocol in any animal, and mean temperatures were similar between groups (baseline values = 37.4 ± 0.1, 37.5 ± 0.1, 37.3 ± 0.1, 37.3 ± 0.2, and 37.4 ± 0.1°C in PhrX-LTF, PhrI-LTF, PhrX-Sham, PhrI-Sham, and acute-PhrX, respectively, P = 0.68). The Pet_CO2 recruitment threshold for phrenic bursting was similar between PhrI and PhrX rats: 41 ± 1 and 42 ± 1 Torr, respectively (P = 0.54). Baseline and posthypoxia Pa_O2 values were not different between groups (Table 1). Pa_CO2 was also similar between groups and was maintained within 2 Torr of baseline values throughout the IH or sham protocols (Table 1). The mean arterial blood pressure (MAP) tended to be lower in PhrX-LTF rats than in the other groups. However, this effect was only statistically significant compared with the sham group (Table 1). MAP dropped transiently during hypoxic episodes (P < 0.01), as previously reported in anesthetized rats (8, 18). The relative decrease in MAP during hypoxia was similar in PhrX-LTF and PhrI-LTF rats: −20 ± 3 and −28 ± 6% baseline, respectively (P = 0.25). Similar to prior phrenic LTF studies (8, 14, 41), MAP also tended to decrease slightly over the course of the experimental protocol; however, statistical significance was reached in only PhrI-LTF rats (P = 0.035; Table 1).

Table 1.

MAP, Pa_CO2, Pa_O2, and arterial pH at baseline, first hypoxic episode, and 25 and 55 min posthypoxia

			Posthypoxia
	Baseline	Hypoxia	25 min	55 min
MAP, mmHg
PhrX-LTF	86 ± 5†	66 ± 7*†	85 ± 5	84 ± 4
PhrI-LTF	107 ± 6	78 ± 7*†	103 ± 8	93 ± 8*
PhrX-Sham	111 ± 13	109 ± 11	107 ± 9	107 ± 7
PhrI-Sham	114 ± 9	113 ± 9	101 ± 10	104 ± 9
PaCO2, torr
PhrX-LTF	39 ± 3	37 ± 3	38 ± 3	38 ± 3
PhrI-LTF	41 ± 2	39 ± 2	42 ± 2	41 ± 2
PhrX-Sham	37 ± 4	35 ± 5	39 ± 3	37 ± 4
PhrI-Sham	35 ± 2	36 ± 3	35 ± 3	36 ± 3
PaO2, torr
PhrX-LTF	163 ± 13	34 ± 1*†	155 ± 9	162 ± 9
PhrI-LTF	162 ± 8	33 ± 1*†	152 ± 10	155 ± 11
PhrX-Sham	142 ± 11	152 ± 11	145 ± 14	158 ± 11
PhrI-Sham	185 ± 8	181 ± 9	173 ± 11	173 ± 11
pH
PhrX-LTF	7.37 ± 0.02	7.34 ± 0.02*	7.35 ± 0.03*	7.34 ± 0.03*
PhrI-LTF	7.35 ± 0.01	7.33 ± 0.01*	7.33 ± 0.01*	7.32 ± 0.01*
PhrX-Sham	7.38 ± 0.02	7.39 ± 0.02	7.40 ± 0.02	7.41 ± 0.02
PhrI-Sham	7.36 ± 0.02	7.35 ± 0.02	7.36 ± 0.02	7.38 ± 0.02

Values are means ± SE. MAP, mean arterial pressure; Pa_CO2 and Pa_O2, arterial Pco₂ and Po₂; PhrX, Phrenicotomy; PhI, phenic nerve intact; LTF, long-term facilitation.

^*Different from baseline.

^†Different from corresponding sham data point.

Baseline Phrenic Activity

Figure 1 shows representative neurograms recorded during baseline conditions. An afferent burst that was distinct from the inspiratory phrenic burst was clearly distinguishable in 15 of 17 (88%) PhrI rats. The afferent burst occurred in phase with lung deflation in all cases (Fig. 1) and was immediately abolished by acute PhrX (see below). We observed anecdotally that recordings from the distal stump of the phrenic nerve after PhrX (i.e., removing all efferent bursting) showed a clear rhythmic burst during lung deflation (Fig. 1C). Phrenic inspiratory burst frequency was similar between groups during baseline (47 ± 2 and 49 ± 2 bursts/min in PhrX and PhrI, respectively). The baseline ∫Phr amplitude (AU) tended to be greater in PhrX than PhrI rats, but the difference did not reach statistical significance (t-test, P = 0.076).

Fig. 1.

Representative examples of phrenic neurograms. Moving-averaged or “integrated” phrenic signal (∫Phr) is presented above unprocessed or “raw” signal (Phr). P_Insp, pressure recorded in the inspired line of the ventilator circuit. A: recordings obtained from an intact phrenic nerve (PhrI). Signal is composed of 2 distinct and phasic bursts: larger burst is the typical inspiratory burst, and smaller bursts (arrowheads) occurred in phase with lung deflation, as reflected by P_Insp (dashed vertical lines in C). These afferent bursts were completely eliminated following phrenicotomy (PhrX, B). C: additional recording obtained from a single rat from the distal stump of the cut phrenic nerve. In this example, rhythmic activity cannot reflect activity of phrenic motoneurons and must reflect activity in afferent neurons within the phrenic nerve.

Intermittent Hypoxia

Neurograms depicting a typical response to IH are provided in Fig. 2, A and B. Hypoxic episode 1 evoked a significant increase in ∫Phr inspiratory burst frequency in both groups, as expected. However, burst frequency assessed during the initial 30 s (onset) of hypoxic episode 1 was significantly greater in PhrX-LTF than PhrI-LTF rats: 174 ± 11 vs. 145 ± 9% baseline (P = 0.006; Fig. 2C). The hypoxia onset phrenic burst frequency showed a decrease over subsequent hypoxic episodes (Fig. 2C). In addition, differences in the hypoxia onset burst frequency between PhrX-LTF and PhrI-LTF groups were not observed during hypoxic episodes 2 and 3 (Fig. 2C). Burst frequency assessed over the final minute of hypoxia was similar across all three hypoxic episodes in PhrX-LTF and PhrI-LTF rats (Fig. 2D).

Fig. 2.

Representative examples of phrenic neurograms (A and B) and mean ∫Phr activity during intermittent hypoxia (C–E). ∫Phr signal [arbitrary units (AU)] is shown along with the corresponding arterial blood pressure (P_arterial) trace (mmHg) during 3 bouts of isocapnic hypoxia in rats with phrenic nerves cut (PhrX-LTF, A) and rats with phrenic nerves intact (PhrI, B). Top: ∼35 min of data, including baseline and each of the 3 hypoxic episodes (H1, H2, H3). Middle: ∫Phr bursting with an expanded time scale. These recordings were taken during baseline (i and iv) and at onset (ii and v) and end (iii and vi) of the initial hypoxic episode. C–E: mean ∫Phr burst frequency during the onset of hypoxia (C) and frequency (D) and peak amplitude (E) at the end of hypoxia from PhrX-LTF and PhrI-LTF rats. *P < 0.02 vs. PhrI-LTF group. #P < 0.05 vs. H1.

The amplitude of the ∫Phr burst during hypoxia tended to be higher in PhrX-LTF than PhrI-LTF rats when the data were expressed as arbitrary units (P = 0.07, data not shown). However, the increase in ∫Phr amplitude during hypoxia (%baseline) was significantly greater in PhrX-LTF than PhrI-LTF rats (P < 0.01; Fig. 2E). PhrX-LTF and PhrI-LTF rats tended to show a slight progressive augmentation (54) of ∫Phr amplitude during successive hypoxic episodes (Figs. 2E), but this was statistically significant only in the PhrX rats (P < 0.05). The sham or time-control animals showed no change in ∫Phr amplitude or frequency during mock hypoxia exposures, as expected (data not shown).

Phrenic LTF

Representative examples of phrenic neurograms recorded during baseline and the posthypoxia period are shown in Fig. 3. Phrenic burst frequency [bursts/min (Fig. 4A) and %baseline (Fig. 4B)] was stable during the post-IH period, with no evidence for an increase relative to baseline values. Although no “frequency LTF” was observed, a more detailed analysis of the respiratory cycle indicated that changes in the timing of the phrenic bursts may have occurred, particularly in the PhrX-LTF group. Compared with baseline, inspiratory and expiratory durations (Ti and Te, respectively) tended to be lower after IH in PhrX-LTF and PhrI-LTF rats, but statistical significance was observed only in Ti at 55 min posthypoxia in PhrX-LTF animals (P < 0.05; Table 2). Values of Ti were also, as expected, significantly lower during hypoxia in PhrI-LTF and PhrX-LTF groups than the corresponding sham data points (P < 0.001; Table 2). Analyses of post-IH ∫Phr burst amplitude revealed a significant interaction between treatment (i.e., PhrX-LTF or PhrI-LTF) and time (i.e., 25 or 55 min; P < 0.001; Fig. 4C). Thus, differences in ∫Phr LTF (%baseline) between PhrX-LTF and PhrI-LTF rats were more pronounced at 55 than at 25 min posthypoxia (Fig. 4C). However, ∫Phr amplitude was significantly higher (P < 0.05) in PhrX-LTF rats at both time points (Fig. 4C). Importantly, ∫Phr amplitude was significantly greater in PhrX-LTF and PhrI-LTF rats than in the corresponding sham groups at 55 min post-IH: 106 ± 5 and 105 ± 3% baseline in PhrX-Sham and PhrI-Sham, respectively (both P < 0.05 vs. corresponding LTF group). Therefore, LTF of ∫Phr amplitude was induced in PhrX-LTF and PhrI-LTF rats, but the relative magnitude was much greater in the PhrX-LTF group. We believed that it was more appropriate to assess LTF by comparing ∫Phr amplitude with the values obtained in the corresponding sham group. However, comparison of the posthypoxia ∫Phr amplitude (%baseline) with baseline values (i.e., 100% in all rats) also resulted in the same conclusions. Specifically, relative to baseline, LTF was significant at 55 min in PhrX-LTF and PhrI-LTF groups but was not present in either sham group.

Fig. 3.

Representative examples of ∫Phr bursting and arterial blood pressure during baseline and at 25 and 55 min following intermittent hypoxia. Robust long-term facilitation (LTF) of ∫Phr burst amplitude was seen in rats with cut phrenic nerves (PhrX-LTF, top). A smaller, but statistically significant, LTF (see Fig. 4) was seen in PhrI-LTF rats (bottom).

Fig. 4.

Mean ∫Phr burst frequency and amplitude during LTF protocols in PhrX-LTF and PhrI-LTF rats. A: burst frequency, shown as bursts/min, during baseline (BL) and 25 and 55 min posthypoxia. B and C: frequency and peak ∫Phr burst amplitude relative to baseline values. *P < 0.01 vs. PhrI-LTF group. #P < 0.05 vs. 25 min.

Table 2.

Ti and Te

			Posthypoxia
	Baseline	Hypoxia	25 min	55 min
Ti, s
PhrX-LTF	0.38 ± 0.03	0.25 ± 0.01*†	0.35 ± 0.02	0.32 ± 0.02*
PhrI-LTF	0.35 ± 0.02	0.23 ± 0.01*†	0.33 ± 0.02	0.33 ± 0.01
PhrX-Sham	0.38 ± 0.02	0.37 ± 0.03	0.39 ± 0.03	0.36 ± 0.03
PhrI-Sham	0.39 ± 0.02	0.40 ± 0.02	0.36 ± 0.01	0.36 ± 0.01
Te, s
PhrX-LTF	0.99 ± 0.07	0.88 ± 0.04	0.87 ± 0.04	0.89 ± 0.04
PhrI-LTF	0.94 ± 0.08	0.93 ± 0.05	0.92 ± 0.05	0.93 ± 0.05
PhrX-Sham	0.88 ± 0.07	0.91 ± 0.1	0.90 ± 0.09	0.81 ± 0.07
PhrI-Sham	0.86 ± 0.07	0.88 ± 0.06	0.90 ± 0.06	0.91 ± 0.05

		Post-PhrX
	Pre-PhrX	1 min	30 min	60 min
Ti, s	0.35 ± 0.02	0.35 ± 0.02	0.36 ± 0.02	0.36 ± 0.03
Te, s	0.98 ± 0.12	0.95 ± 0.09	1 ± 0.05	0.97 ± 0.07

Values are means ± SE. Ti and Te, inspiratory and expiratory duration.

^*Different from baseline.

^†Different from corresponding sham data point.

We also quantified changes in the amplitude of the afferent burst in the PhrI-LTF rats following IH. This analysis indicated that the afferent burst was stable (i.e., no “afferent LTF”) following IH. Specifically, the amplitude of the burst was 97 ± 4 and 89 ± 8% baseline at 25 and 55 min posthypoxia, respectively (both P > 0.05).

Impact of Acute PhrX on ∫Phr Bursting

We performed additional studies in which we recorded phrenic activity before and for 60 min after PhrX in the same animals. Great care was taken to prevent any change in the nerve-electrode contact during the PhrX procedure (see methods). Blood samples taken during baseline (pre-PhrX) and at 60 min after PhrX (post) showed that Pa_O2 (132 ± 16 and 134 ± 20 Torr at pre and post, respectively), Pa_CO2 (32 ± 3 and 32 ± 3 Torr at pre and post, respectively), and arterial pH (7.40 ± 0.01 and 7.40 ± 0.02 at pre and post, respectively) were similar before and after PhrX. We were unable to collect blood samples in two animals; however, the phrenic nerve activity in these rats was not different in any way from the others, and, therefore, they were included in the overall group average. Blood pressure did not significantly change after PhrX (P = 0.58).

An example of the impact of PhrX on phrenic bursting is provided in Fig. 5. Acute PhrX abolished afferent bursting and led to an increase in ipsilateral efferent ∫Phr burst amplitude in all rats. In several (4 of 6) cases, the PhrX procedure caused an immediate increase in the amplitude of the inspiratory phrenic burst that was sustained for ∼60 min (Fig. 5). However, in the remaining (2 of 6) instances, there was a more gradual increase in phrenic burst amplitude after PhrX. On average, the mean ∫Phr efferent burst amplitude was significantly greater than the baseline value (i.e., pre-PhrX) only at 30 min post-PhrX (P < 0.05; Fig. 6B). The increase in ∫Phr burst amplitude at 1 min (P = 0.07) and 5 min (P = 0.08) post-PhrX did not reach statistical significance. The PhrX also led to an immediate increase in nonrhythmic tonic bursting in the efferent phrenic recording (Fig. 5). This response occurred in all PhrX experiments and, most likely, reflects high-frequency spike activity in axons and/or motoneurons secondary to axotomy. Indeed, the response shown after PhrX is quite similar to in vitro recordings of efferent bursting following acute axotomy (45; see discussion). The tonic phrenic bursting decreased gradually over time post-PhrX (Fig. 5, B and C). Phrenic inspiratory burst frequency was not influenced by PhrX (Fig. 6A). Analyses of the respiratory cycle before and after PhrX revealed no significant changes in Ti and Te (Table 2; P > 0.05).

Fig. 5.

Phrenic activity before, during, and after PhrX. Top: ∼1 h of ∫Phr data, including pre-PhrX baseline period, moment of PhrX (arrow), and post-PhrX period. Note immediate increase in the ∫Phr signal after PhrX. PhrX procedure always caused an increase in phrenic burst amplitude, and this response was abrupt (as shown in this example) in 60% of the experiments but was more gradual in the remaining 40%. Middle: expanded time scale traces showing several neural breaths for time points indicated by A, B, C, and D at top. Botom: further expanded time scale showing raw phrenic bursting at the points indicated by i–viii. Traces depict post-PhrX increase in phrenic activity during inspiratory burst (iii and v) and suggest an increase in tonic activity during the expiratory period (iv and vi).

Fig. 6.

Impact of PhrX on ∫Phr burst frequency and amplitude during pre-PhrX baseline period (pre) and at 1, 5, 30, and 60 min post-PhrX. No significant changes in burst frequency were noted (A). While most animals showed an abrupt increase in phrenic bursting after PhrX (see Fig. 5), a few showed a more gradual increase. Accordingly, increase in ∫Phr burst amplitude (%pre-PhrX amplitude, B) did not reach statistical significance until 30 min post-PhrX. *P < 0.05 vs. baseline (pre-PhrX).

These results do not contradict our data indicating that PhrX time-control rats do not show evidence of facilitation of inspiratory burst amplitude. During time-control experiments, there was a delay of ≥30 min after PhrX and before beginning the baseline period. In addition, the Pet_{CO2 apneic and recruitment thresholds were established after the PhrX procedure in the time-control rats.}

DISCUSSION

This is the first comprehensive analysis of the impact of acute PhrX on the expression of hypoxia-induced respiratory plasticity. Our primary finding was that PhrX is associated with a substantial increase in the magnitude of the acute hypoxic response and IH-induced LTF. Efferent phrenic motor output and the capacity for respiratory plasticity are thus influenced by axotomy of afferent and/or efferent axons in the phrenic nerve, and the integrity of the phrenic nerve should be taken into account when interpreting mechanisms of phrenic motor plasticity in anesthetized animals. The detailed mechanism(s) underpinning the effect of PhrX on LTF is unknown but may reflect an increase in phrenic motoneuron excitability. PhrX could change excitability by removing inhibitory afferent input, by injury-related mechanisms that lead to motoneuron depolarization, or by a change in intrinsic membrane properties.

Implications for the Study of Respiratory LTF

Many published studies of respiratory LTF have used in vivo (19, 43), in situ (67), or in vitro (10) preparations with recordings of efferent motor output from the distal end of the cut phrenic and/or hypoglossal (XII) nerves. These studies demonstrate predictable variability across laboratories, preparations, and species, as well as genetic variations between rat strains and substrains (20). However, a review of the literature indicates that LTF of efferent phrenic burst amplitude is considerably more robust in the aforementioned preparations than LTF assessed in spontaneously breathing animals (diaphragm EMG data) and humans (tidal volume data) with intact respiratory nerves (reviewed in Ref. 21). Indeed, LTF of respiratory output in spontaneously breathing animals is more often expressed as a persistent increase in breathing frequency, rather than inspiratory volume or EMG burst amplitude (6, 52). The difference in LTF expression between spontaneously breathing and ventilated preparations probably does not represent the impact of anesthesia on plasticity (30, 46). For example, Janssen et al. (30) could not induce LTF of diaphragm EMG burst amplitude in urethane-anesthetized and spontaneously breathing rats, despite their use of an anesthetic and IH regimen that evokes robust LTF in ventilated rats with PhrX (3, 22). Similarly, LTF of diaphragm EMG activity is not evident in anesthetized and spontaneously breathing cats following intermittent stimulation of the carotid sinus nerve (46). Interestingly, LTF of upper airway (e.g., genioglossus) muscle activity can be evoked in spontaneously breathing animals in the absence of diaphragm LTF (46, 68). A similar result was inferred from measurements of decreased pulmonary airflow resistance following IH in sleeping humans (1, 60). Therefore, the mechanisms that restrain phrenic/diaphragm LTF during spontaneous breathing (30) may not exert a parallel influence on hypoglossal or other upper airway respiratory motor outputs.

Janssen et al. (30) hypothesized that the (comparative) lack of phrenic/diaphragm LTF expression during spontaneous breathing reflects the relatively higher Pa_CO2 values in spontaneously breathing than ventilated animals. Since even small increases in Pa_CO2 will increase the overall output of phrenic motoneurons (32, 65), elevated Pa_CO2 may impair or reduce subsequent LTF expression via a “ceiling effect.” In other words, if phrenic motor output is relatively high during baseline (pre-IH) conditions, capacity for increased motoneuron recruitment or rate coding during the posthypoxic period may be reduced (14). Consistent with this idea, phrenic LTF is difficult to evoke in phrenic neurograms recorded contralateral to cervical spinal cord hemisection injury (14), a condition that results in robust compensatory increases in contralateral phrenic output (51). In contrast, Lee and colleagues (35) recently demonstrated that raising Pet_CO2 by ∼4 Torr above eupneic values is a prerequisite for induction of ventilatory LTF (including increased frequency and tidal volume) in spontaneously breathing humans. Thus, in some circumstances, Pa_CO2 elevations appear to be necessary for respiratory LTF, and factors other than Pa_CO2 levels may be responsible for differences in LTF between spontaneously breathing and ventilated preparations. Based on our data, we suggest that the condition of the phrenic nerve (i.e., cut vs. intact) contributes to the LTF differences. We therefore put forward the working hypothesis that PhrX creates preconditions that enhance the subsequent induction of phrenic LTF. To be clear, we are not suggesting that PhrX is a necessary precondition for LTF but, rather, that PhrX primes the phrenic motor system, possibly by increasing phrenic motoneuron excitability, and enables more robust increases in phrenic burst amplitude following IH or other stimuli (e.g., intermittent apnea) (44, 68).

Many studies of phrenic LTF in rats have used unilateral (vs. bilateral) PhrX (4, 5, 8, 40–42, 44, 67). In the current study, we used a bilateral PhrX to remove all phrenic afferent inputs to the spinal cord. It will be interesting in future work to examine the impact of unilateral PhrX on bilateral phrenic LTF (i.e., one phrenic nerve cut and the other intact). Based on the immediate impact of cutting the phrenic nerve [i.e., increased phasic and tonic bursting in that nerve (Fig. 5)], we speculate that such an experiment would result in enhanced LTF only in the cut phrenic nerve. In any case, our data underscore the importance of considering the effect(s) of PhrX when we interpret and compare LTF data from different experimental setups.

Below we briefly discuss two potential candidate mechanisms to explain the impact of PhrX on phrenic motor output and LTF.

Axotomy of Phrenic Motoneurons and Afferent Neurons: Potential Impact on LTF

The PhrX procedure used in the current and prior LTF studies will sever all the axons in the main trunk of the phrenic nerve. To interpret the effects of PhrX, the impact of axotomy on neuronal discharge and membrane properties should be considered. Long-term changes in neuronal membrane properties after chronic axotomy have been extensively studied (reviewed in Ref. 69). Over time frames ranging from days to weeks, axotomy results in changes in neuronal properties that favor increased discharge (e.g., increased resistance, decreased rheobase). Fewer studies have investigated the potential for acute changes in neuronal properties immediately following axotomy. Mandolesi et al. (45) showed that transecting the axons of cultured rat neurons initiates a rapid depolarization at the injury site followed by a burst of action potentials. This effect appears to be initiated at the site of axotomy, and then the depolarization travels back from the lesion site to the soma, where it triggers vigorous spiking activity and sustained depolarization lasting up to 10 min (45). Similarly, peripheral nerve crush in Aplysia californica results in a transient (∼60 s) bursting of action potentials in motoneurons that were silent prior to the injury (39).

The acute, axotomy-induced alterations in neuronal activity are associated with disruption of ionic regulation across the axon and/or soma membrane. For example, depolarization of the cell membrane after axotomy activates voltage-sensitive Ca²⁺ channels, triggering a rise of intracellular Ca²⁺ (47, 57, 59, 66, 70). There is also a gradual increase of intracellular Na⁺ after axotomy of in vitro mammalian neurons (45). Increased Ca²⁺ has been linked to activation of several molecular mechanisms that regulate various functions in the cell, including modulation of firing patterns and neuronal excitability (2). Furthermore, blocking Ca²⁺ influx after axotomy reduces excitability and prevents firing of in vitro mouse neurons (27).

Axotomy may also trigger a change in neuromodulatory inputs to motoneurons. Chronic axotomy of cervical spinal afferents (including phrenic afferents) via dorsal rhizotomy enhances serotonergic innervation of phrenic motoneurons and augments serotonin-dependent LTF of phrenic motor output (31). However, little is known about potential changes in serotonin receptor expression acutely following axotomy of mammalian neurons. Interestingly, serotonin exposure for 20 min causes molecular changes (activation of mitogen-activated protein kinase) that are similar to those observed after axons are severed (39). These data present the possibility that axotomy-evoked electrical discharge or molecular changes might promote serotonin receptor expression in motoneurons or reorganization of premotor serotonergic input (39). Interruption of axonal continuity also disrupts normal retrograde transport of trophic signals from the axon terminal, which can affect the synthesis of cAMP and gene expression (9, 37, 38). The potential for rapid changes in neurotrophic factor expression after axotomy is of particular interest in regard to phrenic LTF. Baker-Herman et al. (5) demonstrated that spinal brain-derived neurotrophic factor is both necessary and sufficient for phrenic LTF in PhrX rats. Accordingly, if phrenic axotomy influences brain-derived neurotrophic factor expression in or around phrenic motoneurons, this could have a profound effect on the subsequent induction of LTF.

The PhrX procedure will also abruptly eliminate inputs associated with activation of phrenic afferent fibers. It is not always appreciated that a large portion (i.e., 40–45%) of the axons in the phrenic nerve are afferent in origin (33, 34). These afferent fibers carry information from the proprioceptors (muscle spindles and tendon organs), rapidly adapting mechanoreceptors (Pacinian corpuscles), and nociceptors in the diaphragm, as well as free nerve endings in the pericardium and pleural surface of the diaphragm (28, 56). Many studies have reported that electrical stimulation of phrenic afferents has an inhibitory effect on phrenic motoneuron activity (15, 23, 29, 55, 63). Jammes et al. (29) showed that stimulation of large-diameter and thin afferent fibers in the phrenic nerve of anesthetized cats caused a contralateral reduction of phrenic motoneuron impulse frequency and duration of phrenic activity, respectively. Similarly, an inhibitory effect on ipsilateral phrenic activity in response to phrenic afferent stimulation has also been demonstrated (55). Additionally, Cheeseman et al. (11) showed that a sudden change in diaphragm length causes a reflex reduction of integrated diaphragm EMG amplitude in anesthetized cats, and this effect was not observed after interruption of afferent input via cervical dorsal rhizotomy. On the other hand, activation of phrenic afferents may stimulate breathing in some cases. Speck and Revellete (64) showed that about one-fourth of dorsal respiratory group respiratory-modulated neurons are excited by phrenic afferents. Based on conduction velocity measurements, they attributed this effect to activation of predominantly small, type III myelinated fibers. However, despite this excitation, the overall effect of phrenic afferent stimulation was inhibition of phrenic motoneuron activity. In any case, it is clear that sensory feedback from the diaphragm can modulate the respiratory drive. However, the impact of phrenic afferents on phrenic motor output in the anesthetized, ventilated, paralyzed rat preparation is unclear. Our data demonstrate that acute PhrX eliminates an afferent signal originating distal to the recording site (Fig. 6). This phrenic afferent bursting occurred in phase with lung deflation, as indicated by the tracheal pressure recordings (Fig. 1). Therefore, it is possible that cutting the phrenic nerve removes phrenic afferent-mediated modulation of phrenic output at the level of the spinal cord (29) or via ascending projections to medullary respiratory centers (62).

Finally, our data also showed that the increase in phrenic burst frequency (bursts/min) during the initial hypoxic episode was significantly greater in PhrX than PhrI rats. Accordingly, the PhrX procedure may have removed inhibitory inputs to the brain stem respiratory rhythm/pattern generator, which in turn could potentiate the acute hypoxic response. However, the effects of PhrX were only transient in this case, as differences in the hypoxic frequency response between PhrX and PhrI rats were not observed after the initial hypoxic episode.

Conclusion

Our data indicate that PhrX affects phrenic output during baseline and hypoxic conditions and the magnitude of LTF in response to IH in anesthetized rats. Therefore, along with other parameters, including the integrity of the vagus nerves (24), anesthesia, and mechanical ventilation, the effect of PhrX should be taken into consideration when LTF data from PhrX animals are interpreted. The impact of PhrX may be especially important in studies investigating the molecular mechanisms of phrenic LTF. An investigation of the specific mechanism(s) through which PhrX affects phrenic motoneuron excitability and LTF should be the subject of future studies.

GRANTS

This work was funded by National Institute of Child Health and Human Development Grant 1R01 HD-052682-01A1 (D. D. Fuller).

DISCLOSURES

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