Inactivity-induced phrenic and hypoglossal motor facilitation are differentially expressed following intermittent vs. sustained neural apnea

Abstract

Reduced respiratory neural activity elicits a rebound increase in phrenic and hypoglossal motor output known as inactivity-induced phrenic and hypoglossal motor facilitation (iPMF and iHMF, respectively). We hypothesized that, similar to other forms of respiratory plasticity, iPMF and iHMF are pattern sensitive. Central respiratory neural activity was reversibly reduced in ventilated rats by hyperventilating below the CO₂ apneic threshold to create brief intermittent neural apneas (5, ∼1.5 min each, separated by 5 min), a single brief massed neural apnea (7.5 min), or a single prolonged neural apnea (30 min). Upon restoration of respiratory neural activity, long-lasting (>60 min) iPMF was apparent following brief intermittent and prolonged, but not brief massed, neural apnea. Further, brief intermittent and prolonged neural apnea elicited an increase in the maximum phrenic response to high CO₂, suggesting that iPMF is associated with an increase in phrenic dynamic range. By contrast, only prolonged neural apnea elicited iHMF, which was transient in duration (<15 min). Intermittent, massed, and prolonged neural apnea all elicited a modest transient facilitation of respiratory frequency. These results indicate that iPMF, but not iHMF, is pattern sensitive, and that the response to respiratory neural inactivity is motor pool specific.

the neural network controlling breathing expresses remarkable plasticity in response to changing environmental or physiological conditions (7, 13, 19, 36, 39, 44, 53). We recently demonstrated that reduced respiratory neural activity (without hypoxia or hypercanpia) elicits rebound plasticity in respiratory activity (7, 31, 52), a feature that is not necessarily surprising since this critical, homeostatic control system must be continuously active to maintain life. Briefly, a prolonged (30 min) central neural apnea in ventilated rats elicits a rebound increase in inspiratory motor output, known as inactivity-induced phrenic (iPMF) (31) and hypoglossal (iHMF) motor facilitation (7). Since iPMF is induced by multiple forms of central apnea, including hyperventilation, increased inhibitory sensory feedback, and anesthesia-induced respiratory depression (31), we hypothesized that iPMF (and by extension, iHMF) arises from a common factor: reduced respiratory neural activity (7). Although initiated by similar stimuli, inactivity-induced facilitation is differentially expressed in the phrenic vs. hypoglossal motor pools after prolonged neural apnea. Specifically, iPMF is long lasting (>60 min) (31), whereas iHMF is more modest and transient (∼15 min) (7).

Many forms of respiratory plasticity are sensitive to stimulus pattern (2, 3, 29, 53). For example, intermittent hypoxia (IH) elicits a serotonin-dependent increase in phrenic and hypoglossal inspiratory motor output known as phrenic and hypoglossal long-term facilitation (pLTF and hLTF, respectively) (1, 23). By contrast, a similar cumulative duration of sustained hypoxia does not elicit pLTF (3, 16) or hLTF (7). Similarly, intermittent, but not sustained, disruptions in vagal feedback elicit long-lasting increases in genioglossus activity (53). Mechanisms giving rise to pattern sensitivity in respiratory plasticity are unknown, but may involve differential activation patterns of protein kinases/phosphatases during intermittent vs. sustained stimuli (61).

Here, we tested the hypothesis that, similar to other forms of respiratory plasticity, iPMF and iHMF are pattern sensitive. Although we know that prolonged neural apnea (∼30 min) elicits iPMF and iHMF (7, 31, 52), we now report that intermittent exposure to five brief (1.5 min) neural apneas (separated by ∼5 min) elicits iPMF, whereas sustained exposure to a single “massed” apnea of a similar cumulative duration (7.5 min) does not. By contrast, iHMF is not elicited by brief neural apnea when presented in either intermittent or massed patterns. Thus, with short neural apneas, iPMF exhibits pattern sensitivity, whereas iHMF does not. Further, we demonstrate that patterns of neural apnea that induce long-lasting iPMF also increase phrenic responses to severe hypercapnia. Last, brief massed, brief intermittent, and prolonged neural apnea all elicit a modest facilitation of respiratory frequency, whose magnitude appears to be sensitive to apnea duration. Collectively, these data demonstrate that the response to reduced respiratory neural activity is motor pool specific, and expand our understanding of inactivity-induced respiratory motor plasticity.

METHODS

Animals.

Experiments were performed on 2.5- to 5-mo-old male Sprague-Dawley rats obtained from Harlan Laboratories (n = 33; colony 217; Indianapolis, IN). All experimental protocols were approved by the Animal Care and Use Committee at the University of Wisconsin, Madison.

Experimental preparation.

Isoflurane anesthesia was induced in a closed container. Rats were then transferred to a nose cone with 2.5–3.5% isoflurane (50% O₂-N₂ balance). A tail vein catheter (Surflo iv catheter and injection plug) was placed for delivery of intravenous fluids (lactated Ringer solution; ∼20% sodium bicarbonate, as necessary) to maintain blood pressure and blood pH homeostasis throughout surgery and experimental protocols. Body temperature was monitored with a rectal thermometer (Physitemp, model 700 1H) and maintained near 37.0°C with a custom heated surgery table. The trachea was isolated and cannulated and mechanical ventilation was begun (∼70 breaths/min, tidal volume = 2.5–3.5 ml; Harvard Apparatus, model 683; 50% O₂-N₂ balance). End-tidal CO₂ (ETCO₂) was measured continuously from the expired limb of the ventilator circuit as an index of arterial pCO₂ with a flow-through capnograph (Capnogard, Respironics). In preliminary studies, we found that, under our anesthesia protocol, spontaneously breathing rats typically have an ETCO₂ of ∼45 mmHg; thus ETCO₂ was maintained near 45 mmHg throughout the surgery by adding CO₂ to the inspired gas mixture to prevent suppression of respiratory neural activity during mechanical ventilation. Tracheal pressure was monitored, and inspired CO₂ and/or ventilator frequency was adjusted to ensure the animal continued to generate inspiratory efforts. The vagus nerve was cut bilaterally at the cervical level to prevent entrainment of respiratory neural activity with the ventilator. The right femoral artery was isolated and catheterized to measure arterial blood pressure and draw blood samples (∼0.3 ml) for pH and blood gas analysis (ABL800; Radiometer, Copenhagen, Denmark). Urethane anesthesia was then begun (1 ml/100 g of 0.175 g/ml infused at 6 ml/h iv), and isoflurane was slowly withdrawn. Adequate depth of anesthesia was confirmed by testing pressor responses to toe-pinch. The left hypoglossal (XII) and phrenic nerves were dissected via a dorsal approach, cut distally, de-sheathed, placed on bipolar silver electrodes, and submerged in mineral oil. After conversion to urethane anesthesia, pancuronium bromide was infused (2.5 mg/kg iv) to induce neuromuscular paralysis.

Experimental protocols.

One hour after conversion to urethane anesthesia, phrenic and hypoglossal nerve burst amplitude and frequency were monitored under isocapnic conditions until activity remained stable for at least 15–20 min (baseline). At the end of baseline, an arterial blood sample was drawn to determine baseline partial pressure of arterial CO₂ (Pa_CO2), partial pressure of arterial O₂ (Pa_O2), pH, and standard base excess (SBE; temperature corrected) measurements. Rats were then exposed to one of three patterns of reduced respiratory neural activity: five brief intermittent neural apneas (∼1.5 min each, separated by 5 min of baseline conditions; n = 9), a single brief massed neural apnea of a similar cumulative duration (7.5 min; n = 9), or a single prolonged neural apnea (30 min; n = 8). A fourth group underwent similar surgical procedures but received no neural apnea to control for potential time-dependent changes in neural activity over the course of the protocol (time controls; n = 6). Central neural apnea was induced by lowering inspired CO₂ until all rhythmic phrenic and hypoglossal activity ceased (CO₂ apneic threshold); occasionally, it was necessary to also increase ventilator frequency to lower Pa_CO2 enough to fully induce neural apnea. In rats receiving brief intermittent neural apnea, neural apnea was quickly reversed shortly after induction by increasing inspired CO₂ until respiratory neural activity resumed (CO₂ recruitment threshold) and baseline Pa_CO2 was restored. This process was repeated five times. In rats receiving brief massed or prolonged neural apnea, central neural apnea was induced in a similar manner but was then maintained by holding ETCO₂ ∼2 mmHg below the CO₂ apneic threshold for 7.5 or 30 min (brief massed or prolonged neural apnea, respectively) and then reversed as described above. Importantly, in all groups, Pa_O2 was maintained with the ventilator during neural apnea.

Arterial blood samples were drawn before, during (prolonged neural apnea group only) and 5, 15, 30, and 60 min following central neural apnea to confirm Pa_CO2 was within 1.5 mmHg of baseline and to ensure maintenance of Pa_O2, pH, and SBE throughout the protocol. At the end of each protocol, the response to a hypercapnic challenge (90 mmHg < ETCO₂ < 100 mmHg) was tested to determine whether the ability to respond to a respiratory challenge was preserved in the face of plasticity (i.e., no “ceiling” effect). To be included in the study, Pa_CO2 had to be maintained within 1.5 mmHg and MAP maintained within 50 mmHg from baseline at all time points postneural apnea; 33 of 35 protocols met these criteria.

Data analysis.

Phrenic and hypoglossal burst activity was amplified (×10,000), band-pass filtered (0.3–10 kHz; AM Systems), integrated (time constant 50 ms), and rectified. The resulting signal was digitized and analyzed with PowerLab (AD Instruments; Lab Chart 7.0 software). Sixty-breath bins were taken immediately prior to blood gas sampling at baseline and 5, 15, 30, and 60 min postneural apnea and analyzed for peak amplitude and burst frequency (neural correlates of tidal volume and breathing frequency, respectively). Phrenic and hypoglossal amplitude was also quantified during the hypercapnic challenge at the end of each protocol. Phrenic and hypoglossal nerve burst amplitude was expressed as a percent change from baseline (% baseline) and burst frequency was expressed as an absolute change from baseline (Δ baseline). Statistical differences between groups and individual time points were determined using two-way repeated measures ANOVA and Bonferroni post hoc tests, respectively (Prism 5, Graph Pad software). Group differences in maximum phrenic and hypoglossal amplitude elicited during hypercapnic challenge were determined using a one-way ANOVA and Fisher's least significant difference post hoc test (Sigma Plot 12 software). Linear regression analysis was used to determine the relationship between iPMF magnitude and the magnitude of the phrenic response to a hypercapnic challenge (Prism 5, Graph Pad software). A significance level of 0.05 was set for all comparisons. Data are shown as means ± SE.

RESULTS

Respiratory neural activity was reversibly reduced by lowering Pa_CO2 below the CO₂ apneic threshold for breathing (neural apnea). Reduced respiratory neural activity was presented in three different patterns: five brief neural apneas (∼1.5 min each) separated by 5 min of resumed respiratory neural activity (brief intermittent neural apnea), a single brief massed neural apnea of similar cumulative duration (7.5 min), or a prolonged neural apnea (30 min) which approximated the duration of the stimulus period (neural apneas plus intervening rest periods). The total cumulative duration of brief intermittent neural apnea (8:43 ± 0:48, min:s) was not significantly different from the duration of brief massed neural apnea (7:30 ± 0:39, min:s; P > 0.05), and the total stimulus period of brief intermittent neural apnea (28:56 ± 0:59, min:s) was not significantly different from the duration of prolonged neural apnea (30:05 ± 0:19, min:s; P > 0.05). There were no differences in ETCO₂ during neural apnea between groups (in mmHg: brief intermittent 27.8 ± 1.6, brief massed 26.2 ± 1.5, prolonged 28.4 ± 1.6; P > 0.05). Following intermittent, massed, or prolonged neural apnea, Pa_CO2 was restored to baseline levels and phrenic/hypoglossal nerve activity was monitored for 60 min. Table 1 shows Pa_O2, Pa_CO2, pH, and mean arterial pressure (MAP) during baseline and 60 min postneural apnea. There were no time-dependent changes in Pa_CO2 in any group (P > 0.05). There were significant time-dependent changes in Pa_O2 in time control rats, pH in rats receiving prolonged neural apnea, and MAP in rats receiving brief intermittent neural apnea (P < 0.05). These changes were not considered to be physiologically relevant in the context of our study since: 1) Pa_O2 was maintained well above (>200 mmHg) levels that stimulate breathing, 2) the drop in pH was small (from 7.34 to 7.31) and previous reports have found iPMF without a significant change in pH (31, 52), and 3) decreases in MAP were small (∼20 mmHg) and changes in MAP of this magnitude have minimal long-lasting effects on respiratory activity in rats (58). Indeed, changes in Pa_O2, pH, and MAP were not significantly related to the magnitude of iPMF 60 min following neural apnea (R² = 0.308, R² = 0.001, and R² = 0.354, respectively; all P > 0.05). Thus these effects are not likely to have influenced our results.

Table 1.

Pa_O2, Pa_CO2, pH, and mean arterial pressure (MAP) during baseline and 60 min postneural apnea

	Prolonged		Intermittent		Massed		Time Control
	Baseline	60 min	Baseline	60 min	Baseline	60 min	Baseline	60 min
PaCO2, mmHg	47.6 ± 0.7	48.1 ± 0.7	45.0 ± 0.9	45.0 ± 0.9	45.4 ± 0.9	45.5 ± 0.9	46.8 ± 0.9	46.6 ± 1.2
PaO2, mmHg	289.9 ± 7.9	292.4 ± 8.0	296.9 ± 4.3	287.1 ± 5.0	291.9 ± 5.0	284.4 ± 5.8	296.3 ± 13.1	282.3 ± 16.3*
pH	7.344 ± 0.01	7.309 ± 0.01*	7.375 ± 0.01	7.362 ± 0.01	7.347 ± 0.01	7.332 ± 0.01	7.338 ± 0.00	7.323 ± 0.01
MAP, mmHg	149.3 ± 2.9	150.6 ± 4.5	144.0 ± 4.8	122.6 ± 5.1*	148.3 ± 6.1	142.7 ± 4.7	151.8 ± 3.5	146.7 ± 3.1

Values are means ± SE. Pa_O2, partial pressure of arterial O₂; Pa_CO2, partial pressure of arterial CO₂; MAP, mean arterial pressure.

^*Significantly different than baseline (P < 0.05).

iPMF is pattern sensitive.

Representative compressed phrenic neurograms at baseline, during, and 60 min following either prolonged, brief intermittent, or brief massed neural apnea are shown in Fig. 1A. The average change in phrenic burst amplitude for 60 min following prolonged, brief intermittent, or brief massed neural apnea, or an equivalent duration in time control rats not receiving neural apnea is shown in Fig. 1B. There were no time-dependent changes in phrenic burst amplitude from baseline in time controls (5 min: 6 ± 2, 15 min: 10 ± 2, 30 min: 13 ± 4, 60 min: 16 ± 6% baseline, P > 0.05; Fig. 1B), suggesting our preparation was stable during the recording period. As previously reported (31, 52), a 30-min prolonged neural apnea elicited a long-lasting (>60 min) increase in phrenic burst amplitude at all time points following resumption of respiratory activity (5 min: 63 ± 11, 15 min: 74 ± 13, 30 min: 72 ± 12, 60 min: 69 ± 11% baseline), which was significantly greater than time controls (P < 0.01; Fig. 1, A and B). Similarly, five brief neural apneas (∼1.5 min) presented intermittently over a 30-min period resulted in increased phrenic burst amplitude at all time points postneural apnea (5 min: 51 ± 13, 15 min: 64 ± 13, 30 min: 67 ± 14, 60 min: 83 ± 19% baseline), which was significantly greater than time controls (P < 0.01; Fig. 1, A and B). There were no significant differences in phrenic burst amplitude postneural apnea between rats receiving prolonged or brief intermittent neural apnea (P > 0.05). By contrast, a brief (7.5 min) massed neural apnea did not result in increased phrenic burst amplitude at any time point postneural apnea (5 min: 16 ± 4, 15 min: 18 ± 3, 30 min: 20 ± 3, 60 min: 12 ± 3% baseline), compared with time controls (P > 0.05). Further, phrenic burst amplitude following a brief neural apnea was significantly less than phrenic burst amplitude following brief intermittent or prolonged neural apnea (P < 0.05). Collectively, these data indicate that, although iPMF is observed following a prolonged period of reduced respiratory activity, iPMF is more efficiently induced by repeated reductions in respiratory neural activity than by a similar cumulative duration of sustained reduced activity.

Fig. 1.

Inactivity-induced phrenic motor facilitation (iPMF) is pattern sensitive. A: representative compressed phrenic neurograms depicting integrated nerve burst amplitude before, during, and 60 min following either prolonged (top), brief intermittent (middle), or brief massed (bottom) neural apnea. B: mean integrated phrenic amplitude 5, 15, 30, and 60 min following brief intermittent, prolonged, brief massed, or no neural apnea (time control). Changes are expressed as a percentage of baseline amplitude. Brief intermittent and prolonged neural apnea elicit an increase in phrenic amplitude (i.e., iPMF) with a similar magnitude and duration, whereas brief massed neural apnea has no effect. #Significantly different from time control and brief massed neural apnea.

iHMF is not pattern sensitive.

Representative compressed hypoglossal neurograms at baseline, during, and 60 min following either a prolonged, brief intermittent, or brief massed neural apnea are shown in Fig. 2A. The average change in hypoglossal burst amplitude for 60 min following prolonged, brief intermittent, or brief massed neural apnea, or an equivalent duration in time control rats not receiving neural apnea is shown in Fig. 2B. There were no time-dependent changes in hypoglossal burst amplitude from baseline in time controls (5 min: −5 ± 6, 15 min: −6 ± 7, 30 min: −5 ± 7, 60 min: −4 ± 4% baseline; P > 0.05; Fig. 2B). Consistent with previous reports (7), a 30-min prolonged neural apnea elicited a transient increase in hypoglossal burst amplitude, which was only significantly different from time controls for ∼15 min following the resumption of respiratory neural activity (5 min: 27 ± 12, 15 min: 36 ± 12% baseline; P < 0.05); by 30 min postneural apnea, hypoglossal burst amplitude had returned to baseline levels (30 min: 13 ± 9, 60 min: 0 ± 8% baseline; P > 0.05; Fig. 2, A and B). No change in hypoglossal burst amplitude was observed following brief intermittent neural apnea (5 min: 3 ± 6, 15 min: 7 ± 4, 30 min: −1 ± 4, 60 min: −3 ± 3% baseline) or brief massed neural apnea of an equivalent cumulative duration (5 min: 0 ± 6, 15 min: 12 ± 5, 30 min: 15 ± 5, 60 min: 6 ± 5% baseline), compared with time controls (P > 0.05). These data suggest that prolonged, but not brief (intermittent or massed), neural apnea elicits iHMF. Thus, unlike iPMF, iHMF may not be sensitive to the pattern of reduced respiratory neural activity.

Fig. 2.

Inactivity-induced hypoglossal motor facilitation (iHMF) is not pattern sensitive. A: representative compressed hypoglossal (XII) neurograms depicting integrated nerve burst amplitude before, during, and 60 min following either prolonged (top), brief intermittent (middle), or brief massed (bottom) neural apnea. B: mean integrated hypoglossal amplitude 5, 15, 30, and 60 min following brief intermittent, prolonged, brief massed, or no neural apnea (time control). Changes are expressed as a percentage of baseline amplitude. Hypoglossal amplitude is transiently increased following prolonged neural apnea, but is unchanged following brief intermittent or brief massed neural apnea. #Significantly different from time control and brief massed neural apnea; *significantly different from time control; ϕsignificantly different from prolonged neural apnea.

iPMF is associated with increased maximum phrenic burst amplitude.

To determine if expression of inactivity-induced plasticity occluded further increases in inspiratory burst amplitude (i.e., a ceiling effect), rats were exposed to severe hypercapnia at the end of the protocol to elicit a maximum phrenic and hypoglossal burst amplitude response. Average phrenic and hypoglossal burst amplitude during a severe hypercapnic challenge following prolonged, brief intermittent, or brief massed neural apnea, or an equivalent duration in time controls is shown in Fig. 3A. There were no significant differences in hypoglossal burst amplitude in severe hypercapnia relative to time controls following brief intermittent, brief massed, or prolonged neural apnea (179 ± 34, 166 ± 23, 148 ± 16, and 155 ± 39% baseline, respectively; P > 0.05), treatments that do not result in long-lasting facilitation. Similarly, no significant difference was observed in phrenic burst amplitude during severe hypercapnia in rats receiving brief massed neural apnea, relative to time controls (111 ± 18 and 103 ± 10% baseline, respectively; P > 0.05). By contrast, rats receiving either prolonged or brief intermittent neural apnea had significantly increased phrenic burst amplitude during severe hypercapnia (226 ± 42 and 259 ± 59% baseline, respectively), relative to time controls or rats receiving brief massed neural apnea (both P < 0.05), suggesting that the expression of iPMF is associated with an increase in the maximum phrenic response to a respiratory challenge. Indeed, a significant positive relationship between the magnitude of iPMF 60 min following neural apnea and the maximum phrenic amplitude elicited during hypercapnia was observed following prolonged or brief intermittent neural apnea (% baseline; R² = 0.634, P < 0.001; Fig. 3B). Collectively, these data suggest that the expression of iPMF is associated with a proportional increase in the phrenic dynamic range.

Fig. 3.

Reduced respiratory neural activity increases maximum phrenic burst amplitude during hypercapnia. A: average maximum integrated phrenic and hypoglossal (XII) amplitude elicited during a hypercapnic challenge ∼60 min following brief intermittent, brief massed, prolonged, or no neural apnea (time control). Changes are expressed as a percentage of baseline amplitude. Rats exposed to brief intermittent or prolonged neural apnea (inducing iPMF) have a significantly greater maximum phrenic amplitude during hypercapnic challenge than rats receiving brief massed neural apnea or time controls. By contrast, no differences in maximum hypoglossal amplitude were observed between any treatment groups. Shaded bars indicate patterns of neural apnea that result in significant iPMF. #Significantly different from time control and brief neural apnea. B: linear regression of maximum phrenic amplitude elicited during a hypercapnic challenge vs. the magnitude of iPMF expressed 60 min following brief intermittent and prolonged neural apnea.

Frequency facilitation is sensitive to the duration, but not the pattern, of neural apnea.

The average change in respiratory burst frequency at 5, 15, 30, and 60 min following either a prolonged, brief intermittent, or brief massed neural apnea, or an equivalent duration in time control rats is shown in Fig. 4. No time-dependent changes in burst frequency were observed in time control rats not receiving neural apnea (5 min: −2 ± 1, 15 min: −2 ± 1, 30 min: 0 ± 1, 60 min: 1 ± 2 change from baseline; P > 0.05). Following prolonged neural apnea, burst frequency was significantly increased at 5 and 15 min (11 ± 2 and 6 ± 1 change from baseline, respectively) relative to time controls (P < 0.05); however, by 30 min following restoration of respiratory neural activity, burst frequency was no longer significantly different from time controls (30 min: 3 ± 1, 60 min: 4 ± 1 change from baseline; P > 0.05). Modest transient increases in burst frequency were observed following brief intermittent (5 min: 4 ± 1, 15 min: 3 ± 1 change from baseline) and brief massed (5 min: 5 ± 1, 15 min: 5 ± 1 change from baseline) neural apnea; both responses were significantly greater than equivalent time points in time controls (P < 0.05), but were significantly less than those observed following prolonged neural apnea (P < 0.05). By 30 min following brief neural apnea, burst frequency had returned to levels not significantly different from time controls (brief intermittent: 5 ± 1 and 4 ± 1; brief massed: 3 ± 1 and 3 ± 1 change from baseline, 30 and 60 min, respectively; P > 0.05). Collectively, these data suggest that the magnitude of inactivity-induced facilitation of respiratory frequency is sensitive to the cumulative duration of neural apnea, but not its pattern.

Fig. 4.

Mean respiratory nerve burst frequency 5, 15, 30, and 60 min following intermittent, prolonged, brief, or no neural apnea (time control). Changes are expressed as an absolute change from baseline (Δ baseline). All neural apnea transiently increases nerve burst frequency; however, a continuous 30-min neural apnea elicits a greater increase in nerve burst frequency than 5 min of intermittent or massed neural apnea. #Significantly different from time control and brief neural apnea; *significantly different from time control; ϕsignificantly different from prolonged neural apnea.

DISCUSSION

Here we demonstrate that a novel form of respiratory plasticity induced by reduced respiratory neural activity (i.e., iPMF) is pattern sensitive. Although a prolonged central neural apnea is sufficient to elicit iPMF (7, 31, 52), iPMF is more efficiently induced by repeated (intermittent) neural apneas vs. a sustained neural apnea of similar cumulative duration. Further, iPMF is associated with a proportional increase in the phrenic burst amplitude response to a hypercapnic challenge, suggesting that the ability to respond to a respiratory challenge is preserved in the presence of inactivity-induced plasticity. The effects of neural apnea appear to be motor pool specific since facilitation in hypoglossal activity is either absent or transient following brief intermittent or prolonged neural apnea, respectively. Last, we show that changes in respiratory frequency following reduced respiratory neural activity are transient and sensitive to the duration of neural apnea, but not its pattern. These results extend our understanding of plasticity induced by reduced activity within the respiratory neural network and may have implications for conditions that result in episodic reductions in central respiratory drive, such as sleep at altitude, chronic heart failure, and central sleep apnea.

Methodological considerations.

To assess the role of reduced respiratory neural activity in eliciting phrenic and hypoglossal inspiratory motor plasticity, we used a well-established anesthetized rat preparation commonly used to study respiratory plasticity (20, 37). This preparation offers distinct advantages that enable investigations of the neural control of breathing independent from the process of blood gas homeostasis. For example, mechanical ventilation enables strict control over arterial blood gases and avoids attendant hypercapnia and hypoxia—stimuli that induce plasticity in and of themselves (1, 2)—that would normally accompany reduced respiratory neural activity. Further, afferent feedback is minimized by vagotomy, paralysis, and cutting the recorded phrenic and hypoglossal nerves. Collectively, these techniques enable isolation of the effects of reducing brainstem respiratory neural output from many other factors that modify it. However, anesthesia (12) and afferent feedback from the vagus (22) and phrenic (49) nerves are capable of modulating the expression of plasticity. Nevertheless, our findings demonstrate that neural networks underlying breathing sense and respond rapidly to reduced respiratory neural activity, independent of the effects of hypercapnia and hypoxia. The conditions under which these mechanisms manifest as a phenotypic physiological response are of clear clinical and biological interest.

Expression of iPMF is sensitive to the pattern of inactivity.

Plasticity within neural networks controlling motor behaviors is often more effectively induced by intermittent than sustained stimuli (3, 9, 10, 35, 53). For instance, LTF of sensorimotor synapses in Aplysia is induced by both intermittent (five applications over 1.5 h) and prolonged (1.5 h) application of serotonin (18, 27, 38), whereas a relatively brief (∼25 min) application of serotonin is less likely to induce LTF (34).

In many studies describing pattern-sensitive plasticity, the relevant stimulus is often directly or indirectly associated with an increase in neuronal and/or synaptic activity. However, reduced neuronal activity also elicits plasticity, typically by altering the properties of a neuron or network of neurons in a non-synapse specific manner (15, 57). This “synaptic scaling” is thought to have a compensatory or homeostatic function since prolonged reductions in neuronal activity alter cellular properties to increase excitation and/or decrease inhibition (25, 26, 50, 55, 56, 60), presumably to restore a target level of activity and preserve “normal” function. For example, in cultured hippocampal neurons subjected to a prolonged (48 h) activity blockade via application of tetrodotoxin (blocking axon conduction), postsynaptic alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor expression is increased through a tumor necrosis factor alpha-dependent mechanism (51). Although it is clear that long-lasting plasticity can be induced by a single sustained period of reduced activity within a neural network, few studies have tested the effect of repeated or intermittent inactivity. Our data suggest that, in some motor pools (but not in others), intermittent patterns of reduced respiratory neural activity may be more efficient than prolonged reductions to induce plasticity. The detailed, and potentially unique, mechanisms that impart pattern sensitivity on inactivity-induced plasticity of phrenic motor output remain unknown.

Expression of iPMF increases phrenic dynamic range.

Respiratory motor output operates within a limited dynamic range beginning at zero (neural apnea) and spanning to a maximum value that can be elicited during a potent respiratory challenge, such as high CO₂. Under normal conditions, only a fraction of this range is utilized (32). Our observation that iPMF proportionally increases the maximum phrenic response to high CO₂ is consistent with other reports (53), and suggests that the ability of the respiratory control system to respond to a respiratory challenge is preserved in the face of at least certain forms of plasticity. Thus we hypothesize that the iPMF phenotype is the result of a compensatory shift in phrenic excitability to restore motor output when respiratory neural activity is reduced. Because the dynamic range of all inspiratory motor output was not increased (i.e., hypoglossal maximum output remains constant), it is likely that inactivity-induced changes in phrenic dynamic range occur downstream of sites of chemosensation, central respiratory rhythm generation, and (brainstem) premotor processing. Together, these data are consistent with the idea that iPMF is not due to an increase in brainstem respiratory drive, but is a process occurring locally within individual respiratory motor pools.

Effects of neural apnea on respiratory motor output are motor pool specific.

Phrenic and hypoglossal motor pools provide excitatory drive to their target muscles (diaphragm and tongue, respectively) during the inspiratory phase of respiration; however, the physiological functions of these motor pools differ greatly. Regular rhythmic activity of phrenic motor neurons is predominantly involved in breathing (negative pressure ventilation) and is essential to maintain life, while hypoglossal motor neurons are active during a variety of behaviors including chewing, swallowing, and vocalization, as well as maintaining upper airway patency during breathing (28). Further, reduced hypoglossal activity does not generally pose an immediate threat to life (42, 62).

Despite differences in functionality, phrenic and hypoglossal motor pools are both controlled by the same brainstem regions that initiate, coordinate, and modulate respiratory motor neuron activity (19). Because of this shared input, different respiratory motor pools often respond similarly when the system is challenged. However, there may also be benefits in local regulation specific to each motor pool, particularly with regard to long-lasting changes in system performance (i.e., plasticity) (3, 7, 53), a characteristic that may allow motor pools to respond in a stimulus-specific manner. For example, moderate IH elicits a serotonin-dependent increase in inspiratory motor output known as LTF (1, 6). Similar to inactivity-induced facilitation, IH-induced LTF is differentially expressed in phrenic and hypoglossal motor output (4, 7, 20, 21). One important question is: How does differential expression of plasticity arise if both phrenic and hypoglossal motor pools experience similar inducing stimuli? Evidence suggests that local mechanisms operating in or near the motor pool may ultimately underlie subsequent responses, even though central mechanisms may be an initial trigger. With regard to LTF following moderate IH, it is suggested that the stimulus (hypoxia) is “sensed” by peripheral chemoreceptor stimulation which initiates a common (centralized) trigger: episodic activation of medullary raphe neurons and subsequent release of serotonin in the phrenic and hypoglossal motor pools (in addition to other disparate sites), resulting in LTF. However, local mechanisms (e.g., serotonin receptor expression and terminal density, presence of other neuromodulator systems) then appear to modify IH-induced LTF in a motor pool specific manner (7, 24, 41).

iPMF and iHMF are also initiated by a common trigger: central neural apnea. However, we hypothesize that local mechanisms within individual respiratory motor pools sense and respond to the initiating stimulus: reduced respiratory related synaptic inputs. Mechanisms that sense reduced respiratory activity and initiate iPMF or iHMF are unknown; however, downstream mechanisms that give rise to enhanced inspiratory burst amplitude are beginning to be elucidated (52). At least for iPMF, these mechanisms ultimately activate atypical protein kinase C (aPKC) isoforms PKCζ or PKCι within or near the phrenic motor pool and induce the formation of a signaling complex between PKCζ/ι and the scaffolding molecule p62/ZIP (52). aPKCs represent one of three subfamilies of PKC (classical, novel, and atypical), and have been implicated in multiple forms of plasticity (for review, see Refs. 47, 48, 54). Mechanisms of transient iHMF are unknown, and we do not understand why inactivity-induced facilitation differs so greatly in phrenic and hypoglossal motor pools. Inactivity-induced plasticity may operate in time domains tailored to the specific characteristics of a motor pool, such as the regularity and extent of normal activity, the diversity of function, or the physiological consequences of reduced activity. Consistent with this idea, the phrenic motor pool responds to reduced respiratory neural activity fairly rapidly, whereas the hypoglossal motor pool may require a longer duration of inactivity to consolidate long-lasting plasticity. In this sense, rapid and robust iPMF may preserve a vital, time-sensitive (i.e., adjustments must be made within minutes) physiological function, whereas a similar response within the hypoglossal motor pool may hinder normal function when required for other nonrespiratory behaviors. Alternatively, it is possible that the transient nature of iHMF represents a faster induction of a counter mechanism that reduces an inappropriately facilitated motor output back toward normal levels.

Frequency facilitation and neural apnea.

All patterns of neural apnea tested here elicited frequency facilitation that differed only in magnitude, but not duration. It is noteworthy that frequency facilitation was more similar to iHMF than iPMF in its transient (∼15 min) duration; however, the reason(s) for this similarity is unknown. Because prolonged neural apnea elicited greater frequency facilitation than brief neural apnea, we hypothesize that inactivity-induced frequency facilitation is more sensitive to the duration of inactivity than its pattern. The cellular mechanisms responsible for frequency facilitation following neural apnea are not known, but likely occur in brainstem respiratory rhythm generating neurons (5, 11). Medullary neurons responsible for respiratory rhythm generation exhibit plasticity, such as frequency LTF following intermittent anoxia in vitro (11). Although it has not yet been directly shown, frequency LTF following intermittent hypoxia in vivo is hypothesized to similarly occur in medullary respiratory rhythm generating neurons (5, 11, 45). Similar to frequency facilitation following reduced respiratory neural activity, frequency LTF is small (∼10%) and relatively insensitive to the pattern of hypoxia (5).

Significance of repeated neural apnea and iPMF.

Recurrent apnea or hypopnea, particularly during sleep, is experienced by millions of people worldwide with varying degrees of severity (14, 63) and may be classified as central, obstructive, or mixed (both central and obstructive). Central apnea is characterized by a reduction in respiratory neural activity, whereas obstructive apnea is characterized by continued respiratory efforts against a collapsed airway. While the effects of intermittent hypoxia that occurs with apnea/hypopnea have received considerable attention (8, 30, 33, 40, 43, 46, 59), little work has addressed the impact of repeated fluctuations in respiratory neural activity. Both central apnea (7, 31, 52) and airway obstruction (53) elicit plasticity in respiratory control independent of hypoxia and hypercapnia that would normally accompany an apneic event. Mechanisms giving rise to these different forms of plasticity are not well known, but may involve reduced respiratory neural activity in efferent respiratory motor output (data presented here and see Refs. 31 and 52) or afferent respiratory feedback (53), respectively. Interestingly, central neural apnea appears to preferentially elicit phrenic plasticity in phrenic burst amplitude, whereas airway obstruction appears to preferentially elicit plasticity in genioglossus (and presumably hypoglossal) activity. Although the significance of these motor pool-specific responses is unknown, these differential effects may represent appropriate stimulus-specific responses to these very different threats to ventilation.

We hypothesize that iPMF represents one component in a continuum of endogenous responses to reduced respiratory neural activity. Given the prevalence of central apnea and the associated health risks (e.g., cardiovascular complications, frequent nighttime awakenings, excessive daytime sleepiness) (8, 17, 14), further investigation of the causes and potential endogenous compensatory responses to central apnea are warranted. To fully understand the physiology or pathophysiology of repeated central apnea, one must consider all associated physiological stimuli (e.g., chemical feedback, mechanical feedback, sympathoexcitation, and reduced network activity), as all are potential sources of plasticity, which may occur simultaneously (perhaps with important interactions) to benefit or hinder stable breathing. A greater understanding of conditions that efficiently recruit these mechanisms may guide treatment and prevention strategies for conditions that destabilize breathing.

GRANTS

Research support was provided by National Heart, Lung, and Blood Institute Grant HL-105511.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

N.A.B. and T.L.B.-H. conception and design of research; N.A.B. performed experiments; N.A.B. and T.L.B.-H. analyzed data; N.A.B. and T.L.B.-H. interpreted results of experiments; N.A.B. prepared figures; N.A.B. and T.L.B.-H. drafted manuscript; N.A.B. and T.L.B.-H. edited and revised manuscript; N.A.B. and T.L.B.-H. approved final version of manuscript.