| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
J Physiol (2003), 551.3, pp. 981-991
© Copyright 2003 The Physiological Society
DOI: 10.1113/jphysiol.2003.048157
 |
ABSTRACT |
|---|
 TopAbstract IntroductionMethodsResultsDiscussionReferences |
|---|
All respiratory long-term facilitation (LTF) is induced by inspiratory-excitatory stimulation, suggesting that LTF needs inspiratory augmentation and is the result of a Hebbian mechanism (coincident pre- and post-synaptic activity strengthens synapses). The present study examined the long-term effects of episodic inspiratory-inhibitory vagus nerve stimulation (VNS) on phrenic nerve activity. We hypothesized that episodic VNS would induce phrenic long-term depression. The results are compared with those obtained following serotonin receptor antagonism or episodic carotid sinus nerve stimulation (CSNS). Integrated phrenic neurograms were measured before, during and after three episodes of 5 min VNS (50 Hz, 0.1 ms), each separated by a 5 min interval, at a low (~50 µA), medium (~200 µA) or high (~500 µA) stimulus intensity in anaesthetized, vagotomized, neuromuscularly blocked and artificially ventilated rats. Medium- and high-intensity VNS eliminated rhythmic phrenic activity during VNS, while low-intensity VNS only reduced phrenic burst frequency. At 60 min post-VNS, phrenic amplitude was higher than baseline (35 ± 5 % above baseline, mean ± S.E.M., P < 0.05) in the high-intensity group but not in the low- (-4 ± 4 %) or medium-intensity groups (-10 ± 15 %), or in the high-intensity with methysergide group (4 mg kg-1, I.P.) (-11 ± 5 %). These data, which are inconsistent with our hypothesis, indicate that phrenic-inhibitory VNS induces a serotonin-dependent phrenic LTF similar to that induced by phrenic-excitatory CSNS (33 ± 7 %) and may require activation of high-threshold afferent fibres. These data also suggest that the synapses on phrenic motoneurons do not use the Hebbian mechanism in this LTF, as these motoneurons were suppressed during VNS.
(Received 27 May 2003; accepted after revision 17 July 2003; first published online 18 July 2003)
Corresponding author L. Ling: Division of Sleep Medicine, Brigham and Women's Hospital, Harvard Medical School, 221 Longwood Avenue, Boston, MA 02115, USA. Email: lling{at}partners.org
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Accumulating evidence suggests that the respiratory control system exhibits an impressive degree of plasticity, as do many other neural networks in the central nervous system (Eldridge & Millhorn, 1986; McCrimmon et al. 1995; Ling et al. 1997a; Powell et al. 1998; Mitchell et al. 2001). For example, episodic hypoxia induces a persistent augmentation of respiratory activity (Cao et al. 1992; Bach & Mitchell, 1996; Turner & Mitchell, 1997; Olson et al. 2001; McGuire et al. 2002), known as long-term facilitation (LTF). Episodic carotid sinus nerve (CNS) stimulation (CSNS) also elicits phrenic LTF in anaesthetized animals (Millhorn et al. 1980a; Hayashi et al. 1993; Ling et al. 1997b), which is not abolished by decerebration or spinal transection at the C7-T1 level (Eldridge & Millhorn, 1986). These results suggest that LTF is elicited by central mechanisms located in the brainstem and/or cervical spinal cord, and that the carotid body, respiratory mechanics, systemic hypoxia, forebrain and the lower spinal cord are not necessary for its expression (Eldridge & Millhorn, 1986). Both hypoxia- and CSNS-induced LTF are serotonin-dependent (Millhorn et al. 1980b; Bach & Mitchell, 1996). Recent evidence further suggests that spinal serotonin receptors (Baker-Herman & Mitchell, 2002) and the synapses transmitting bulbospinal, inspiratory drive to the phrenic motoneurons (Fuller et al. 2002) play key roles in phrenic LTF.
In these experiments, respiratory LTF is always preceded by repeated inspiratory augmentation. This fact suggests that induction of respiratory LTF is somehow related to inspiratory augmentation, and more specifically, that phrenic LTF relies heavily on an activity-dependent Hebbian mechanism (coincident pre- and post-synaptic activity strengthens synapses) in the synapses on the phrenic motoneurons. However, as all LTF to date has been induced by episodic inspiratory-excitatory stimulation, it has been difficult to directly test these hypotheses by separating the LTF from inspiratory augmentation.
In contrast, vagus nerve (VN) stimulation (VNS) suppresses inspiration during stimulation, and induces a reduction (< 1 min duration) in phrenic amplitude and frequency after stimulation (0.5 min duration; Eldridge & Millhorn, 1986). We speculate that a relatively longer post-stimulation inhibitory memory is possible if using episodic and longer VNS. VNS has also been used as a tool in brain research (e.g. evoked potentials recorded from the cerebral cortex, hippocampus, thalamus and cerebellum; Rutecki, 1990) and neurophysiological studies of several reflexes (e.g. cough, swallow and Hering-Breuer reflex or inspiratory off-switch mechanisms) because vagal afferents provide an easily accessible, peripheral route by which to modulate the central nervous system function. However, these investigators all used brief VNS and focused mainly on immediate or short-term (< several min) effects, during and/or after VNS. For many years, episodic VNS has been used clinically as a common treatment for patients with medically intractable epilepsy (Schachter, 2002). However, the precise underlying mechanisms and the consequence of long-term VNS remain unclear.
The aims of the present study were to (1) examine the long-term effects of episodic phrenic-inhibitory VNS on phrenic nerve activity and compare them with those elicited by phrenic-excitatory CSNS and (2) explore the possibility of using VNS as a tool to suppress phrenic motor neurons during LTF elicitation. We hypothesized that episodic VNS would induce phrenic long-term depression, but that phrenic activity would eventually return to baseline in less than 60 min.
|
METHODS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The Harvard Medical Area Standing Committee on Animals approved all experimental procedures used herein. Experiments were conducted on 33 adult male Sprague-Dawley rats (265-380 g, colony 236, Harlan, Madison, WI, USA). The present study includes three sets of experiments. The first set examined the long-lasting effects on phrenic nerve activity of episodic VNS at three different stimulus intensity levels. The second set examined the effects of high-intensity episodic VNS on the phrenic activity following serotonin receptor antagonism. The third set examined the effects of episodic CSNS on phrenic activity, for comparison.
Experimental preparation
The rats were anaesthetized initially with isoflurane in a closed chamber and then maintained in an anaesthetized state via a nose cone (2.5-3.0 % isoflurane; inspired O2 fraction (FI,O2) = 0.5; balance N2). The trachea was cannulated, and the rats were ventilated mechanically (Harvard Apparatus, Holliston, MA, USA) while maintaining the inspired isoflurane concentration. A bilateral vagotomy was performed in the mid-cervical region. Two femoral venous catheters were inserted bilaterally for administration of anaesthetic and fluid. A catheter was also placed into the femoral artery to allow blood pressure measurement and blood sample withdrawal for measurement of arterial blood gases and pH analysis. Rats were slowly converted from isoflurane to urethane anaesthesia (1.6 g kg-1 in distilled water, I.V., supplemented as needed), and the adequacy of anaesthesia was assessed periodically by testing corneal reflexes and blood pressure responses to toe pinch (after neuromuscular blockade). Rats were then neuromuscularly blocked using pancuronium bromide injection (2.5 mg kg-1, I.V., supplemented as needed). A slow infusion of sodium bicarbonate (5 %) and Ringer solution with sodium lactate (50:50, ~1.7 ml kg-1 h-1) was initiated approximately 1 h after induction of anaesthesia to maintain fluid and acid-base balance. Rectal temperature was monitored and maintained near 37.5 °C with a servo-controlled heat blanket.
End-tidal CO2 partial pressure (PET,CO2) was monitored in the expired line of the ventilator circuit using a flow-through capnograph (Novametrix; Wallingford, CT, USA) with sufficient response time (< 75 ms) to measure PET,CO2 in rats. PET,CO2 values obtained with this method approximate the CO2 partial pressure in arterial blood (Pa,CO2) in most rats (usually within 1-2 mmHg). Inspired gases were 50 % O2 (balance N2, FI,O2 = 0.50) under baseline conditions to improve tolerance to experimental stresses and prolong the viability of the preparation. At the end of the experiments, rats were killed by a urethane overdose.
Isolation of nerves for stimulation and recording
In the VNS experiments, after performing a bilateral cervical vagotomy, the proximal part of the left VN was isolated via a dorsal approach and hooked onto a stimulating bipolar electrode (diameter ~0.5 mm). In the CSNS experiments, the left CSN was also isolated via a dorsal approach, cut distally and mounted on a fine bipolar silver wire electrode (diameter ~0.15 mm).
In all experiments, the left phrenic nerve was dissected via a dorsal approach, cut distally, desheathed and prepared for recording (bipolar silver wire electrode). Phrenic nerve activity was filtered (300-10 000 Hz) and amplified (2000 
, BMA-200 AC/DC Bioamplifier, CWE, Ardmore, PA, USA). The amplified signal was full-wave rectified and integrated (Paynter Filter, BAK Electronics, USA; time constant = 100 ms). The integrated phrenic signals were digitized and acquired with computer software (LabView 5.0, National Instruments, USA), and analysed with a program that was developed in our laboratory. This program determines the amplitude and timing of integrated phrenic activity, from which the minute phrenic activity can be calculated.
Experimental protocols
Following completion of the surgical procedures, the preparation was allowed to stabilize for at least 1 h under hyperoxic (FI,O2 = 0.5, Pa,O2 > 150 mmHg) and hypocapnic conditions (PET,CO2 ~30 mmHg, no phrenic nerve discharge). The phrenic discharge was then resumed by manipulating the respiratory pump rate and/or volume to slowly decrease ventilation and raise the PET,CO2. The CO2-apnoeic threshold was defined as the PET,CO2 at which the respiratory rhythmic activity started to reappear in the phrenic nerve recording. Baseline phrenic nerve activity was set at 3 mmHg above this threshold and was allowed to stabilize for about 30 min. One or two arterial blood samples (~0.3 ml collected into a 1 ml heparinized syringe; unused blood was returned to the animal) were drawn for blood gases and pH analysis (ABL-700; Radiometer, Copenhagen, Denmark) with correction for rectal body temperature. All subsequent blood samples were compared to this initial baseline value.
In all experiments, integrated phrenic nerve activity was measured before, during and up to 60 min (at 15, 30 and 60 min post-stimulation) after each episodic nerve stimulation protocol to determine the baseline phrenic response to nerve stimulation and phrenic LTF, respectively. Blood samples were also taken at all these points to ensure an isocapnic condition. If any deviations in PET,CO2 (or Pa,CO2) from isocapnic conditions were noted, corrections were made by adjusting the inspired CO2 fraction, thus assuring that Pa,CO2 was generally within 1 mmHg of the baseline value. At the end of the protocol, the phrenic responses to hypercapnia (PET,CO2 = 90-95 mmHg) were recorded to obtain approximately maximal phrenic nerve activity.
VNS protocols. These stimulation protocols consisted of three episodes of 5 min VNS, each separated by a 5 min interval. Electrical pulse stimulation (50 Hz, 0.1 ms pulse duration) was delivered to the central end of the left cervical VN (Grass S88 Stimulator, Grass Instrument, W. Warwick, RI, USA) at three different intensity levels. The low-intensity group (n = 5) received an average constant current of 50 ± 5 µA (Isolation unit PSIU 6; Grass Instrument, W. Warwick, RI, USA), the medium-intensity group received 200 ± 18 µA (n = 6) and the high-intensity group received 507 ± 53 µA (n = 8). The purpose of using three different stimulus intensity levels was to differentially activate the vagal afferent nerve fibres with different thresholds. In most rats of the high-intensity group, it was necessary to progressively increase the stimulus intensity during the 5 min stimulation and along the three episodes to completely eliminate rhythmic phrenic activity. The reason for selecting three episodes of 5 min stimulation in the protocols was because these are the most commonly used protocols (three episodes of 5 min hypoxia) in studies with hypoxia-induced LTF (Hayashi et al. 1993; Bach & Mitchell, 1996; Ling et al. 2001; Bavis & Mitchell, 2003). VNS would have been used as a tool to suppress phrenic activity during induction of LTF using episodic hypoxia if VNS per se had caused no long-lasting effects on phrenic activity. The following two protocols were added after we observed that episodic VNS induced phrenic LTF, a result that was inconsistent with our original hypothesis.
VNS with methysergide protocol. This protocol (n = 6) also consisted of three episodes of 5 min VNS at a high intensity (50 Hz, 0.1 ms, 517 ± 6 µA), separated by 5 min intervals. Methysergide was injected systemically (4 mg kg-1, I.P.) about 20 min before the VNS. Methysergide maleate (chemical formula: C21H27N3O2.C4H4O4) is a widely used, broad-spectrum serotonin (5-HT1, 2, 5, 6, 7) receptor antagonist (RBI Sigma, USA) that easily crosses the blood-brain-barrier. Baseline phrenic activity increased after methysergide in each experiment (24 ± 3 % above the original baseline level), which was also reflected by an increased ratio of phrenic baseline to maximal CO2 response (48.9 ± 2.4 % versus a range of 28.9-36.8 % in the other four groups). This might cause a ceiling effect since a higher baseline usually reduces LTF magnitude or even abolishes LTF in some extreme cases. To rule out this possible confounding factor, the post-drug phrenic baseline was intentionally brought down to the original baseline level in one rat by carefully adjusting ventilation to reduce the PET,CO2. The results from that rat were the same as those from other rats of the group.
CSNS protocol. This protocol (n = 8) consisted of five episodes of 2 min CSNS (20 Hz, 0.1 ms duration, 30-100 µA (3 phrenic activity threshold)), separated by 4 min intervals, which is similar to the protocol used in other studies of CSNS-induced phrenic LTF (Eldridge & Millhorn, 1986; Hayashi et al. 1993; Fregosi & Mitchell, 1994; Ling et al. 1997b). The purpose of this protocol was to allow a quantitative comparison of VNS-induced phrenic LTF with the well-known CSNS-induced LTF under the same experimental conditions (experimenter and setting). Since the serotonin dependency of CSNS-induced phrenic LTF has been well documented (Millhorn et al. 1980b; Eldridge & Millhorn, 1986), a corresponding CSNS with methysergide group was not included. The difference in episode number and duration between the VNS and CSNS protocols was unlikely to significantly affect the comparison results as the magnitude of LTF was probably already saturated in both cases (cf. Eldridge & Millhorn, 1986; McGuire et al. 2002).
Data analysis
Results included in the analysis were collected from successful experiments in which the arterial blood pressure was always > 80 mmHg, Pa,O2 was > 150 mmHg and Pa,CO2 was within 1.5 mmHg of the baseline value throughout an experiment. Integrated phrenic nerve activity was averaged over 50-60 inspiratory bursts at each point (baseline, stimulation episodes (CSN only), and 15, 30 and 60 min post-stimulation). Variables determined include: the peak amplitude of integrated phrenic nerve activity (Phr), phrenic nerve burst frequency (f, bursts min-1) and minute phrenic nerve activity (Phr f). For phrenic LTF and phrenic responses to CSNS, changes from baseline in Phr and Phr f were normalized as a percentage of the baseline (%baseline). Changes from baseline in f were measured in absolute units (bursts min-1). Phrenic responses to CSNS were recorded during the last minute of stimulation when Phr reached a plateau, and were averaged over the five CSNS episodes since there was little difference among them.
For LTF, the changes from baseline within each group and the differences between groups in the post-stimulation Phr, f or Phr f were determined using a two-way ANOVA with repeated measures, followed by a post hoc test (Student-Newman-Keuls). Only baseline (value set at zero for each rat) and post-stimulation data (relative to baseline) were included in this two-way ANOVA (data recorded during stimulation not included). For blood gases, blood pressure values and phrenic responses during CSNS, a one-way ANOVA was used to determine the differences between and within groups. P < 0.05 was considered significant. All values are expressed as means ± S.E.M.
|
RESULTS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Blood gas and blood pressure responses
Pa,CO2 was unchanged after each of the five stimulation protocols (Table 1), demonstrating that isocapnia was maintained throughout the experiment in all five groups. The post-stimulation Pa,O2 values were also similar to baseline in all stimulation groups except for the low-intensity VNS and CSNS groups, in which the post-stimulation Pa,O2 values were significantly decreased from baseline but still above 150 mmHg in each experiment (Table 1).
Arterial blood pressure responded to VNS by an initial increase that lasted approximately 30 s and then gradually declined to a relatively stable level for the remainder of the stimulation (Fig. 1). This dynamic blood pressure response appeared to be similar in all three intensity VNS groups (Fig. 1) and in the VNS with methysergide group. However, in the low-intensity group, the average stable blood pressure value (138 ± 2 mmHg) was significantly (P < 0.05) higher than baseline (121 ± 5 mmHg). The average stable blood pressure values were not significantly (all P > 0.1) different from baseline in the medium-intensity (134 ± 6 versus 119 ± 6 mmHg), high-intensity (119 ± 9 versus 124 ± 5 mmHg) or the VNS with methysergide (122 ± 4 versus 123 ± 4 mmHg) groups. In contrast, blood pressure responded to CSNS by an initial decrease that lasted 20-40 s and then returned to near-baseline levels for the remainder of the stimulation (Fig. 1). The average value during stimulation (126 ± 3 mmHg) was not significantly different from baseline (126 ± 2 mmHg).
 |
View larger version [in this window] [in a new window] |
|
|
Figure 1. Integrated phrenic neurograms recorded before, during and after episodic electrical stimulation of the VN (A) or CSN (B) A, integrated phrenic nerve (Phr) and mean arterial blood pressure (MAP) responses to three episodes of 5 min electrical stimulation (50 Hz, 0.1 ms duration) delivered to the proximal end of the left cervical VN in three individual rats, at a low (50 µA), medium (200 µA) and high (500 µA) stimulus intensity. B, Phr and MAP responses to five episodes of 2 min CSNS (20 Hz, 0.1 ms duration, 30 µA) in one rat. Notice that data from only one episode during nerve stimulation are presented in these recordings and note that the recording paper speed during the 5 min of stimulation is different from the rest. | ||
Phrenic nerve responses
VNS groups. For phrenic responses during stimulation, the low-intensity VNS substantially reduced the phrenic burst frequency but did not eliminate rhythmic phrenic activity. These peak phrenic amplitudes were even greater than those in baseline (Fig. 1). Both medium- and high-intensity VNS, including VNS with methysergide, eliminated rhythmic phrenic activity in each experiment (Fig. 1). Tonic phrenic activity gradually increased in 9/14 high-intensity rats, but not in low- or medium-intensity rats. The stimulus intensity required to eliminate rhythmic phrenic activity was progressively increased during the 5 min stimulation period and during the three episodes in 8/14 high-intensity rats, 300-400 µA in the first, 400-500 µA in the second and 500-900 µA in the third episode. In 11/14 high-intensity rats, phrenic activity resumed immediately after termination of VNS and quickly returned to (or slightly above) baseline level. In 3/14 rats, however, it took 5-10 s for phrenic activity to reappear, slowly returning to near-baseline level over several minutes. There appeared to be no difference in phrenic responses during stimulation between the VNS and the VNS with methysergide groups.
For phrenic LTF of VNS groups of all three intensities, there was a significant interaction between intensity and time factors in integrated phrenic amplitude (Phr, F(6, 45) = 3.85; P = 0.0035) and minute phrenic activity (Phr f, F(6, 45) = 2.85; P = 0.02) but not in phrenic burst frequency (f, F(6, 45) = 0.6; P = 0.7). The low-intensity episodic VNS did not produce long-lasting effects on the phrenic nerve activity at any post-stimulation time point (all P > 0.75, Fig. 2). The medium-intensity VNS appeared to reduce post-stimulation Phr, f and Phr f; however, these values were not significantly different from baseline (Fig. 2). The high-intensity VNS induced a progressive increase in post-stimulation Phr and Phr f (i.e. a progressively augmenting pattern). Phr (35.4 ± 4.5 % above baseline) and Phr f (39.5 ± 6.0 % above baseline) at 60 min post-stimulation were significantly greater than the baseline (both P < 0.05, Fig. 2) and the medium- (-10.1 ± 14.7 % and -9.9 ± 22.6 %) and low-intensity (-3.6 ± 4.4 % and -1.1 ± 5.6 %) groups (all P < 0.05). However, the high-intensity VNS caused little change in post-stimulation f (all P > 0.50; Fig. 2). These data indicate that high-intensity, but not medium- or low-intensity, episodic VNS induces phrenic (both Phr and Phr f) LTF, thus disproving our original hypothesis. The medium-intensity group data also indicated that repeated elimination of phrenic activity alone could not elicit LTF.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 2. The effects of episodic VNS on phrenic nerve activity A, average changes from baseline in peak amplitude of integrated phrenic nerve activity normalized to a percentage of the baseline ( | ||
In comparing VNS only and the VNS with methysergide groups (both using high-intensity stimulation), there was a significant interaction between drug and time factors in Phr (F(3, 36) = 6.97; P = 0.0008) and Phr f (F(3, 36) = 5.37; P = 0.004). The Phr (-11.4 ± 3.5 %) and Phr f (-10.8 ± 4.8 %) at 60 min post-stimulation in the VNS with methysergide group were not significantly different from their baseline values (both P > 0.26), but were significantly lower than the values in the VNS only group (both P < 0.05, Fig. 3). The lower post-stimulation Phr and Phr f values might result partially from the drug efficacy decline as methysergide per se somehow increased phrenic activity (see Methods). These data indicate that the VNS-induced LTF in Phr and Phr f was abolished by pre-treatment with methysergide (4 mg kg-1, I.P.), suggesting that this VNS-induced phrenic LTF is serotonin dependent.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 3. LTF of phrenic nerve activity elicited by VNS in untreated control (n = 8, open squares) and methysergide-pre-treated (n = 6, filled squares) rats A, average changes from baseline in peak amplitude of integrated phrenic nerve activity normalized to a percentage of the baseline ( | ||
CSNS group. In contrast to VNS, CSNS augmented phrenic nerve activity during stimulation. CSNS produced an immediate increase in phrenic activity, which then quickly reached a plateau in Phr (176 ± 32 % above baseline), and the more stable f (10 ± 1 bursts min-1 above baseline) and Phr f (239 ± 43 % above baseline) during stimulation (all P < 0.05, Fig. 1). CSNS also induced a significant increase from baseline in Phr at all post-stimulation time points, and in Phr f at 30 and 60 min post-stimulation (all P < 0.05, Fig. 4), but caused little change in f at any post-stimulation point. The Phr f (40.2 ± 8.1 %) at 60 min post-stimulation was also significantly higher than the values at 15 and 30 min post-stimulation (P < 0.05; Fig. 4). These data suggest that episodic CSNS induces phrenic LTF with a progressively augmenting pattern.
 |
View larger version [in this window] [in a new window] |
|
|
Figure 4. LTF of phrenic nerve activity elicited by CSNS (triangles) and VNS (squares) A, average changes from baseline in peak amplitude of integrated phrenic nerve activity normalized to a percentage of the baseline ( | ||
In comparing the CSNS and high-intensity VNS groups, the interaction between nerve and time factors in Phr (F(3, 42) = 0.34; P = 0.8) or Phr f data (F(3, 42) = 0.003; P = 0.99) was not significant (Fig. 4). There was also little difference between groups in all post-stimulation Phr and Phr f (all P > 0.40; Fig. 4). These data indicate that the VNS- and CSNS-induced LTF (in Phr and Phr f) are similar in both magnitude and pattern.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The results of the present study demonstrate that episodic VNS induced phrenic LTF that was similar in magnitude, augmenting pattern and serotonin dependency to that elicited by episodic CSNS, thus disproving our original hypothesis. In contrast to CSNS, which greatly enhances phrenic activity during stimulation, VNS induced LTF despite the elimination of phrenic activity during stimulation. This VNS-induced phrenic LTF is likely to require activation of high-threshold afferent fibres, as low- and medium-intensity VNS failed to elicit LTF. These results suggest that the induction of phrenic LTF can be dissociated from inspiratory augmentation and that the synapses on phrenic motoneurons do not use the Hebbian mechanism in this LTF, as these motoneurons were suppressed during VNS.
Phrenic LTF
Respiratory (phrenic) LTF was first demonstrated by Millhorn et al. (1980a) and was elicited by episodic CSNS in anaesthetized, vagotomized, neuromuscularly blocked and artificially ventilated cats. Manifestation of this LTF requires the activation of serotonin receptors, since it can no longer be elicited following serotonin receptor antagonism by systemic injection of methysergide, and is impaired following application of the tryptophan hydroxylase (the rate-limiting enzyme in serotonin biosynthesis) inhibitor para-chlorophenylalanine or the serotonergic neurotoxin 5,7-dihydroxytryptamine (Eldridge & Millhorn, 1986). Our results demonstrate that VNS-induced phrenic LTF is abolished by systemic injection of methysergide, thereby also indicating a serotonin dependency. In addition, our results demonstrate that both the magnitude and pattern of VNS-induced phrenic LTF are very similar to those of CSNS-induced LTF.
Despite the similarity between the two types of LTF, we argue strongly against the possibility that high-intensity VNS might indirectly stimulate the CSN through current spread and thus induce a similar LTF. First, the low- and medium-intensity (~200 µA) VNS failed to induce LTF, but CSNS could induce LTF using very low stimulus intensity (30-100 µA). Second, VNS induced the same LTF in one experiment in which the left CSN had been severed. Finally, we used a rather thick mixture of mineral oil and Vaseline to cover and protect the VNs, and the vagal stimulating electrode was physically separated from other tissues at the exposed jugular foramen (~10 mm in the shortest distance and ~15 mm to the tissue containing CSN), thereby limiting the spread of stimulation to other tissues.
Episodic hypoxia-induced LTF in anaesthetized rats persists for at least 1 h and shows an augmenting pattern (Bach & Mitchell, 1996; Zabka et al. 2001). However, in two previous studies on anaesthetized rats, CSNS-induced phrenic LTF showed a decrementing pattern (Hayashi et al. 1993; Ling et al. 1997b), suggesting that the two types of LTF have different patterns (Mitchell et al. 2001). Our CSNS results showed that although post-stimulation Phr appeared to decline at 30 min, it resumed a progressive increase afterwards. Therefore, it is likely that the augmenting pattern would have been obtained in the two previous studies if the data had been collected 60 min, instead of just 30 min, after stimulation.
Dissociation of LTF induction from inspiratory enhancement
Respiratory LTF has been elicited in many animal species by CSNS or (both isocapnic and poikilocapnic) hypoxia (cf. Mitchell et al. 2001). All such stimuli enhance inspiration and are delivered intermittently. Sustained hypoxia (even with longer exposures) does not induce LTF in either anaesthetized or awake animals (Dwinell et al. 1997; Baker & Mitchell, 2000). As far as we are aware, LTF has never been induced by inspiratory-inhibitory stimulation, which leads to several questions: (1) is respiratory enhancement during stimulation necessary for LTF? (2) Is LTF just a residual respiratory enhancement after stimulation? (3) Can LTF be elicited by stimuli that inhibit inspiration during stimulation? (4) Does the size of inspiratory enhancement during stimulation determine the magnitude of LTF? (5) Does the Hebbian plasticity play a role in the synapses on phrenic motoneurons? To directly address these issues, we need to be able to dissociate between LTF and inspiratory or phrenic augmentation, and the VNS protocol used here provides such a unique opportunity.
CSNS induces a short-term potentiation of phrenic activity (a progressive increase over 1-2 min, followed by a similar progressive decrease when the stimulus is removed). If CSNS is delivered several times consecutively, the residual phrenic activity level will progressively increase and stay elevated for a period of time (i.e. LTF; Eldridge & Millhorn, 1986). Ventilatory LTF induced in awake animals also shows this pattern and gradually declines to the original baseline level (Cao et al. 1992; Turner & Mitchell, 1997; McGuire et al. 2002), suggesting that LTF is a residual respiratory enhancement. However, this concept is in some ways challenged by the augmenting pattern of phrenic LTF elicited by episodic hypoxia in anaesthetized rats (Bach & Mitchell, 1996; Zabka et al. 2001). Our results demonstrate that phrenic LTF can be elicited by stimuli that inhibit inspiration during stimulation, suggesting that LTF is not just a residual enhancement, but requires additional mechanisms. There is a weak but significant correlation between the acute hypoxia ventilatory response and the magnitude of LTF (Fuller et al. 2000), suggesting that the size of the hypoxic enhancement somehow determines the magnitude of the LTF. Our results, however, totally dissociated these two events.
Homosynaptic Hebbian or heterosynaptic modulatory mechanisms
In 1949, Donald Hebb postulated that synapses are strengthened when the pre- and post-synaptic elements are synchronously active. Over several decades, this Hebbian postulate has become a cornerstone in our understanding of activity-dependent neural development (e.g. ocular dominance plasticity; Wiesel, 1982) and the cellular basis of learning and memory. Certain synapses exhibit a long-lasting increase in their efficacy after they are heavily used, which is termed long-term potentiation (LTP) and has been identified in many areas of the brain (e.g. the hippocampus and amygdala; Bliss & Collingridge, 1993). Many believe that LTP serves as a model for the cellular mechanisms of learning and memory (Bliss & Collingridge, 1993). LTP in the Schaffer collateral pathways of the hippocampus is induced as a consequence of coincident pre- and post-synaptic activity (Bliss & Collingridge, 1993), and this LTP can no longer be induced if the post-synaptic depolarization is suppressed (cf. Kandel, 2000), in conformity with Hebb's rule.
It is tempting to hypothesize that Hebbian plasticity also plays a role in phrenic LTF, which is similar in some ways to hippocampal LTP, and more specifically that the synapses directly on the phrenic motoneurons may use this activity-dependent Hebbian mechanism in phrenic LTF. Several lines of evidence (Baker-Herman & Mitchell, 2002; Fuller et al. 2002) have suggested that these synapses are crucial for phrenic LTF. Their pivotal role in LTF is further supported by our recent experiments (McGuire & Ling, unpublished results), in which microinjection of the NMDA receptor antagonist MK-801 into the phrenic motor nucleus region totally abolished phrenic LTF. However, our results in the present study unequivocally reject the notion that a Hebbian mechanism is involved in the synapses on the phrenic motoneurons, as phrenic LTF could be induced by episodic VNS while the post-synaptic, phrenic motoneurons were totally suppressed. This conclusion does not, however, rule out the possibility that a Hebbian mechanism may play a role in other synapses involved in LTF (e.g. those synapses on the pre-motor neurons or on the raphe neurons). There is other indirect evidence against the role of Hebbian mechanism in the synapses on phrenic motoneurons. Episodic calf muscle stimulation (Eldridge & Millhorn, 1986) or episodic hypercapnia (Eldridge & Millhorn, 1986; Baker et al. 2001) did not induce LTF, despite similar coincident pre- and post-synaptic activity enhancement. On the other hand, although hypoxic phrenic response was greatly reduced after CSN section, LTF of a smaller magnitude could still be elicited by episodic hypoxia (Bavis & Mitchell, 2003).
There are generally two basic categories of synaptic plasticity: homosynaptic, Hebbian activity dependent and heterosynaptic, modulatory input dependent. In Hebbian plasticity, the events responsible for inducing the plastic change occur at the same synapse that is being strengthened. In heterosynaptic plasticity, the plastic change often involves new protein synthesis and can occur in the absence of activity in the synapse being strengthened, but instead, as a result of modulatory inputs from a third neuron (Bailey et al. 2000). Our data suggest that the VNS-induced LTF uses heterosynaptic plasticity in those synapses on phrenic motoneurons and that the raphe serotonergic neurons are the source (third neurons), providing the required modulatory inputs.
It is known that phrenic LTF is serotonin dependent and thus definitely uses modulatory mechanisms. However, this does not necessarily rule out the involvement of a Hebbian mechanism somewhere in this complex LTF, which requires multiple synapses and appears to depend on a functioning neural network, because the two basic categories of plasticity are not mutually exclusive. On many occasions, both homosynaptic and heterosynaptic mechanisms participate at the synaptic level and a non-additive (synergetic) interaction between the two mechanisms is required (Bailey et al. 2000). For example, the long-lasting, late phase of homosynaptic LTP in the hippocampus requires a heterosynaptic modulatory input. The classical conditioning of the gill- and tail-withdrawal reflexes in Aplysia also involves both homo- and heterosynaptic plasticity at the same synapses between the sensory neurons and the motoneurons (Bailey et al. 2000).
Potential mechanisms
The hypothetical mechanism of phrenic LTF has been proposed and refined many times (Eldridge & Millhorn, 1986; Fuller et al. 2000; Mitchell et al. 2001). Briefly, carotid body chemo-afferents, activated by episodic hypoxia or CSNS, project through multiple synapses to both the raphe nuclei and the integrative centres responsible for respiratory rhythm generation and burst pattern formation. These medullary nuclei then activate each other as well as the phrenic motoneurons. Serotonin released from the raphe serotonergic neurons activates 5-HT2 receptors on the phrenic motoneurons, which initiate a series of intracellular signalling events, leading to phrenic LTF. It is unclear how VNS induces LTF. However, based on our results and other available information, we speculate that VNS stimulates high-threshold unmyelinated C fibres that transmit vagally mediated chemoreceptor inputs (Martin-Body et al. 1985; Berthoud & Neuhuber, 2000), which then activate the raphe serotonergic neurons and the same 'downstream machinery' described above, thereby generating a similar LTF.
The VN is often considered a parasympathetic nerve, but is actually a mixed nerve, containing about 80 % sensory fibres (left vagus), most of them small-diameter unmyelinated C fibres with high thresholds (Rutecki, 1990; George et al. 2000). Through multisynaptic projections, these fibres have a wide distribution throughout the central nervous system, including the midline raphe nuclei and locus coeruleus (George et al. 2000). There are numerous paraganglia associated with the abdominal VN (McDonald & Blewett, 1981; Kummer & Neuhuber 1989). Many believe that these paraganglia are arterial chemoreceptors like the carotid body, since they also contain clusters of glomus cells (which are capable of catecholamine synthesis), neurons and extensive capillary beds, receive sensory innervation from the nodose ganglion and respond electrophysiologically to hypoxia (McDonald & Blewett, 1981; Kummer & Neuhuber 1989; Berthoud & Neuhuber, 2000). In the present study, LTF was probably elicited by stimulating the abdominal chemoreceptor C fibres (aortic chemoreceptors are less important than abdominal chemoreceptors in rats; Martin-Body et al. 1986), as LTF was not elicited by low- or medium-intensity VNS and the threshold for activation of C fibres is 10-100 times greater than that for myelinated A and B fibres with larger diameters (Woodbury & Woodbury, 1990). Both clinical and animal studies also suggest that VNS stimulates serotonin release, as the cerebrospinal fluid levels of 5-hydroxyindoleacetic acid (a metabolite of serotonin) is increased after VNS (Ben-Menachem et al. 1995).
VNS also activates many inhibitory mechanisms (e.g. releasing GABA and glycine in widespread brain structures; Woodbury & Woodbury, 1990). Thus, the observed post-VNS change in phrenic activity could be the outcome of both inhibitory and excitatory mechanisms, with phrenic LTF probably resulting from the excitatory mechanisms that eventually outlast the inhibitory ones. It should be noted, however, that these mechanistic discussions are mostly speculative and await more direct experimental verification. For example, we may exclusively stimulate the abdominal chemoreceptor afferents (not the whole VN) to induce LTF, or selectively block this LTF by microinjecting (not systemically) a serotonin receptor antagonist into the phrenic motor nucleus region. We are aware that the conclusions (dissociation of LTF induction with inspiratory enhancement and non-Hebbian mechanism involved in the synapses on the phrenic motoneurons) may also apply to other LTF forms (e.g. hypoxia- or CSNS-induced ones). However, caution should be used for such an extrapolation since the VNS-induced LTF might be distinct from others.
In conclusion, although VNS totally eliminates phrenic activity during stimulation, episodic VNS induces a phrenic LTF that is similar in many ways to that induced by episodic, phrenic-excitatory CSNS. These data suggest that induction of respiratory LTF can be dissociated from inspiratory enhancement and that LTF requires separate mechanisms other than those responsible for the respiratory responses to short-term hypoxia. These results also suggest that the Hebbian activity-dependent plasticity plays no role in this phrenic LTF, at least at the synapses on the phrenic motoneurons, since those post-synaptic neurons are suppressed during VNS.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| Bach KB , & Mitchell GS (1996). Hypoxia-induced long-term facilitation of respiratory activity is serotonin dependent. Respir Physiol 104, 251-260 | [Medline] |
| Bailey CH, Giustetto M, Huang YY, Hawkins RD & Kandel ER (2000). Is heterosynaptic modulation essential for stabilizing Hebbian plasticity and memory? Nat Rev Neurosci 1, 11-20 | [Medline] |
| Baker TL , & Mitchell GS (2000). Episodic but not continuous hypoxia elicits long-term facilitation of phrenic motor output in rats. J Physiol 529, 215-219 | [Abstract/Full Text] |
| Baker-Herman TL , & Mitchell GS (2002). Phrenic long-term facilitation requires spinal serotonin receptor activation and protein synthesis. J Neurosci 22, 6239-6246 | [Abstract/Full Text] |
| Bavis RW , & Mitchell GS (2003). Plasticity in respiratory motor control: selected contribution: intermittent hypoxia induces phrenic long-term facilitation in carotid-denervated rats. J Appl Physiol 94, 399-409 | [Abstract/Full Text] |
| Ben-Menachem E, Hamberger A, Hedner T, Hammond EJ, Uthman BM, Slater J, Treig T, Stefan H, Ramsay RE, Wernicke JF & Wilder BJ (1995). Effects of vagus nerve stimulation on amino acids and other metabolites in the CSF of patients with partial seizures. Epilepsy Res 20, 221-227 | [Medline] |
| Berthoud HR , & Neuhuber WL (2000). Functional and chemical anatomy of the afferent vagal system. Auton Neurosci 85, 1-17 | [Medline] |
| Bliss TV , & Collingridge GL (1993). A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361, 31-39 | [Medline] |
| Cao K-Y, Zwillich CW, Berthon-Jones M & Sullivan CE (1992). Increased normoxic ventilation induced by repetitive hypoxia in conscious dogs. J Appl Physiol 73, 2083-2088 | [Abstract] |
| Dwinell MR, Janssen PL & Bisgard GE (1997). Lack of long-term facilitation of ventilation after exposure to hypoxia in goats. Respir Physiol 108, 1-9 | [Medline] |
| Eldridge FL , & Millhorn DE (1986). Oscillation, gating, and memory in the respiratory control system. In Handbook of Physiology section 3, The Respiratory System, vol. II, Control of Breathing, ed. Cherniack NS & Widdicombe JG, pp. 93-114. American Physiological Society, Bethesda, MD, USA | |
| Fregosi RF , & Mitchell GS (1994). Long-term facilitation of inspiratory intercostal nerve activity following carotid sinus nerve stimulation in cats. J Physiol 477, 469-479 | [Abstract] |
| Fuller DD, Bach KB, Baker TL, Kinkead R & Mitchell GS (2000). Long term facilitation of phrenic motor output. Respir Physiol 121, 135-146 | [Medline] |
| Fuller DD, Johnson SM & Mitchell GS (2002). Respiratory long term facilitation (LTF) is associated with enhanced spinally evoked phrenic potentials (abstract). Soc Neurosci Abstr 363.1., George MS, Sackeim HA, Rush AJ, Marangell LB, Nahas Z, Husain MM, Lisanby S, Burt T, Goldman J & Ballenger JC | |
| Hayashi F, Coles SK, Bach KB, Mitchell GS & McCrimmon DR (1993). Time-dependent phrenic nerve responses to carotid afferent activation: intact vs. decerebellate rats. Am J Physiol 265, R811-819 | [Medline] |
| Hebb D, (1949). The Organization of Behavior. Wiley, New York | |
| Kandel ER, (2000). Cellular mechanisms of learning and the biological basis of individuality. In Principles of Neural Science, ed. Kandel ER, Schwartz JH & Jessell TM, pp. 1247-1279. McGraw-Hill, New York | |
| Kummer W , & Neuhuber WL (1989). Vagal paraganglia of the rat. J Electron Microsc Tech 12, 343-355 | [Medline] |
| Ling L, Fuller DD, Bach KB, Kinkead R, Olson EB Jr & Mitchell GS (2001). Chronic intermittent hypoxia elicits serotonin-dependent plasticity in the central neural control of breathing. J Neurosci 21, 5381-5388 | [Abstract/Full Text] |
| Ling L, Olson EB Jr, Vidruk EH & Mitchell GS (1997a). Developmental plasticity of the hypoxic ventilatory response. Respir Physiol 110, 261-268 | [Medline] |
| Ling L, Olson EB, Vidruk EH & Mitchells GS (1997b). Integrated phrenic responses to carotid afferent stimulation in adult rats following perinatal hyperoxia. J Physiol 500, 787-796 | [Abstract] |
| Martin-Body RL, Robson GJ & Sinclair JD (1985). Respiratory effects of sectioning the carotid sinus glossopharyngeal and abdominal vagal nerves in the awake rat. J Physiol 361, 35-45 | [Abstract] |
| Martin-Body RL, Robson GJ & Sinclair JD (1986). Restoration of hypoxic respiratory responses in the awake rat after carotid body denervation by sinus nerve section. J Physiol 380, 61-73 | [Abstract] |
| McCrimmon DR, Dekin MS & Mitchell GS (1995). Glutamate, GABA, and serotonin in ventilatory control. In Lung Biology in Health and Disease: Regulation of Breathing, ed. Dempsey JA & Pack AI, pp 151-218. Dekker, New York | |
| McDonald DM , & Blewett RW (1981). Location and size of carotid body-like organs (paraganglia) revealed in rats by the permeability of blood vessels to Evans blue dye. J Neurocytol 10, 607-643 | [Medline] |
| McGuire M, Zhang Y, White DP & Ling L (2002). Effect of hypoxic episode number and severity on ventilatory long-term facilitation in awake rats. J Appl Physiol 93, 2155-2161 | [Abstract/Full Text] |
| Millhorn DE, Eldridge FL & Waldrop TG (1980a). Prolonged stimulation of respiration by a new central neural mechanism. Respir Physiol 42, 87-103 | [Medline] |
| Millhorn DE, Eldridge FL & Waldrop TG (1980b). Prolonged stimulation of respiration by endogenous central serotonin. Respir Physiol 42, 171-198 | [Medline] |
| Mitchell GS, Baker TL, Nanda SA, Fuller DD, Zabka AG, Hodgeman BA, Bavis RW, Mack KJ & Olson EB Jr (2001). Invited review: Intermittent hypoxia and respiratory plasticity. J Appl Physiol 90, 2466-2475 | [Abstract/Full Text] |
| Olson EB Jr, Bohne CJ, Dwinell MR, Podolsky A, Vidruk EH, Fuller DD, Powell FL & Mitchel GS (2001). Ventilatory long-term facilitation in unanesthetized rats. J Appl Physiol 91, 709-716 | [Abstract/Full Text] |
| Powell FL, Milsom WK & Mitchell GS (1998). Time domains of the hypoxic ventilatory response. Respir Physiol 112, 123-134 | [Medline] |
| Rutecki P, (1990). Anatomical, physiological, and theoretical basis for the antiepileptic effect of vagus nerve stimulation. Epilepsia 31, S1-6 | |
| Schachter SC, (2002). Vagus nerve stimulation therapy summary: five years after FDA approval. Neurology 59, S15-20 | |
| Turner DL , & Mitchell GS (1997). Long-term facilitation of ventilation following repeated hypoxia episodes in awake goats. J Physiol 499, 543-550 | [Abstract] |
| Wiesel TN, (1982). Postnatal development of the visual cortex and the influence of environment. Nature 299, 583-591 | [Medline] |
| Woodbury DM , & Woodbury JW (1990). Effects of vagal stimulation on experimentally induced seizures in rats. Epilepsia 31, S7-19 | [Medline] |
| Zabka AG, Behan M & Mitchell GS (2001). Long term facilitation of respiratory motor output decreases with age in male rats. J Physiol 531, 509-514 | [Abstract/Full Text] |
Acknowledgements
This work was supported by National Institutes of Health (NIH) grant HL64912. We wish to thank Dr Steven A. Shea for his careful critique of the manuscript.
This article has been cited by other articles:
|
|
 |
|
  A. Tadjalli, J. Duffin, Y. M. Li, H. Hong, and J. Peever Inspiratory activation is not required for episodic hypoxia-induced respiratory long-term facilitation in postnatal rats J. Physiol., December 1, 2007; 585(2): 593 - 606. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. P. Harris, A. Balasubramaniam, M. S. Badr, and J. H. Mateika Long-term facilitation of ventilation and genioglossus muscle activity is evident in the presence of elevated levels of carbon dioxide in awake humans Am J Physiol Regulatory Integrative Comp Physiol, October 1, 2006; 291(4): R1111 - R1119. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. McGuire, Y. Zhang, D. P. White, and L. Ling Phrenic long-term facilitation requires NMDA receptors in the phrenic motonucleus in rats J. Physiol., September 1, 2005; 567(2): 599 - 611. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. D. Fuller Episodic hypoxia induces long-term facilitation of neural drive to tongue protrudor and retractor muscles J Appl Physiol, May 1, 2005; 98(5): 1761 - 1767. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. McGuire, Y. Zhang, D. P. White, and L. Ling Serotonin receptor subtypes required for ventilatory long-term facilitation and its enhancement after chronic intermittent hypoxia in awake rats Am J Physiol Regulatory Integrative Comp Physiol, February 1, 2004; 286(2): R334 - R341. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Phr, %baseline). B, phrenic burst frequency (