Main

The presence of synchronized bursts of activity in pre- and postganglionic nerve traffic is a well known feature of spontaneous sympathetic discharges in mature mammals. This phenomenon was described first by Adrian et al.(1) as a “grouping” of waves of activity during the respiratory and cardiac cycles. Respiratory modulation of sympathetic activity is likely due to complex synaptic relationships between brain stem neurons involved in respiratory rhythmogenesis and those generating sympathetic activity(2–4). The cardiac-related component, on the other hand, is a product of baroreceptor afferent activity which entrains an endogenous 2-6 Hz rhythm in the activity of brain stem sympathetic neurons(4–6). Entrainment of the 2-6 Hz rhythm occurs very early in postnatal life (<1 d old) because spectral analyses of preganglionic sympathetic and blood pressure signals exhibit a high degree of coherence (i.e. correlation)(7, 8). However, it is not known whether the 2-6 Hz rhythm appears at the same time in pre and postganglionic nerve discharges, or whether this rhythm appears first in preganglionic activity, and later in postganglionic activity. Indeed, there is evidence suggesting a lag in the appearance of postganglionic activity, possibly due to transmission failure in sympathetic ganglia(9–11). A delay in the appearance of the 2-6 Hz oscillation, or of respiration-related activity may have important physiologic consequences. For example, the target organs of postganglionic innervation may not receive functionally appropriate sequences of activity.

The objective of the present investigation was to determine whether the appearance of cardiac- and respiratory-modulated discharges is developmentally delayed in postganglionic nerves. We selected the peroneal nerve for study because its sympathetic activity has been used as an index of autonomic function in adult humans during normal physiologic states (e.g. exercise, sleep) as well as pathophysiologic states (e.g. obstructive apnea, heart failure)(12–19). Furthermore, there is little information regarding developmental changes of postganglionic nerve activities during hypoxic stimulation.

METHODS

This study was approved by the institutional animal use committee. All experimental procedures are in compliance with federal and state regulations as well as the “Guide for the Use and Care of Laboratory Animals” approved by the council of the American Physiologic Society. The swine used in this study were born and nurtured by their respective sows in the Division of Laboratory Animal Resources of the Health Science Center at Brooklyn (SUNY). Experiments were performed on 20 piglets ranging in age from <1 to 40 d old and in weight from 1.11 to 8.45 kg; regression analyses revealed a highly significant correlation coefficient (r = 0.66, p < 0.01) between age and weight.

General surgical procedures. Piglets were initially anesthetized with an injection of the steroidal anesthetic Saffan (Vet Drug Ltd., York, UK) (12 mg/kg, intramuscularly). Next, an external jugular vein was cannulated for continuous infusion of Saffan (12-18 mg/kg/hr). The trachea, femoral vein, and artery were then cannulated for artificial ventilation, infusion of 5.0% glucose in 0.9% saline (2 mL/kg/h), and for measurement of arterial blood gases and pH, respectively. Tracheal and arterial cannulae were attached to pressure transducers for measurements of intratracheal and arterial pressures, respectively. At this time both cervical vagus nerves were isolated and loose ligatures placed around them for later transections. Animals were then paralyzed with decamethonium bromide and artificially ventilated with 100% O2. A rectal probe measured core temperature which was maintained at 39 ± 0.5°C using a servo-controlled heating blanket. Arterial blood gas tensions and pH were measured at 20-min intervals throughout each experiment. Pco2 was maintained in the normal range by adjustments of ventilatory rate and volume. In the event of metabolic acidosis, sodium bicarbonate was administered by slow i.v. infusion.

Preparation of nerves for recording. With the piglet in the supine position, the left C5 phrenic root was dissected from surrounding tissues, cut and immersed in a pool of warm mineral oil. Next, the gastrocnemius and peroneus longus muscles were retracted to visualize the left common peroneal nerve near the posterior aspect of the knee at the apex of the head of the fibula. At this location, the peroneal nerve traverses the lateral aspect of the leg, turns anteriorly, and descends below several muscle groups. The peroneus brevis, anterior tibialis, and external digitorum longus were removed to identify that part of the nerve where it divided into smaller fascicles. Each fascicle was cut and immersed in warm mineral oil.

Monophasic recordings were obtained from the cut, crushed central ends of the C5 phrenic root and peroneal fascicles using bipolar platinum hook electrodes. The bandpass for phrenic and peroneal recordings was 10 Hz to 3 kHz and 3 Hz to 1 kHz, respectively. In three piglets, the ganglionic blocker, trimethaphan camsylate (Arfonad; Roche, Nutley, NJ) was diluted in saline and administered by slow i.v. infusion (15 mg/kg) to demonstrate that peroneal nerve activity contained postganglionic discharges.

Experimental protocols. At the conclusion of all surgery the Saffan infusion rate was lowered to 6 mg/kg/h. At least 1 h was allowed to pass to permit physiologic stabilization. Recordings were then taken of nerve and pressure signals during a control hyperoxia period of 5-min duration, followed by a test period of hypoxia (10% O2, balance nitrogen) for 5-6 min, and then a return to hyperoxia. This sequence of control and test trials was presented before and after bilateral vagotomy. Arterial blood gas tensions and pH were measured during hyperoxia, toward the end of the hypoxic exposure, and after a suitable recovery period.

Signal processing and analysis. All neural and pressure signals were stored on video cassettes for off-line computer analyses. The manner of signal conditioning as well as the types of analyses have been described in detail elsewhere(7, 8, 20, 21). Briefly, both low pass filtered (40 Hz) and integrated (100 ms time constant) phrenic and peroneal signals were acquired along with the low pass filtered(40 Hz) blood pressure signal using a sampling rate of 128 Hz. Three types of analysis were carried out. 1) Power and coherence spectral analyses were performed to demonstrate the coupling of the peroneal discharge to the cardiac cycle. A FFT routine was applied to the low pass filtered peroneal and blood pressure signals in eighty contiguous 4 s epochs; the number of FFT points (512) was set to equal the duration of a single epoch. The mean and any other trend were then subtracted from each data point; next, each point was multiplied by a Hamming function. Power spectral estimates were then calculated for each epoch and the final power spectrum was obtained by ensemble averaging of all individual epochs. To detect correlated(i.e. coherent) frequencies between peroneal and blood pressure signals, the cross-spectrum was first calculated followed by computation of the coherence spectrum i.e., the ratio of the squared magnitude of the average cross-power estimate at each frequency to the product of the individual spectral estimates at the same frequency(22). 2) Cardiac-related histograms were constructed because coherence could also result from cardiac-related artifacts mechanically coupled to nerve activity. These signal average histograms of low pass filtered peroneal activity were triggered by pulses placed at the peak of systole and represented the average peroneal activity in 1000 cardiac cycles.3) Inspiratory-triggered histograms were constructed from integrated nerve signals and were triggered by marks corresponding to the onsets of phrenic discharges. These marks were generated by a software algorithm and were used as synchronizing events to compute histograms of average nerve activity in 20-30 central inspiratory phases.

Statistical analyses. To determine whether coherence estimates were statistically significant, a test for significance of nonzero coherence(23) was carried out using the values of coherence for peaks appearing in these spectra. This procedure establishes the upper 95% point of the confidence interval for the coherence estimate at any selected frequency. Estimates above the 95% point are considered to be statistically significant (p < 0.05) and indicated that the two signals were nonrandomly related, whereas those below the 95% point were due to chance. In the present series of experiments, a coherence value >0.1 was found to exceed the 95% point of the confidence interval and was considered to be significant.

RESULTS

Spontaneous discharges. Simultaneous recordings of C5 phrenic and postganglionic peroneal activities were obtained from 20 piglets with intact vagi; in four animals, recordings were obtained from a second and third peroneal branch as well. A consistent feature of peroneal discharges in these experiments was that both spontaneous and hypoxic discharges were relatively unchanged after vagotomy. As well, spectral characteristics i.e., locations and amplitudes of primary peaks, were not altered by vagotomy. Consequently, the data presented herein will be those obtained after vagotomy. Peroneal activity was identified as being sympathetic by the presence of pulse synchronous activity, and, in a few cases, by the observation of spontaneous respiration-related discharges. Additional confirmation of the postganglionic origin of peroneal activity was obtained from three animals which received i.v. infusions of the ganglionic blocker Arfonad. Figure 1 shows that the prominent expiratory burst in integrated peroneal discharge (A, Int Per) was abolished by ganglionic blockade(B). This persisted for at least 30 min (C).

Figure 1
figure 1

Effect of ganglionic blockade on postganglionic peroneal discharges. Digital traces (sampling rate 128 Hz) of integrated (100 ms time constant) phrenic (Int PHR) and postganglionic peroneal(Int PER) discharges before (A), during (B), and after (C) ganglionic blockade by i.v. infusion of Arfonad. Thehorizontal line beneath each set of traces is a time mark and is equal to 1 s.

Full size image

Baroreceptor modulation during hyperoxia and hypoxia. The role of baroreceptor afferents in modulating peroneal discharges was evaluated by power and coherence spectral analyses. To qualify for such processing, records of peroneal activities were required to be relatively free of artifacts which were mainly due to artificial ventilation. Of 24 nerves in 20 piglets, the discharges of 16 nerves in 10 piglets (ranging in age from 3 to 37 d) were fee of artifact and were processed accordingly. Spectral analyses of peroneal and blood pressure signals in these animals revealed prominent peaks over the range of cardiac frequencies (3.25-5.0 Hz). For example, the autospectra ofFigure 2A were computed from signals obtained during hyperoxia and showed: 1) a large primary peak at 5 Hz in the blood pressure autospectrum (top) and 2) a peak at 5 Hz in the peroneal autospectrum (middle). Both autospectra contain a peak at 1 Hz (due to ventilatory rate) and at 10 Hz which probably represents the second harmonic of the 5 Hz peak. The coherence spectrum of Figure 2A (bottom) shows a statistically significant coherence estimate of 0.9 at 5 Hz which is suggestive of a strong linear relationship between baroreceptor afferent inputs and the sympathetic rhythm generating networks in this animal. That the baroreceptor-related discharge in peroneal activity was a relatively constant feature is demonstrated by the presence of a large summation waves to the left and right of the vertical line at time zero (the peak of systole) in the average histogram of Figure 2B. It is important to note that these waves do not straddle the vertical line at time zero, thereby demonstrating their nonartifactual origin.

Figure 2
figure 2

Evidence of baroreceptor modulation of postganglionic peroneal discharge. (A) Autospectra of blood pressure(top) and peroneal (middle) signals demonstrating peaks at the cardiac frequency, in this case at 5 Hz. This peak is highly coherent as indicated by a 0.9 coherence estimate in the coherence spectrum(bottom). Bind width in all spectral plots is 0.25 Hz. (B) Average triggered by pulses positioned at the peak of the blood pressure wave shows that the baroreceptor-related peroneal discharge occurs in diastole and is nonartifactual. Horizontal bar at bottom left is a time mark equal to 100 ms.

Full size image

During hyperoxia, 11/16 nerve-blood pressure coherence spectra contained significant coherence estimates at the cardiac frequency; during hypoxia, 14/16 nerve-blood pressure coherence estimates were significanti.e., the five cases, which lacked coherence in hyperoxia, became significantly coherent, whereas two cases of significant coherences in hyperoxia became nonsignificant during hypoxia. The descriptive statistics for coherence values will be given as the median and the semi-interquartile range(25th and 75th percentiles) because there was a substantial difference between the mean and median (0.3 versus 0.2, respectively) in the hyperoxia data set; the 25th and 75th percentiles were 0.13 and 0.35, respectively. During hypoxia, the median value increased to 0.39 (0.24, 0.55; 25th and 75th percentiles, respectively). Such increases of coherence in hypoxia were found to be highly significant (p < 0.0007, Wilcoxon matched-pairs signed ranks test)(24). That this result was solely due to hypoxia and not to a concomitant state of hypercapnia and/or acidosis is verified by the data in Table 1 which shows that only the values of Po2 were altered in a statistically significant manner(paired t test, p < 0.05).

Table 1 Summary of pH and arterial blood gas measurements during normocapnic hyperoxia (control) and normocapnic hypoxia (test) epochs
Full size table

Respiratory modulation during hyperoxia and hypoxia.Table 2 summarizes the types of discharge patterns observed during hyperoxia and hypoxia as well as the long term (min) responses to hypoxic stimulation. Spontaneous respiratory activity was observed only in peroneal recordings taken from animals ≥20 d old (5/9 cases); of the remaining 4/9 cases, 3 exhibited respiration-related activity with hypoxic stimulation (Table 2). In many <20-d-old animals, respiration-related discharges were elicited by hypoxic stimulation (7/11, seeTable 2).

Table 2 Summary description of peroneal nerves with respiratory discharges, their type of discharge pattern (I, inspiratory; eE, early expiratory; E, expiratory), and the overall pattern of response to hypoxia
Full size table

Respiration-related discharge patterns were classified according to the time in the central respiratory phases when the maximum amplitude of activity was observed. For example, Figure 3A (top trace) represents a pattern of peroneal activity classified as inspiratory: the onset of peroneal activity occurs just before that of phrenic discharge(indicated by the dashed vertical line) and maximum amplitude occurs during inspiration. In Figure 3, B and C traces were taken from different animals and are representative of early expiratory and expiratory discharge patterns, respectively. As indicated inTable 2, the most common type of peroneal activity observed in the different conditions of these experiments was early expiratory(16/20 piglets); purely inspiratory and expiratory discharge patterns were relatively rare, and were observed with the same frequency i.e., twice each in the remaining 4/20 piglets.

Figure 3
figure 3

Respiration-related discharge patterns in postganglionic peroneal activity. Averages of integrated nerve signals in three different animals: (A) inspiratory pattern, (B) early expiratory pattern, and (C) expiratory pattern.

Full size image

An unexpected finding in these experiments was that the peroneal activity of many animals continued to exhibit an immature (biphasic) response to hypoxia: a transient early 1-2-min period of increase of discharge amplitude, followed by a long lasting decrease of amplitude toward or below that of hyperoxic activity, and sometimes cessation of respiration-related activity. Despite the fact that phrenic nerve activity exhibited sustained increases of discharge amplitude in all animals, a peroneal biphasic pattern was as likely to be observed as a sustained response in animals <20 d old(Table 2: 4/11 versus 3/11 cases, respectively).

DISCUSSION

The results of the present investigation do not support the hypothesis that the appearance of a 2-6 Hz rhythm in postganglionic sympathetic discharges is linked to a developmental process. Indeed, our findings indicate that postganglionic sympathetic activity contains a 2-6 Hz oscillation which is synchronized to the cardiac cycle in piglets as early as 3 d old. Moreover, the fact that values of coherence were significantly increased as a result of hypoxic stimulation demonstrates a functional chemoreceptor reflex.

Unlike the results obtained with baroreceptors, respiration-related modulation of peroneal discharges was not apparent in spontaneous activities until the third postnatal week. Even then about half of the animals did not have a respiration-related discharge. Although such activity was readily elicited by hypoxic stimulation, peroneal activity in many <20-d-old animals had an immature biphasic response pattern, even though phrenic activity exhibited a mature response pattern. These qualitative findings suggest that the development of synaptic relationships between sympathetic and respiratory neural networks continues well into the neonatal period. Another possible explanation for the absence of respiratory activity during normoxia, as well as the failure to maintain respiratory activity during hypoxic stimulation, may be transmission failure at ganglionic synapses, possibly due to insufficient amounts of neurotransmitters. Nevertheless, the presence of respiration-related peroneal activity in hypoxic epochs does not appear to be of significance because values of coherence in blood pressure-peroneal coherence spectra increased in most cases regardless of the respiratory response pattern. The finding of decreased coherence values during hypoxia in two animals may represent a hypoxia mediated uncoupling of sympathetic networks from baroreceptor entrainment. Such a loss of a normal physiologic control mechanism may actually underlie some pathophysiologic states of autonomic function. Thus, although these data do not suggest a functional role for respiration-related postganglionic discharge, such modulation may certainly be necessary for the expression of functionally appropriate activity in response to other types of stimulation.

Peroneal discharges in young piglets manifested many similarities to those in humans. In both species, peroneal activity is related to both the cardiac cycle and to the expiratory phase of the respiratory cycle(15, 25–30). Furthermore, like the piglet, peroneal activity in humans is unaffected by vagal afferent activity during eupnea(28). However, there may be differences in the discharge features of human and piglet peroneal activities. One possible difference is whether peroneal cardiac-related activity has a probabilistic or deterministic relation to the cardiac cycle. Single or pauci-fiber recordings of peroneal activity in humans using microelectrodes suggest a probabilistic relationship because action potentials are discharged in only a fraction of cardiac cycles(16, 30, 31). On the other hand, our recordings of large fascicles branching from the common peroneal nerve argue for a deterministic relationship because activity occurred in most cardiac cycles. Such a difference in discharge characteristics is probably due to a sampling bias inherent in human studies where data are provided by relatively small numbers of recorded axons. The benefit of an animal preparation is the increased sample size obtained by recordings of large numbers of axons as was done in the present study. Thus, differences between humans and piglets in degree of cardiac modulation of peroneal activity are attributable to different recording techniques, and do not deter from the similarities of discharge features in the two species. In conclusion, these results indicate that the pig may be an appropriate species to model disorders of central regulation of sympathetic discharge, for example, abnormalities of sympathetic outflows which may occur in sudden infant death syndrome(32–34).