Impact of lung inflation cycle frequency on rat muscle and skin sympathetic activity recorded using suction electrodes.

Microneurography has been used in humans to study sympathetic activity supplying targets within skeletal muscle and skin. Comparable animal studies are relatively few, probably due to the technical demands of traditional fibre picking techniques. Here we apply a simple suction electrode technique to record cutaneous (CVC) and muscle (MVC) vasoconstrictor activities and describe and investigate the basis of the frequency dependence of lung inflation related modulation. Hindlimb MVC and CVC activities were recorded concurrently. The magnitude of MVC and CVC activities at the lung inflation cycle frequency was significantly less at 2.0 Hz than at lung inflation cycle frequencies < or =1.0 Hz. As lung inflation cycle frequency was increased the coherence between lung inflation cycle or BP and MVC or CVC waveforms decreased. Consistent with the hypothesis that much of the coherence between lung inflation cycle and nerve activity waveforms is secondary to oscillating baroreceptor activity attributable to BP waves, partialization with the BP waveform significantly decreased the coherence between lung inflation cycle and nerve waveforms, and there was an absence of coherence between these waveforms following sinus and aortic denervation. Our data extend findings from other laboratories and establish the value of a suction electrode technique for recording MVC and CVC activities. Furthermore, our observations describe the rates of positive pressure ventilation that avoid strong and regular gating of sympathetic activity.

Introduction

Over the last four decades microneurography has been used to study human sympathetic activity supplying targets within skeletal muscle and skin (Hagbarth and Vallbo, 1968; Vallbo et al., 2004; Montano et al., 2009). Sympathetic activity supplying skeletal muscle, considered to be mainly vasomotor, is elevated in many disease states (see Fisher et al., 2009; Wallin and Charkoudian, 2007). There is a relative paucity of equivalent animal studies with most focussing on activity recorded from visceral nerves (see Montano et al., 2009). This probably reflects, at least in part, the higher technical demands of the fibre picking technique (used to record sympathetic activity to muscle and skin) over the whole visceral nerve technique (e.g., Dampney and McAllen, 1988; Futuro-Neto and Coote, 1982a,b; Jänig and Häbler, 2003; Yusof and Coote, 1988a,b). Huang and Gilbey (2005) acquired data on sympathetic activity to muscle and skin in the anaesthetised rat using a suction electrode recording technique, which is technically less demanding than fibre picking. This methodological innovation has the potential to increase muscle and skin sympathetic activity data flow from animal models of diseases (see Gilbey and Huang, 2006).

Breathing cycle-related sympathetic discharges are found in rats and other species including humans (Häbler et al., 1994b). Such modulation emerges from CNS rhythm generating networks and from the influence of peripheral feedback on CNS processes (e.g., Adrian et al., 1932; Badra et al., 2001; Barman and Gebber, 1976; Gregor et al., 1977; Häbler et al., 1996; Jänig and Häbler, 2003; Seals et al., 1993; St Croix et al., 1999); this paper focusses on the latter.

In both spontaneously breathing and positive pressure ventilated animals and humans the magnitude of the arterial blood pressure waves associated with the lung inflation cycle decreases as the frequency of the lung inflation cycle increases (Daly, 1986; Kuo et al., 1996). Such blood pressure waves lead to oscillating arterial baroreceptor activity (Häbler et al., 1993, 1994b, 1996). The oscillating baroreceptor activity may produce rhythmic lung inflation related sympathetic activity through its influence on tonic activity and/or by entraining an autonomous sympathetic rhythm (see Gilbey, 2004). In a preliminary study in rats we noted that sympathetic activities regulating hindlimb muscle and cutaneous circulations could have lung inflation related discharges, consistent with observations on cervical sympathetic preganglionic neuronal activity (Häbler et al., 1996). However, Häbler et al. (1993) failed to demonstrate routinely lung inflation related discharges in hindlimb sympathetic activities, which led them to consider the possible factors underlying such inconsistencies (Häbler et al., 1992, 1996). The variable modulation may either indicate differential control of sympathetic activities regulating particular targets or reflect different experimental protocols. The differences have been attributed to the latter (Häbler et al., 1994b).

This study maps the relationship between the frequency of positive pressure ventilation, the magnitude of arterial blood pressure waves and lung inflation related rhythms in simultaneously recorded cutaneous (CVC) and muscle (MVC) vasoconstrictor sympathetic activities and probes the mechanisms underlying the associated sympathetic modulation. Experiments were conducted in the absence of rhythmic central respiratory drive. Population sympathetic activity was recorded from a gastrocnemius nerve and the ipsilateral tibial nerve at a point distal to the main muscle branches (plantar branch). These served as indices of MVC and CVC sympathetic activities, respectively (Huang and Gilbey, 2005). Some of the data have been published as an abstract (Huang et al., 2004).

Materials and methods

Anaesthesia and animal maintenance

Experiments were performed on 9 male Sprague Dawley rats (290–340 g) under Project and Personal licences issued by the Home Office and local ethics committee approval. Anaesthesia was induced with sodium pentobarbitone (60 mg kg− 1 i.p.) and supplemented with α-chloralose (5–10 mg i.v.) as required. The depth of anaesthesia was monitored continuously and judged using a number of criteria: i) stability of heart rate, blood pressure, frequency and depth of spontaneous breathing (when present); (ii) pupil size; and (iii) palpebral and paw pinch reflexes. At the end of experiments animals were killed with an overdose of sodium pentobarbitone (i.v.).

The right femoral artery and vein were cannulated to monitor arterial blood pressure (BP) and to administer drugs, respectively. The trachea was cannulated to enable positive pressure ventilation (Harvard 683, volume controlled ventilator, Harvard Apparatus, Holliston, Massachusetts, USA). Tracheal pressure was used as a surrogate for the lung inflation cycle waveform. Animals were ventilated at various frequencies (see “Protocol”) using a pump tidal volume set at approximately 2 ml (O2-enriched room air, inspiratory time/expiratory time ratio of 1). Arterial blood samples (75 µl) were taken after each data collection period to monitor pH (range 7.35–7.55), blood gases (range PaCO2, 15–35 mm Hg; PaO2, 200–400 mm Hg). Rectal temperature was monitored (control temperature 36.0 ± 0.5 °C). The bladder was cannulated to allow free passage of urine.

Recordings of diaphragmatic EMG and nerve activity

Details have been provided previously (Huang and Gilbey, 2005). Briefly, the absence of rhythmic diaphragmatic EMG, appropriate blood gases and pH was used to assess central apnoea (hypocapnic/hyperoxic). Using a suction electrode recording technique population sympathetic activity was recorded concurrently from a gastronemius nerve and the ipsilateral tibial nerve at a point distal to the main muscle branches — “plantar branch”: these served as indices of MVC and CVC activities, respectively. The sympathetic nature of discharges was determined by their abolition following ganglionic blockade. Levels of activity following administration of antagonists of ganglionic transmission were comparable to those seen following death (intrinsic noise in recording system).

Data analysis

The magnitude of rhythmic components and the degree of linear correlation between waveforms were assessed using frequency domain analysis (300 s data sets sampled at 100 Hz). Spike 2 data files were converted into text files and analysed using Matlab computer software (Maths Works). The spectral resolution was 0.049 Hz/bin. Autospectra were generated as described previously (Huang and Gilbey, 2005). Those of nerve activities were plotted as percentage total power whereas the BP waveform was plotted as absolute power.

Ordinary and partial coherence spectra (Bendat and Piersol, 1999) were generated. Ordinary coherence was used to reveal the linear correlation between signals at a particular frequency; estimated by normalizing the cross spectrum between two nerve activities. A coherence value of 1 indicates perfect correlation, whereas a value < 1 indicates the presence of uncorrelated linear, and/or non-linearly related signals or extraneous noise. A 95% confidence level threshold value of 0.1 for non-zero coherence was adopted (Smith and Gilbey, 2000). Partial coherence is the residual coherence between two waveforms when the effect of other input(s) is removed. A partial coherence value of 1 indicates that neither of the two waveforms can be predicted from the other (partializing) input. Conversely a partial coherence value of 0 indicates that the coherence between the waveforms can be explained totally by the partializing waveform.

The power density and coherence at a particular lung inflation cycle frequency were assessed (using Matlab software) using a cursor positioned by an operator. The peak value was then read.

Preparations

All animals underwent bilateral cervical vagotomy (n = 9), three with sino-aortic denervation (SAD). SAD was achieved and assessed as described by Huang and Gilbey (2005). All experiments were conducted with animals in central apnoea to remove the confounding influences of central respiratory drive (see Chang et al., 2000).

Protocol

Lung inflation cycle frequency was adjusted to 0.8, 0.9, 1.0, 1.25, 1.5 or 2.00 Hz. Data were collected at each lung inflation cycle frequency. No particular sequence of lung inflation cycle frequency was used. Data at each lung inflation cycle frequency were collected over at least 10 min.

Statistical analysis

All statistical analyses were computed using GraphPad Prism (version 4.00 for Windows, GraphPad Software, San Diego, CA, USA). Results are expressed as mean ± S.E.M. One way and two way ANOVA (Bonferroni post test) were applied as indicated. P values < 0.05 were considered significant.

Results

Cardiac-related activity

As reported by Huang and Gilbey (2005) cardiac-related activity was present in the activity of MVCs but not CVCs of animals with sinus and aortic nerves intact (SA intact). Cardiac-related activity was absent in SAD animals (data not shown).

Arterial blood pressure waves and lung inflation cycle frequency

In SA intact and SAD animals lung inflation cycle and BP waveforms were highly correlated over the range of lung inflation cycle frequencies (coherence values 0.9–1.0). Population data revealed, as evidenced by the analysis of autospectra, that as lung inflation cycle frequency was increased the magnitude of the BP waveform at lung inflation cycle frequency decreased in both SA intact and SAD groups (Fig. 1A and B). Sample data from the SA intact group illustrate that BP waves occurred at lung inflation cycle frequency when lung inflation cycle frequency was 0.9 Hz, (Fig. 1C). This was confirmed, by analysis of 300 s data sets, as a peak in autospectra at 0.9 Hz (Fig. 1C). In contrast, it can be seen that at the highest lung inflation cycle frequency (2.0 Hz) BP waves at lung inflation cycle frequency were not apparent (Fig. 1D).

Fig. 1

BP waves and nerve activities at lung inflation cycle frequency. A, Absolute power of BP waves, in SA intact animals, across the range of lung inflation cycle frequencies. Note that as lung inflation frequency is increased the magnitude (absolute power) at lung inflation frequency decreases; one-way ANOVA: “star” = P < 0.05 against values at all other lung inflation cycle frequencies; “circle” = P < 0.05 against 2.0 Hz; “diamond” = P < 0.05 against 1.5 and 2.0 Hz. Numbers above columns indicate the number of animals. B, Absolute power of BP waves at lung inflation cycle frequency, in SAD animals, across the range of lung inflation cycle frequencies. Numbers above columns indicate the number of animals. As for the SA intact group the magnitude of BP waves tended to decrease as the lung inflation cycle frequency was increased. C and D, Real time recordings and autospectra, from an SA intact rat, of lung inflation cycle, BP and rectified and smoothed MVC and CVC nerve activity waveforms during lung inflation cycle frequencies of 0.9 and 2.0 Hz, respectively. In the time domain it can be seen that at 0.9 Hz there are BP waves and “bursts” of nerve activity at 0.9 Hz. This feature of all waveforms can be seen in the autospectra (left column) where power is concentrated at 0.9 Hz. In contrast at a lung inflation frequency of 2.0 Hz neither BP waves nor “bursts” of nerve activity are observed at 2.0 Hz. Dot above peak in autospectrum of CVC identifies the T-rhythm (see text for details). E and F, Group data from SA intact animals showing the magnitude (percentage power) of MVC (E) and CVC (F) activities at lung inflation cycle frequency over the range of lung inflation cycle frequencies. Note that as lung inflation frequency increases the percentage power decreases; one-way ANOVA; “circle” = P < 0.05 against 2.0 Hz; “diamond” = P < 0.05 against 1.25, 1.50 and 2.0 Hz. Numbers above columns indicate the number of animals.

Sympathetic activity and lung inflation cycle frequency

Group data demonstrate that in the SA intact animals the magnitude of MVC and CVC activities at the lung inflation cycle frequency was significantly less at 2.0 Hz than their values at lung inflation cycle frequencies ≤ 1.0 Hz (Fig. 1E and F). Sample data show lung inflation cycle-related rhythms in MVC and CVC activities when lung inflation cycle frequency was 0.9 Hz (Fig. 1C). In contrast, when lung inflation cycle frequency was 2.0 Hz, a lung inflation cycle-related rhythm was not apparent (Fig. 1D). Note the presence of a “free run” T-rhythm in CVC activity (see Huang and Gilbey, 2005). At lower lung inflation cycle frequencies the lung inflation cycle-related peaks in CVC activity are due to entrainment of this rhythm (see Chang et al., 2000). In SAD rats lung inflation related rhythms were not seen at any lung inflation cycle frequency (data not shown).

In summary, although pump tidal volume was held constant, in SA intact animals the magnitude of MVC and CVC activities at lung inflation cycle frequency tended to decrease as lung inflation cycle frequency was increased. Although lung inflation related blood pressure oscillations remained in SAD animals, lung inflation related MVC and CVC activities were absent at all lung inflation cycle frequencies.

Blood pressure waves and lung inflation related sympathetic activity

As evidenced above both the magnitude of BP waves and sympathetic activity at lung inflation cycle frequency decreased as frequency was increased. In the SA intact group, as lung inflation cycle frequency was increased the strength of the linear correlation (coherence) between lung inflation cycle and both MVC and CVC decreased (Fig. 2A and B) as did that for BP to nerve activity coherence (Fig. 2C and D); N.B., coherence (correlation) is a measure independent of the magnitude of the two variables that are coupled. These relationships were not observed in the SAD group where coherence values across the range of lung inflation cycle frequencies were close to 0.1. In line with the effect of SAD on coherence the mathematical removal of the component of the linear coupling between lung inflation cycle and nerve activities that can be explained by the BP waves generated partial coherence values that were significantly less than the respective ordinary coherence values (Fig. 2A and B). This result is consistent with the idea that much of the coherence between lung inflation cycle and nerve activity waveforms is secondary to the oscillations of baroreceptor activity attributable to BP waves generated by the lung inflation cycle.

Fig. 2

Coherence between lung inflation cycle and nerve activity waveforms and the effect of partialization using the BP waveform. A and B, Ordinary coherence (open bars) of the lung inflation cycle waveform with both MVC (LIC-MVC) and CVC (LIC-CVC) activities was within the range 0.9–1.0 at lung inflation cycle frequencies 0.8–1.0 Hz. Thereafter ordinary coherence decreased (see Fig. 1C and D). Partialization with the BP waveform (LIC-MVC/BP and LIC-CVC/BP) significantly decreased coherence values across all lung inflation cycle frequencies (dot filled bars), which is consistent with the finding that in SAD animals ordinary LIC-MVC and LIC-CVC coherences were reduced to approximately 0.1; the threshold for significance. Two-way ANOVA; “star” = P < 0.05 ordinary coherence values compared to partial coherence values; “circle” = P < 0.05 ordinary coherence at designated lung inflation cycle frequency compared to 2.0 Hz; “ diamond” = P < 0.05 at designated lung inflation cycle frequency compared to 1.5 and 2.0 Hz. C and D, Ordinary coherence of the BP waveform with both MVC (BP-MVC) and CVC (BP-CVC) activities was within the range 0.9–1.0 at lung inflation cycle frequencies 0.8–1.0 Hz. Thereafter ordinary coherence decreased. Key to symbols as in A and B and one-way ANOVA.

Discussion

Using a suction electrode recording technique we have demonstrated that rat hindlimb MVC and CVC activities can be modulated by lung inflation related inputs. We found that in SA intact animals the magnitude of the sympathetic rhythm at lung inflation cycle frequency, recorded simultaneously from MVCs and CVCs, decreased as the lung inflation cycle frequency was increased, as did the magnitude of the BP waves. No sympathetic rhythms at lung inflation cycle frequency were observed in SAD animals. Therefore the magnitude of lung inflation related modulation of MVC and CVC activities is dependent upon the frequency of ventilation and intact sinus and aortic nerves. This is consistent with, and extends the findings of other studies (see Häbler et al., 1994b) and therefore establishes the suction electrode recording technique as a viable alternative to conventional fibre picking. The arterial baroreceptor dependent lung inflation related modulation if present is weak at frequencies of positive pressure ventilation similar to the frequency of spontaneous breathing in conscious animals, consistent with observations in conscious rats where: i) SAD was not observed to alter breathing cycle-related rhythms in renal nerve activity (Julien et al., 2003); ii) breathing cycle-related rhythms were not a regular feature of renal sympathetic nerve activity (Brown et al., 1994). However, in a study on conscious rats Kunitake and Kannan (2000) observed an arterial baroreceptor mediated component of breathing cycle-related activity.

The magnitude of lung inflation related rhythms in MVC and CVC activities at lung inflation cycle frequency was sustained at lung inflation cycle frequencies ≤ 1 Hz, which coincided with the robust BP waves. Therefore in experiments where animals are artificially ventilated, in an attempt to clamp blood gases, the lung inflation cycle frequency and tidal volume used may substantially influence the responses of sympathetic outflows to various inputs. This can be observed anecdotally by examining the influence of lung inflation cycle frequency on the burst pattern of MVC activity and the entrainment of the CVC T-rhythm (see Fig. 1C and D): it can be seen that its influence is greatest at lower frequencies.

The data demonstrate that with constant pump tidal volume (inspiratory time/expiratory time ratio fixed at 1), as lung inflation cycle frequency was increased the amplitude of BP waves at lung inflation cycle frequency decreased in both SAD and SA intact animals. This lung inflation related influence on BP is similar to that observed in other studies (Kuo et al., 1996; Yang and Kuo, 1999).

The linear correlation between lung inflation cycle and nerve activities decreased as lung inflation cycle frequency was increased, similar to observations made in a previous study (rats with intact vagi) in population recordings of cutaneous vasoconstrictor neurones innervating the tail (thermoregulatory) circulation (Chang et al., 2000). Following partialization with the BP waveform, group data demonstrated that the correlation between lung inflation cycle signal and nerve activities was significantly reduced at all lung inflation cycle frequencies. Therefore in agreement with the findings from other laboratories (see Häbler et al., 1996) it may be concluded that in this preparation it is the BP waves, secondary to the lung inflation cycle, that predominantly provide the coupling between lung inflation cycle and nerve activities at lung inflation cycle frequency and hence the lung inflation cycle-related rhythms in MVC and CVC activities. Sino-aortic nerves are involved in this coupling as it is disrupted by their section. Clearly other factors can be involved under some conditions and when recording from other outflows (Collins and Gilbey, 2003; Julien et al., 2003). In humans there is evidence that cardiopulmonary baroreceptor and pulmonary vagal afferents are also involved in MVC respiratory rhythm generation (Wallin and Charkoudian, 2007).

The involvement of the arterial baroreceptors in the lung inflation cycle-related modulation of CVC in addition to MVC activity may at first appear questionable, as the probability of CVC discharge does not vary strongly with the cardiac cycle, whereas MVC activity does (Häbler et al., 1994a). However, as demonstrated by other workers, CVC activity is influenced by lowering or raising arterial blood pressure and by aortic nerve stimulation (Häbler et al., 1994a; Johnson and Gilbey, 1998).

In conclusion, our data demonstrate clearly that as frequency of positive pressure ventilation is increased the magnitude of the lung inflation related hindlimb MVC and CVC activities declines as does the magnitude of BP waves. This modulation is not apparent following SAD. Although CVC activity lacks robust cardiac-related modulation (see Huang and Gilbey, 2005) our observations indicate that the parameters of positive pressure ventilation influence the lung inflation related modulation of both MVC and CVC activities, albeit through separate mechanisms (phasic inhibition and entrainment of the T-rhythm, respectively) which are primarily dependent upon SA afferent inputs. In humans most studies have failed to identify an influence of arterial baroreceptors on skin sympathetic nerve activity (see Wallin and Charkoudian, 2007). Further studies are required to assess whether in rat models of human diseases MVC activity changes in a similar manner to that seen in the human condition. If such is the case it will be possible to use rat models to identify potential factors that underlie disease associated aberrant muscle sympathetic activity in humans.