Responses evoked in single sympathetic nerve fibres of the rat tail artery by systemic hypoxia are dependent on core temperature

Abstract

No direct evidence exists of the changes evoked by systemic hypoxia in sympathetic nerves to the rat cutaneous circulation, and of the concomitant changes in cutaneous blood flow. Here we investigated responses evoked by two levels of systemic hypoxia (12% and 8% inspired O₂) in single sympathetic units supplying tail caudal ventral artery (CVA) in spontaneously breathing anaesthetized rats, whilst simultaneously recording tail blood flow and vascular resistance (TVR) from the CVA, under conditions of modest hypothermia and hyperthermia. During modest hypothermia and normoxia, TVR was high and CVA unit activity was present, with marked respiratory modulation and a rhythmictiy (T-rhythm) that was often independent of respiration. Hypoxia evoked a graded fall in TVR indicating vasodilatation, but there were no consistent changes in CVA unit firing rate or T-rhythm frequency, although respiratory modulation increased. By contrast, during hyperthermia, TVR was low and CVA unit activity was absent. Systemic hypoxia evoked graded increases in TVR, indicating vasoconstriction, and in 8% O₂ there was recommencement of firing in some CVA units, at low discharge rate, with respiratory modulation but no T-rhythm. These results indicate that the changes evoked by systemic hypoxia in TVR and sympathetic nerve activity to CVA are dependent on core temperature. During modest hypothermia, hypoxia-induced cutaneous vasodilatation in the tail is independent of sympathetic activity, whereas during hyperthermia, when sympathetic activity is ‘switched off’, severe hypoxia initiates respiratory-related low level activity, causing cutaneous vasoconstriction.

In human subjects, cats and rats, systemic hypoxia evokes an increase in muscle sympathetic nerve activity (MSNA) (e.g. Gregor & Jänig, 1977; Blumberg et al. 1980; Fukuda et al. 1989; Somers et al. 1989; Hudson et al. 2002). Nevertheless, the vasoconstrictor effect of increased MSNA is overcome to yield muscle vasodilatation that in humans and rats is largely attributable to the actions of locally released adenosine and nitric oxide (NO) released from the endothelium (Blitzer et al. 1996; Leuenberger et al. 1999; Edmunds et al. 2003,Ray et al. 2002). Sympathetic vasoconstriction in muscle is also blunted during systemic hypoxia by mechanisms that include the particular vulnerability of the vasoconstrictor influence of noradrenaline on α₂-adrenoreceptors (Tateishi & Faber, 1995; Coney & Marshall, 2007).

By contrast, the effect of systemic hypoxia upon the cutaneous circulation, which has a rich sympathetic innervation and is heavily involved in thermoregulation (Rowell, 1983), is much less clear. It was deduced from the effects of sympathetic denervation and body warming in the rabbit, that although graded systemic hypoxia evokes vasodilatation in the ear and hindlimb skin, sympathetic vasoconstrictor activity increases to arterial resistance vessels of hindlimb skin in severe hypoxia, but decreases to arteriovenous anastomes (AVAs) of the ear, with a weak underlying sympathetic vasoconstriction (Chalmers & Korner, 1966), presumably in resistance vessels. It was also reported that systemic hypoxia evokes cutaneous vasoconstriction in the hand of human subjects (Abrahamson et al. 1943). Later, it was shown that systemic hypoxia evoked a decrease in sympathetic activity to the rabbit ear unless the hypoxia was severe, when an increase in cutaneous sympathetic activity occurred (Iriki & Kozawa, 1975). Further experiments on decerebrated rabbits led to the conclusion that suprabulbar structures, and by implication central thermoregulatory regions, are required for the cutaneous inhibitory response to systemic hypoxia (Iriki & Kozawa, 1976). However, these sympathetic recordings were made from multifibre preparations and so gave no indication of whether different fibres showed directionally different responses and/or supplied different types of blood vessels. On the other hand, in the cat, systemic hypoxia or direct stimulation of carotid chemoreceptors evoked a decrease in many of the sympathetic fibres that supplied the skin of the hindlimb and paw, but no change, or an increase in activity in others (Gregor & Jänig, 1977; Blumberg et al. 1980). These fibres were ‘split’ from the whole nerve with no direct evidence of the vessels they were supplying. However, it was speculated that they were destined for AVAs (and veins) and arterial resistance vessels, respectively, in accord with the interpretation Chalmers & Korner (1966) placed on their results (Gregor & Jänig, 1977; Blumberg et al. 1980). Subsequently, it has been argued that cutaneous vasodilatation in response to systemic hypoxia is part of a centrally regulated decrease in the thermoregulatory set point (anapyrexia) that leads to heat loss, inhibition of shivering thermogenesis, and a decrease in O₂ consumption and is therefore O₂ sparing (e.g. Gautier & Bonora, 1992; Steiner & Branco, 2002; Madden & Morrison, 2005).

In view of all of these findings and proposals, the primary objective of the present study was to use the focal recording technique to investigate the effects of systemic hypoxia on the activity of single sympathetic nerve fibres on the surface of the caudal ventral artery (CVA) of the rat tail (Johnson & Gilbey, 1994, 1996) whilst simultaneously recording tail blood flow (TBF) from the CVA (Johnson et al. 2001), and to accomplish this at relatively low and high core temperatures. The tail circulation of the rat, like the cutaneous circulation, plays a major role in thermoregulation (Torrington, 1966).

By using the focal recording technique it has already been shown that the activity in sympathetic fibres that supply the CVA often has respiratory-related rhythmicity (Johnson & Gilbey, 1996, 1998) as is common in sympathetic fibres (Häbler et al. 1994). However the dominant rhythm was not necessarily identical with the respiratory rhythm and has been termed the T-rhythm (Johnson & Gilbey, 1996), with a frequency of 0.4–1.2 Hz under normocapnic, normothermic conditions. The T-rhythym was more likely to become entrained to the respiratory rhythm when central respiratory drive (CRD) was increased by hypercapnia (Chang et al. 1999), but it persisted when CRD was absent, and was decreased by aortic baroreceptor stimulation, which had no effect on CRD (Johnson & Gilbey, 1996, 1998). Importantly, during hyperthermia, the T-rhythm frequency decreased, as did the mean firing frequency, even though CRD increased (Johnson & Gilbey, 1996, 1998). It is now widely accepted that changes in the rhythmicity of sympathetic activity are important in producing changes in vascular tone (see McAllen & Malpas, 1997). Certainly, rhythmic sympathetic discharge produces greater vasoconstriction than the same number of impulses delivered at constant frequency in both tail and skeletal muscle of the rat in vivo (Johnson et al. 2001; Coney & Marshall, 2003), in other vascular beds and in isolated arteries (e.g. Pernow et al. 1989). This is, at least in part, because the patterning of the activity determines the relative importance and magnitude of effect of the different sympathetic cotransmitters (e.g. Kennedy et al. 1986; Sjöblom-Widfeldt & Nilsson, 1990). Thus, the second objective of the present study was to investigate the effect of systemic hypoxia upon the T-rhythm and respiratory rhythmicity of the sympathetic fibres that supply the CVA at two different core temperatures and to establish the relationship between any changes in rhythmicity and the vascular response in the CVA. We used a core temperature of 36.3°C when sympathetic vasoconstrictor activity to, and vascular tone of, the CVA was relatively high, and 39°C when sympathetic activity to the tail was ‘switched off’ and vascular tone was low (Johnson & Gilbey, 1996; Häbler et al. 2000).

Methods

Anaesthesia and general preparation

Experiments were performed on spontaneously breathing male Wistar rats (270–350 g). All experiments were approved by UK legislation under the Animals (Scientific Procedures) Act 1986. Rats were initially anaesthetized with an oxygen–halothane (3.5%) mixture and then maintained by intravenous infusion of Saffan (7–12 mg kg⁻¹ h⁻¹; Schering-Plough Animal Health, Welwyn Garden City, UK) as previously described (Johnson et al. 2001). Both brachial arteries were cannulated to allow arterial blood pressure (ABP) to be continuously recorded and to provide samples for analysis of the partial pressures of O₂ and CO₂ (P_a,O2, P_a,CO2, respectively) and arterial pH (see below). The trachea was cannulated low in the neck and the animal allowed to breathe spontaneously with room air. Tracheal pressure (TP) was recorded as an index of respiratory frequency by means of a side arm leading from the tracheal cannula to a pressure transducer (NL108A, Digitimer, UK). Arterial blood samples were taken regularly throughout experiments to maintain blood gases at levels reported in Table 1. Core temperature was measured from the oesophagus and maintained at 37°C by means of a homeothermic blanket placed under the animal except when changed during the protocol (see below). A cotton wool blanket was placed over the rat when recordings were being made at a higher core temperature. This was removed when the animal was allowed to cool as part of the protocol (see below for further details). Tail skin temperature was not measured: the bath in which the tail was placed (see below) was at room temperature (21–25°C), and rose slightly (1–3°C) when the animal was heated.

Table 1.

Baseline values (mean ±s.e.m.) recorded at two different core temperatures

	Baseline values in modest hypothermia	Baseline values in hyperthermia
Temperature (°C)	36.3 ± 0. 1	39.2 ± 0.4**
Respiratory frequency (Rf; Hz)	1.54 ± 0.05	1.94 ± 0.06**
(breaths min−1)	(92 ± 3)	(118 ± 4)
Heart rate (HR; Hz)	6.67 ± 0.14	7.13 ± 0.22*
(beats min−1)	(400 ± 8)	(428 ± 13)
Mean arterial blood pressure (ABP) (mmHg)	104 ± 5	99 ± 4
Tail blood flow(TBF) (ml−1 min−1)	0.34 ± 0.05	1.61 ± 0.19***
Tail vascular resistance (TVR) (mmHg−1 ml−1 min−1)	477 ± 63	74 ± 15***
CVA unit firing rate (Hz)	1.37 ± 0.2	0
T-rhythm frequency (Hz)	0.97 ± 0.12	0
Femoral blood flow (FBF) (ml−1 min−1)	1.33 ± 0.70	2.66 ± 1.04***
Femoral vascular resistance (FVR) (mmHg−1 ml−1 min−1)	85 ± 9	40 ± 8**

P < 0.05

^**P < 0.001

^***P < 0.001

below versus above thermoregulatory threshold, 1-way ANOVA.

The preparation of the CVA for recording of sympathetic nerve activity and the lumbar sympathetic chains (LSCs) for stimulation has been previously described (Johnson & Gilbey, 1994, 1996). Briefly, with the tail in a bath filled with physiological saline (0.9%), the surface of the CVA was exposed by a longitudinal incision on the ventral surface of the tail. Sympathetic nerve activity was recorded from fibres on the adventitia of the proximal half of the CVA, by using a glass electrode pulled to a tip with an internal diameter of 20–50 μm and filled with physiological saline. This was placed proximally on the CVA and a ‘seal’ between the electrode tip and vessel was achieved by applying gentle suction. By means of a laparotomy, stimulating electrodes were placed on the LSCs between L3 and L5 to aid in identification of single units (Johnson & Gilbey, 1994) and to confirm recording electrode attachment when nerve activity was switched off during body heating (see below; Fig. 1). Nerve activity was amplified (100 000–200 000 times) and filtered (300–1500 Hz).

Figure 1. Example of cardiovascular and neural responses evoked in an individual animal by 8% O₂ during modest hypothermia (A) and hyperthermia (C), and changes recorded during body warming (B).

A, during modest hypothermia, below the thermoregulatory set point, tail blood flow (TBF) was low (a), reflecting high tail vascular resistance (TVR, c) and ongoing sympathetic nerve activity to the CVA (d). During 5 min of 8% O₂, arterial blood pressure (ABP, b) fell and TVR decreased. In this example, CVA unit activity decreased by the end of hypoxia. In Inset in d, 5 on-going action potentials discriminated in the rate histogram are superimposed to show the singular nature of the unit from which action potentials were recorded. B, as temperature rose above thermoregulatory threshold, TBF increased and TVR decreased, while CVA unit activity decreased and then ‘switched off’. Inset in d shows 3 action potentials evoked from the lumbar sympathetic chains (sweep delay 100 ms, distance between stimulus and recording sites 55 mm, conduction velocity 0.55 ms⁻¹), demonstrating that the recording position was still viable. C, during hyperthermia, when baseline TBF was high and TVR low, and there was no on-going CVA nerve activity. Under this condition, 8% O₂ evoked an increase in TVR and fall in ABP, while the CVA unit that was being recorded from ‘switched on’. The inset in d shows 5 superimposed action potentials evoked from the LSC (see Methods), demonstrating that it was the same unit as in A and B.

In all experiments, blood flow was recorded from the distal half of the CVA, approximately 5–8 cm from the tip, using a perivascular flow probe (0.7 V; Transonic Systems, Ithaca, NY, USA) connected to a dual channel flow meter (T106; Transonic Systems; Johnson et al. 2001). In some experiments, blood flow was simultaneously recorded from a femoral artery (see Johnson et al. 2001) by using a second transonic flow probe that was connected to the second channel of the flow meter.

All signals were stored on computer via a lab interface (1401; Cambridge Electronic Design (CED), Cambridge, UK) compatible with software for on- and off-line monitoring and analysis (Spike2, CED). Software calculated mean arterial blood pressure from the blood pressure trace along with heart rate. Similarly, respiratory frequency (R_f) was calculated on-line from the TP signal. Tail and femoral vascular resistance (TVR, FVR, respectively) were calculated on-line as TBF or FBF divided by ABP with a calculation frequency of 5 Hz. This software also generates trigger signals from nerve action potential, ABP and TP wave-forms, which were used for time series histogram analysis of action potential patterning (interspike interval histograms (ISIHs), auto- and cross-correlograms, see Johnson & Gilbey, 1996). These were calculated over 60 s epochs, so that the 5 min period of hypoxia (see below) generated five histograms for each variable. Modal intervals for the first peak in the ISIHs were used to calculate intraburst intervals. In most units, a second peak, corresponding to interburst interval, was not clearly defined (see Results). Thus, no attempt was made to obtain group data for these intervals. Auto-correlations of nerve activity were used to show the periodicity of sympathetic discharge. TP-triggered cross-correlations of nerve discharge were used to examine periodicity of respiration: this reflected the periodicity of CRD as the animal was spontaneously breathing. Modal frequencies of the respiratory cycle and T-rhythm were calculated from the reciprocal of peak-to-peak time. These histograms allowed comparison between central respiratory and T-rhythm periodicities: we did not attempt to deduce the phase relationships between the two rhythms because TP does not give precise information on the timing of central respiratory activity. Similarly, we did not attempt detailed analysis of the modulation of sympathetic nerve activity by CRD.

Protocol

In early experiments, when the core temperature of the rat was maintained at ∼37°C (∼36.7–37.3°C), TBF was variable (0.3–1.75 ml min⁻¹). When TBF was in the high end of this range, it was not possible to record on-going activity, whereas this was possible when TBF was in the low end of the range. Therefore, for the first part of the protocol in experiments reported here, core temperature was kept slightly below 37°C (see Results) such that TBF was relatively low and TVR relatively high to facilitate recording of ongoing sympathetic activity from the CVA. The singular nature of the recorded, on-going action potentials was confirmed by comparison with an action potential evoked from the LSCs and constancy of shape (Fig. 1, see Johnson & Gilbey, 1996). On-going activity and other variables were monitored for a 5 min control period. The inspirate was then switched from air to 12% O₂ in N₂ for 5 min, and then back to air for 5 min. At the end of each 5 min period, an arterial blood sample was taken for gas analysis. After a recovery period of 10 min, and assuming that the neural recording was still intact, this part of the protocol was repeated except that the rat breathed 8% O₂ rather than 12% O₂.

The animal was then slowly warmed until ongoing-sympathetic activity switched off as part of the thermoregulatory response to an increase in core temperature (see Results). The position of the electrode on the original nerve recording site was confirmed by establishing that an action potential could be evoked from the LSCs. This was particularly important because it was common for the recording to be lost, presumably due to movement associated with respiratory changes, or vasomotion of the blood vessel beneath the electrode tip during temperature changes or hypoxia. The animal was then given 12% and then 8% O₂ to breathe for 5 min each as described above. If unit activity did not resume during either period of hypoxia, further attempts were made to record an action potential evoked by stimulating the LSCs to confirm that the original recording site was intact. Experiments in which no action potential could be evoked are not included in the analysis of this part of the protocol.

Statistical analyses

Results are expressed as means ±s.e.m. At the lower and at the higher core temperature (modest hypothermia and hyperthermia, respectively), the values recorded under control conditions over the minute prior to each period of hypoxia, in each of the 5 min during hypoxia were compared by one-way analysis of variance (ANOVA) followed by Bonferroni's post hoc test when appropriate (P < 0.05). Comparisons were also made between baseline values recorded at the lower and higher core temperatures by using one-way ANOVA.

Results

Baseline values for all cardiovascular and respiratory variables measured when core temperature was 36.3 ± 0.1°C (n= 13) are shown in Table 1, while blood gas values measured in the last minute before and in the final minute of each period of hypoxia are shown in Table 2. Under these conditions, TBF and FBF were low (0.34 ± 0.05, n= 13 and 1.33 ± 0.63 ml min⁻¹, n= 10, respectively), while TVR and FVR were relatively high (477 ± 63 mmHg⁻¹ ml⁻¹ min⁻¹ and 85 ± 9 mmHg ml min⁻¹, respectively, see Figs 1 and 2E, F, H and I). On-going activity was present in all CVA fibres from which recordings were made (mean firing rate 1.37 ± 0.2 Hz, n= 13; Figs 1 and 2D). The ISIHs revealed that units generally fired rhythmically in bursts of action potentials with or without activity between bursts (11/13; see Fig. 3Aa), or sometimes as discrete regularly occurring single action potentials (2/13), similar to unit activity previously recorded from the CVA (e.g. Johnson & Gilbey, 1996). The mean modal intraburst frequency was 22.6 ± 3.31 Hz (n= 11). The autocorrelograms revealed that all units fired rhythmically in the frequency range of the T rhythm (see Johnson & Gilbey, 1994, 1996), the mean T-rhythm frequency being 0.98 ± 0.12 Hz (see Figs 2G and 3Ab). Respiratory-related modulation was also present in all units (e.g. Fig. 3Ad) as revealed by the TP-triggered cross-correlograms. The T-rhythm was synchronized with respiration (R_f 1.54 ± 0.05 Hz) in relatively few cases (8/13) in ratios of either 1 : 1 (3/8) or 1 : 2 (5/8). In the remaining cases, there was no synchrony between T-rhythm and respiratory cycle (Fig. 3Ab and c), tending to decrease the distinction between separate bursts of unit activity. As a consequence, only a few units (7/13) showed an obvious second major peak in the ISIH (e.g. Fig. 3Ba).

Table 2.

Blood gas and pH values (mean ±s.e.m.) recorded in normoxia (control) and at the end of 5 min periods of breathing 12 or 8% O₂ during modest hypothermia or hyperthermia; core temperature and, respectively

	Control		12% O2		8% O2
	Modest hypothermia	Hyperthermia	Modest hypothermia	Hyperthermia	Modest hypothermia	Hyperthermia
Pa,O2(mmHg)	95.0 ± 3.7	94.8 ± 3.1	50.3 ± 1.0***	47.8 ± 1.2***	31.0 ± 0.7***‡‡	33.5 ± 3.0***‡‡
Pa,CO2 (mmHg)	36.0 ± 1.2	36.2 ± 1.2	31.6 ± 1.1**	30.6 ± 2.3**	25.9 ± 2.0***‡	23.6 ± 1.6***‡
pHa	7.37 ± 0.02	7.39 ± 0.02	7.44 ± 0.01*	7.44 ± 0.03*	7.42 ± 0.03	7.43 ± 0.02*
Respiratory rate, Rf (Hz)	1.54 ± 0.05	1.94 ± 0.06††	1.77 ± 0.06**	2.27 ± 0.04***‡	2.00 ± 0.14**†‡	2.42 ± 0.22***†
(breaths min−1)	(92 ± 3)	(118 ± 4)	(106 ± 4)	(136 ± 2)	(120 ± 8)	(145 ± 13)
Heart rate (Hz)	6.67 ± 0.14	7.13 ± 0.22†	6.69 ± 0.28	7.33 ± 0.25†	7.35 ± 0.36‡	8.73 ± 0.15**††‡‡
(beats min−1)	(400 ± 8)	(428 ± 13)	(401 ± 17)	(440 ± 15)	(441 ± 22)	(524 ± 9)

Values of R_f and HR under the same conditions are shown for comparison

^‡P < 0.05

^‡‡P < 0.01

12%versus 8% O₂. P < 0.05

^**P < 0.01

^***P < 0.001

control versus 12% or 8% O₂.

^†P < 0.05

^††P < 0.01

modest hypothermia versus hyperthermia.

Figure 2. Responses evoked in cardiovascular and respiratory variables and in CVA unit activity during 5 min periods of 12 and 8% O₂ during modest hypothermia.

Values for each variable are shown as mean ±s.e.m. during the final 5 min of normoxia, and in the 1st, 3rd and 5th minutes of hypoxia (1, 3 and 5, respectively). Values recorded during 12% O₂ are shown as squares, continuous line; those recorded during 8% O₂ are shown as triangles, dashed line. All abbreviations are described in the text. Note the differences in scales used for FBF and FVR and for TBF and TVR. *P < 0.05, **P < 0.01, ***P < 0.001, control versus 12% O₂. ††P < 0.01, †††P < 0.001, control versus 8% O₂. ‡‡‡P < 0.001, 12%versus 8% O₂.

Figure 3. Typical example of rhythms recorded from a CVA unit in modest hypothermia over a 60 s period immediately before (A) and in the last minute (B) of a 5 min period of 8% O₂.

In A in normoxia, the ISIH (a) of unit activity (169 sweeps) revealed two peaks, reflecting the bursting nature of activity, the first corresponding to the intraburst interval (modal interval 0.05 s), the second to the interburst interval (modal interval 1.15 s). b, the autocorrelation (169 sweeps) showed a clear T-rhythm at 0.83 Hz (periodicity 1.2 s). c, tracheal pressure-triggered autocorrelation (90 sweeps) indicated an R_f of 1.5 Hz (0.65 s periodicity). NB, the T-rhythm frequency and respiratory frequency are completely different. d, tracheal pressure-triggered cross correlation (91 sweeps) indicates a strong respiratory modulation of sympathetic unit activity. Ba ISIH (181 sweeps) shows a small decrease in interburst interval to 0.85 s in hypoxia, but the modal intraburst interval was not changed. b, the autocorrelation (181 sweeps) shows that the T rhythm frequency increased slightly to 1 Hz (periodicity 1 s) in hypoxia. c, tracheal pressure-triggered autocorrelation (127 sweeps) indicated an increase in R_f to 2.2 Hz (0.45 s periodicity). NB the T-rhythm and respiratory frequency remain unrelated. d, the tracheal pressure-triggered cross correlation (127 sweeps) indicates stronger respiratory modulation of sympathetic activity in hypoxia than normoxia.

Representative traces and analyses showing responses evoked by hypoxia in individual animals are presented in Figs 1 and 3, while Fig. 2 shows the grouped data. The R_f increased in response to both levels of hypoxia (more in 8% than 12% O₂), the level reached being maintained, whereas HR rose modestly, but not significantly in the first 3 min and recovered by the fifth minute (Fig. 2A and C). ABP decreased to similar extents during 12 and 8% O₂ (Fig. 2B). Meanwhile, TVR fell during both levels of hypoxia, indicating vasodilatation in tail vasculature, with no change in TBF (Fig. 2E and H). Similarly, FVR decreased, the fall in FVR being greater in 8% O₂ than in 12% O₂ (Fig. 2F and I), with no change in FBF.

Respiratory modulation was clear in all CVA units recorded in both levels of hypoxia, and thus the frequency of the respiratory-related rhythmicity increased with R_f (see Fig. 3Bd for example). However, during 12% O₂, there was no significant change in the mean frequency of CVA unit activity (group mean: 1.38 ± 0.21 versus 1.41 ± 0.26 Hz, n= 13; see Fig. 2D), although 6/13 units showed a slight increase in discharge frequency (from 1.54 ± 0.29 to 1.87 ± 0.30 Hz) and 7/13 a slight decrease (from 1.15 ± 0.31 to 0.81 ± 0.32 Hz). There was no obvious relationship between control discharge rate and response to hypoxia (see Fig. 4). During 12% O₂, there was also no significant change in T-rhythm frequency (Fig. 2G): in three units the T-rhythm (1.00 ± 0.12 Hz) disappeared during hypoxia and had not returned after 5 min of recovery, in two units the T-rhythm frequency moderately increased (from 0.66 ± 0.26 to 0.76 ± 0.28 Hz), in two units it decreased (from 1.11 ± 0.14 to 0.83 ± 0.15 Hz) and in six it was unchanged (0.96 ± 0.43 Hz). There was no obvious relationship between change in rate and T-rhythm frequency, these two variables changing in the same direction in only one unit. Of the eight units with T-rhythms that were locked to that of respiration during normoxia, synchrony was lost in six units during 12% O₂ remained in a ratio of 1 : 1 in one and changed from 1 : 1 to 1 : 2 in the other.

Figure 4. Discharge rate and T-rhythm frequency of activity recorded during modest hypothermia in individual CVA units during normoxia and 12% O₂ (LHS), and during normoxia and 8% O₂ (RHS).

Mean values for each set of data are indicated by squares and dotted lines. It is clear that in responses to 12 or 8% O₂ CVA sympathetic discharge rate and T-rhythm frequency could increase, decrease or remain the same, and the response could not be predicted from the control value.

The effects of 8% O₂ on unit activity were similar to 12% O₂. There was no significant change in mean discharge frequency (1.76 ± 0.28 versus 1.33 ± 0.27; Fig. 2D), as 5/8 units showed a decrease (2.18 ± 0.27 to 1.27 ± 0.36 Hz) and 3/8 an increase (1.32 ± 1.00 to 2.2 ± 1.10 Hz; see Fig. 4). Of the six units with an obvious T-rhythm in normoxia, three showed a reduction in T rhythm frequency during 8% O₂ (1.5 ± 0.26 to 1.04 ± 0.37 Hz), two showed an increase (from 0.83 to 2.2 Hz and from 0.77 to 0.97 Hz), and one showed no change (0.80 Hz), so that overall there was no significant change (see Fig. 2G). Again, there was no relationship between change in rate and T-rhythm frequency. The T-rhythm was locked to respiratory rhythm in 4/6 units in normoxia: in two of these, synchrony was lost during 8% O₂ and in two it was retained.

There was no apparent relationship between the CVA unit response to 12 and 8% O₂: there were changes of rate in the same direction in only 2/8 units, Similarly, there was no apparent relationship between change in T-rhythm frequency in 12% O₂ and 8% O₂. The ISIHs showed that there were no differences in modal intraburst frequencies between control periods and either level of hypoxia: control, 22.6 ± 3.3 Hz (n= 11); 12% O₂, 21.7 ± 2.9 Hz (n= 11); 8% O₂, 22.6 ± 4.5 Hz (n= 6). In two units in which there was an obvious second ISIH peak in normoxia, reflecting the T-rhythm, the interburst interval decreased during hypoxia, and a third smaller peak with a shorter interspike interval possibly corresponding to the respiratory frequency, was also apparent (see Fig. 3Ba). In the remaining units, a second peak was no more obvious in hypoxia than in normoxia (see above).

Responses evoked during hyperthermia

When core temperature was raised to 39.2 ± 0.4°C (n= 11) there was no change in ABP, but there were marked decreases in baseline TVR and FVR and increases in TBF and FBF (Fig. 5E, F, H and I). Concomitantly both baseline HR and R_f increased (Table 1 and Fig. 5A and C). Accompanying the fall in TVR, there was a switch-off of unit activity recorded from the CVA (Figs 1B and 5D).

Figure 5. Responses evoked in cardiovascular and respiratory variables and in CVA unit activity during 5 min periods of 12 and 8% O₂ during hyperthermia.

Values for each variable are shown as means ±s.e.m. during the final 5 min of normoxia, and in the 1st, 3rd and 5th minutes of hypoxia (1, 3 and 5, respectively). Values recorded during 12% O₂ are shown as squares, continuous line; those recorded during 8% O₂ are shown as triangles, dashed line. All abbreviations are described in the text. Note the differences in scales used for FBF and TBF and for FVR and TVR. Note also that CVA unit activity was ‘switched off’ in hyperthermia and was not initiated during 12% O₂. However, in 8% O₂, activity began in 4 out of 8 CVA units (D), with respiratory rhythmicity but no T-rhythm (G). *P < 0.05, **P < 0.01, ***P < 0.001, control versus 12% O₂. †P < 0.05, ††P < 0.01, †††P < 0.001, control versus 8% O₂. ‡P < 0.05, 12%versus 8% O₂.

During 12 and 8% O₂, there was a marked decrease in ABP, while HR and R_f increased and FVR decreased, as occurred during hypoxia at lower core temperature (Fig. 2, cf. Fig. 5). In contrast to the responses observed at lower core temperature, TVR increased during both 12 and 8% O₂ in every test and there was a pronounced fall in TBF (Fig. 5H and E).

Concomitantly, during 12% O₂, none of the 13 units from which recordings were made switched back on: in each case, attachment of the electrode to the nerve was confirmed by evoking a response from the LSCs during hypoxia (see Fig. 1Bd and 5F). However, during 8% O₂, in 4 of the 8 units in which the recording site remained viable, the increase in TVR was accompanied by recommencement of unit firing (Figs 1C and 5I). The frequency of this discharge was considerably lower than that recorded during normoxia or hypoxia at the lower core temperature (cf. Fig. 2Af and 2Bf). Of these four units, two had responded with a slight increase in discharge rate during 8% O₂ in moderate hypothermia, one had responded with a decrease and in one unit there had been no change in rate. No T-rhythm was observable, although strong respiratory-related modulation was apparent in the TP-triggered cross-correlograms (see Fig. 6A and C).

Figure 6. Typical example of rhythmicity recorded in a CVA unit activity over the last 3 min of a 5 min period of 8% O₂ during hyperthermia.

Due to the low firing rate (see text), CVA unit discharge was analysed over the last 3 min of hypoxia so as to reveal potential rhythms. The TP-triggered autocorrelation (B: 210 sweeps) indicated a R_f of 2.8 Hz (periodicity 0.35 s) and the TP-triggered cross correlation with CVA unit discharge (C: 210 sweeps) indicated strong respiratory modulation. However, the CVA unit autocorrelation (A: 52 sweeps) showed no discernable T-rhythm; the possible peak at ∼2.5 s was not visible at 5 s (not shown). The ISIH was identical to the autocorrelation as there were no short interspike intervals.

Discussion

The main new finding of the present study is that responses evoked by systemic hypoxia in the vasculature served by the CVA and in sympathetic nerve fibres that supply the CVA are dependent on core temperature. When core temperature was below normal (36.3°C), mild or severe systemic hypoxia (breathing 12 or 8% O₂) had no significant effect on the sympathetic activity recorded from the CVA, but TVR decreased, indicating vasodilatation in the tail. However, when core temperature was 39.2°C, activity recorded from the CVA switched off, baseline TVR decreased and severe hypoxia initiated activity in some CVA units, while both levels of hypoxia evoked an increase in TVR, indicating tail vasoconstriction.

For the first part of our protocol, core temperature was maintained at 36.3°C so as to ensure ongoing activity in sympathetic units on the CVA and achieve stable levels of TBF and TVR. Although most experiments on anaesthetized rats are performed with core temperature maintained at 37°C, core temperature in the conscious rat is ∼38°C (e.g. Kregel et al. 1990). Thus 36.3°C represents modest hypothermia. Under this condition, CVA unit activity in spontaneously breathing rats anaesthetized with Saffan was comparable to that in spontaneously breathing, or artificially ventilated rats under pentobarbitone/chloralose anaesthesia and maintained at ∼37°C, with or without vagi and/or aortic and sinus nerves cut (see Johnson & Gilbey, 1994, 1996; Chang et al. 1999, 2000). Thus, activity in all units was rhythmic and the dominant rhythm had a frequency of 0.4–1.2 Hz (0.96 Hz), termed the T-rhythm (Johnson & Gilbey, 1996): it was sometimes the same as and sometimes different from the central respiratory rhythm. By contrast, Häbler et al. (1999), who recorded from the ventral collector nerve (VCN) to the tail in artificially ventilated rats maintained at ∼37°C under pentobarbitone anaesthesia with vagi, aortic and sinus nerves cut, reported that central respiratory rhythm was dominant and there was no independent T-rhythm. Reasons for the disparity are not clear given the variety of conditions under which the T-rhythm has been recorded. However, our results are compatible with the proposal that sympathetic neurones that supply the CVA are driven by multiple T-rhythm oscillators (Chang et al. 1999, 2000).

Responses evoked during modest hypothermia

Although neither level of systemic hypoxia evoked a statistically significant change in mean discharge frequency of CVA units, some showed an increase and others a decrease. These findings might seem consistent with those of Jänig and colleagues (Gregor & Jänig, 1977; Blumberg et al. 1980) who reported opposing patterns of behaviour in single sympathetic fibres supplying the paw and skin of the cat during systemic hypoxia, and deduced these represented fibres destined for arteries or AVAs, respectively. However, in the present study, there was no consistency in the responses of individual units to the two levels of hypoxia. Further, we recorded from fibres in the adventitia of the proximal CVA and there is no reason to suppose they supplied any vessel other than the CVA. Branches arise from paravascular bundles and innervate short lengths of the CVA: AVAs have their own nerve supply and are mainly in the distal tail (Sittiracha et al. 1987; Anderson & McLachlan, 1991). Thus, it is unlikely that we recorded from functionally different populations of neurones. Rather, the very small changes in mean frequency, < 1Hz in either direction, probably reflected variable changes in rhythmicity in individual CVA neurones (see below).

The T-rhythm frequency did not change significantly during either level of hypoxia, some individual units showing no change, some an increase, others a decrease and a few losing their T-rhythm. When CVA unit activity is continuously recorded under stable experimental conditions, the rhythmicity shows periods of instability with changes in pattern of discharge and T-rhythm frequency (Johnson & Gilbey, 1996; Chang et al. 1999). Thus, the present variability may reflect this inherent instability. Certainly, the present results suggest that the combination of an increase in CRD caused by peripheral chemoreceptor stimulation and of baroreceptor unloading caused by the hypoxia-induced fall in ABP, has no consistent effect on T-rhythm frequency. This apparently contrasts with the findings that blood volume expansion, or continuous stimulation aortic baroreceptor afferents in artificially ventilated rats, did not affect CRD frequency, but decreased T-rhythm frequency in CVA units (Johnson & Gilbey, 1998). Vascular and cardiac components of baroreceptor reflex are fully effective in Saffan-anaesthetized rats (e.g. Hebert & Marshall, 1988). Thus, the present results indicate that at least in modest hypothermia, either baroreceptor unloading is not as effective as baroreceptor stimulation in altering T-rhythm frequency, or that baroreceptor inputs are less effective when CRD frequency is concurrently increased by peripheral chemoreceptor stimulation.

Most CVA units also had a respiratory-related rhythm, as reported by Johnson & Gilbey (1996) in spontaneously breathing rats. During mild and severe hypoxia, the frequency of this respiratory rhythm increased, i.e. CVA unit rhythmicity followed the increase in CRD frequency evoked by peripheral chemoreceptor stimulation. Due to the variability of the T-rhythm frequency during hypoxia (see above), the synchronization between T-rhythm and respiratory rhythm that existed in ∼50% of units in normoxia, was generally lost during hypoxia. By contrast, under the highly controlled conditions of pneumothorax and constant artificial ventilation with vagi cut, the increase in CRD evoked by hypercapnia led to greater synchronization of the T-rhythm and CRD in individual and pairs of CVA units (Chang et al. 1999). Thus, the present results suggest that in modest hypothermia, synchronization of the influences of CRD and T-rhythm oscillators on CVA unit activity occurs less readily under more natural conditions when increased CRD is associated with an increase in CRD frequency. This suggestion is consistent with the finding that in artificially ventilated rats, synchronization of the T-rhythm and CRD was less likely when the frequency of the lung inflation cycle was moved further away from the T-rhythm frequency (Chang et al. 2000).

Thus, the present findings indicate that during modest hypothermia, mild or severe systemic hypoxia has relatively little effect on the pattern of sympathetic discharge to the CVA. There was no consistent change in T-rhythm frequency, whereas the frequency of the respiratory rhythm increased, but resulted in minor changes in mean discharge frequency. Further, the ISIHs revealed that neither level of hypoxia had any effect on intraburst frequency, which remained at ∼22 Hz. Moreover, interburst intervals were inconsistent in both normoxia and hypoxia, reflecting the presence of a T-rhythm and/or a respiratory rhythm. It is therefore very unlikely that changes in the patterning of the sympathetic discharge to the CVA contributed to the substantial decrease in TVR. Rather, this vasodilatation may reflect blunting of the α₂ component of the sympathetic vasoconstrictor influence on the CVA (Bao et al. 1993), which is especially vulnerable to hypoxia (Tateishi & Faber, 1995; Coney & Marshall, 2007). It may also reflect myogenic dilatation induced by the hypoxia-induced fall in ABP and/or dilatation induced by adenosine and NO that are released by the endothelium, as occurs in hindlimb muscle (Edmunds et al. 2003; Ray et al. 2002).

These proposals are consistent with the report that in normothermia, local dilator influences were totally responsible for vasodilatation evoked by systemic hypoxia (P_a,O2 37–31 mmHg) in the hindlimb skin of rabbits and partly responsible for vasodilatation in the ear (Chalmers & Korner, 1966). We cannot discount the possibility that vasodilatation induced by inhibition of sympathetic activity to AVAs supplied by the CVA contributed to the fall in TVR, as reported for the rabbit ear (Chalmers & Korner, 1966; Iriki & Kozawa, 1975, 1976). However, TBF remained constant at both levels of hypoxia and so heat delivery through the CVA was not changed. Thus, there is no reason to suppose that during modest hypothermia, the effect of mild or severe systemic hypoxia on sympathetic activity to the CVA or its vasculature, was part of an anapyrexic response that reduces core temperature (e.g. Steiner & Branco, 2002). We already know that muscle vasodilatation and increased O₂ extraction allows muscle O₂ consumption to remain constant during the levels of hypoxia used in the present study (Edmunds & Marshall, 2001). Indeed, given muscle is largely responsible for whole body , it is unlikely that even the 5 min period of severe hypoxia used in the present study evoked the O₂-sparing components of anapyresis. For comparison, conscious normothermic rats breathing 13 or 11% O₂ for 20 min showed no fall in body temperature or (Gautier & Bonora, 1992).

Effects of systemic hypoxia during hyperthermia

When core temperature was raised to 39.2°C, R_f increased, sympathetic discharge in CVA units ceased and there was a gradual decrease in TVR indicating vasodilatation attributable to sympathetic withdrawal. This pattern is fully consistent with previous findings (Johnson & Gilbey, 1996; Häbler et al. 2000; Owens et al. 2002). The fall in TVR presumably reflected vasodilatation in the arterial resistance vessels and AVAs of the tail (Torrington, 1966). Concomitantly, baseline FVR fell by ∼50%. Given vascular resistance and blood flow in limb muscles are not changed by hyperthermia (Detry et al. 1972), while MSNA is increased (Niimi et al. 1997; Keller et al. 2006), the decrease in FVR can be attributed to vasodilatation caused by inhibition of sympathetic vasoconstrictor drive to hindlimb skin. Certainly, gross forearm vascular resistance falls during hyperthermia due to cutaneous vasodilatation (Detry et al. 1972; Rowell et al. 1989). As ABP was well maintained in hyperthermia, there must have been a substantial increase in splanchnic and renal vascular resistance and effective baroreceptor regulation of ABP, as in conscious rats (Kregel et al. 1990).

During hyperthermia, mild systemic hypoxia failed to initiate activity in the 13 sympathetic CVA units from which activity had been recorded in normoxia, but severe hypoxia ‘switched on’ 4 out of 8 units so that they discharged at a low mean frequency of < 0.25 Hz. This discharge had no T-rhythm, but did have respiratory rhythmicity, attributable to a further increase in CRD: although R_f was similar at these two levels of hypoxia, previous studies showed that tidal volume was greater in severe hypoxia indicating greater CRD (Marshall & Metcalfe, 1988). Since TVR increased significantly in both mild and severe hypoxia, it seems reasonable to propose that some CVA units began to fire in mild hypoxia, but that more were recruited in severe hypoxia. If the population of CVA units from which we recorded were representative, then < 8% (< 1/13) discharged in mild hypoxia and ∼50% in severe hypoxia. That low discharge in single CVA units was associated with an increase in TVR is consistent with evidence that even single impulses delivered to perivascular nerves of the CVA in vitro evoked substantial contraction (Bao et al. 1993) and that single impulses in CVA units in vivo were accompanied by transient increases in TVR (Johnson et al. 2001). When the rabbit ear was dilated by body heating, severe hypoxia similarly evoked vasoconstriction in the ear (Chalmers & Korner, 1966). Indeed, if this observation is considered together with the present findings, it seems that when the AVAs are dilated by hyperthermia, this facilitates the ability of systemic hypoxia to evoke sympathetically evoked vasoconstriction in the arterial resistance vessels of the cutaneous circulation, by stimulating peripheral chemoreceptors (see Marshall, 1994). It should be noted that during hyperthermia, in the rat tail as well as the rabbit ear, the vasoconstrictor response to systemic hypoxia led to a substantial decrease in cutaneous blood flow and presumably restricted heat loss (cf. anapyresis).

The present study also showed that during hyperthermia, both levels of systemic hypoxia caused a significant decrease in FVR. This can most easily be explained by persistence of the hypoxia-induced vasodilatation in skeletal muscle (see above). The magnitude of the falls in FVR may have been smaller than those evoked by hypoxia during modest hypothermia, because the effect of muscle vasodilatation on FVR was offset by vasoconstriction in the arterial resistance vessels of hindlimb skin, as occurred in the tail (see above). In other words, these results are consistent with the idea that changes in core temperature differentially affect vascular responses evoked by systemic hypoxia in skeletal muscle and in skin.

Integration of influences on CVA unit activity

The very fact that the discharge evoked in CVA units by severe hypoxia in hyperthermia had respiratory rhythmicity, but no T-rhythm, accords with the idea that the oscillators that drive the T-rhythm and respiratory rhythm can be independent of one another (Johnson & Gilbey, 1996; Chang et al. 1999, 2000). They also accord with evidence that sympathetic activity to rat tail is controlled by two discrete regions of the medulla, the medullary raphé and the rostral ventrolateral medulla (RVLM). The RVLM provides the drive that maintains preganglionic neurone excitability, while the medullary raphé is the more powerful and carries the thermoregulatory drive (Smith et al. 1998; Oostuka & McAllen, 2005 and references therein). The RVLM is known to play an important role in mediating several non-thermoregulatory reflexes, including peripheral chemoreceptor and baroreceptor reflexes (see Guyenet, 2000). Thus, on the basis of the present results, we can propose that although sympathetic activity in neurones that supply CVA is switched off during hyperthermia by the raphé circuitry, they are switched on again during systemic hypoxia, in numbers and to extents that are proportional to the peripheral chemoreceptor input to the RVLM, by a drive that has respiratory rhythmicity (see Guyenet, 2000) and which induces vasoconstriction. On the other hand, during modest hypothermia, although the influence of peripheral chemoreceptor stimulation on the respiratory rhythmicity of CVA units can be demonstrated during systemic hypoxia, the dominant influence is that of the thermoregulatory input from medullary raphé on T-rhythm oscillators that drive CVA units (see Chang et al. 1999, 2000). The T-rhythm oscillators may be influenced by other inputs such as baroreceptors, but during modest hypothermia, they apparently have little net effect on pattern of discharge in the CVA units (see above). Thus, vasoconstrictor tone in the cutaneous circulation supplied by the CVA remains constant.

In summary, the present study has shown that in spontaneously breathing rats during modest hypothermia, when there is strong discharge with a T-rhythm and respiratory rhythmicity in sympathetic fibres that supply the CVA, mild or severe systemic hypoxia has no consistent effect on the T-rhythm, but increases the frequency of the respiratory rhymicity, without significantly changing mean discharge frequency, or intraburst frequency. Meanwhile, the relatively high TVR shows a modest decrease during hypoxia mediated by local dilator factors, which keep TBF constant, and maintains O₂ delivery to the cutaneous circulation of the tail without disturbing its role in thermoregulation. By contrast, during hyperthermia, when sympathetic discharge to the CVA is switched off, systemic hypoxia initiates low frequency discharge with respiratory rhythmicity, but no T-rhythm. This increases the relatively low TVR and decreases the high level of TBF, so limiting delivery of O₂ and heat to the tail, but allowing the cutaneous circulation to contribute to maintenance of total peripheral resistance and thereby O₂ delivery to other tissues including skeletal muscle and brain.