In cold defense and fever, activity increases in sympathetic nerves supplying both tail vessels and interscapular brown adipose tissue (iBAT). These mediate cutaneous vasoconstrictor and thermogenic responses, respectively, and both depend upon neurons in the rostral medullary raphé. To examine the commonality of brain circuits driving these two outflows, sympathetic nerve activity (SNA) was recorded simultaneously from sympathetic fibers in the ventral tail artery (tail SNA) and the nerve to iBAT (iBAT SNA) in urethane-anesthetized rats. From a warm baseline, cold-defense responses were evoked by intermittently circulating cold water through a water jacket around the animal's shaved trunk. Repeated episodes of trunk skin cooling decreased core (rectal) temperature. The threshold skin temperature to activate iBAT SNA was 37.3 ± 0.5°C (n = 7), significantly lower than that to activate tail SNA (40.1 ± 0.4°C; P < 0.01, n = 7). A fall in core temperature always strongly activated tail SNA (threshold 38.3 ± 0.2°C, n = 7), but its effect on iBAT SNA was absent (2 of 7 rats) or weak (threshold 36.9 ± 0.1°C, n = 5). The relative sensitivity to core vs. skin cooling (K-ratio) was significantly greater for tail SNA than for iBAT SNA. Spectral analysis of paired recordings showed significant coherence between tail SNA and iBAT SNA only at 1.0 ± 0.1 Hz. The coherence was due entirely to the modulation of both signals by the ventilatory cycle because it disappeared when the coherence spectrum was partialized with respect to airway pressure. These findings indicate that independent central pathways drive cutaneous vasoconstrictor and thermogenic sympathetic pathways during cold defense.
- vasoconstrictor fiber
- interscapular brown adipose tissue
- spectral analysis
in rats, interscapular brown adipose tissue (iBAT) is a major organ of heat production by nonshivering thermogenesis, whereas the rat's tail is a major organ of heat dissipation. Specific sympathetic nerves regulate each function, by activating β3-adrenoceptors in the BAT (6, 14) and by regulating the degree of vasoconstriction in tail vessels (22). In cold environments, rats maintain their core temperature by constricting their tail and skin vessels to reduce heat dissipation and by producing extra heat (nonshivering thermogenesis) (20). Similar mechanisms raise core temperature during fever. These mechanisms involve activation of both tail (12) and iBAT sympathetic nerves (16).
Besides their common patterns of functional activation, iBAT thermogenesis and tail vasoconstriction share similar central pathways. Sympathetic premotor neurons for both iBAT and tail nerves are believed to be located in the medullary raphé. Injecting neuroexcitatory substances into the raphé activates both nerves (13, 29). Injecting inhibitory neurotransmitter agonists into the raphé can block the excitatory effect on both nerves of experimental fever (12, 16). Moreover, cold-activated tail sympathetic nerve activity (SNA) depends on raphé neurons (24, 25), and it is believed that the same is true for cold-activated BAT (7, 15). Finally, viral tracer injections into either the BAT or the tail cause early trans-synaptic labeling of raphé neurons (7, 21, 38).
The anatomical and physiological similarities linking these two sympathetic outflows raise the possibility that they might be controlled by common central neuron groups. To approach this issue, we have made the first simultaneous recordings from iBAT and tail sympathetic nerve fibers in anesthetized rats. We examined 1) the concordance between iBAT and tail nerve responses to cooling and 2) the coherence between the frequency components expressed in the activity patterns of the two nerves.
Experiments adhered to the guidelines of the National Health and Medical Research Council of Australia and were approved by the Animal Experimentation Ethics Committee of the Howard Florey Institute. They were performed on seven male Sprague-Dawley rats (280–450 g). Animals were initially anesthetized with an intraperitoneal administration of Nembutal (pentobarbital sodium, initial dose 40–80 mg/kg; Boehringer Ingelheim). The trunk and limbs were shaved, and the trachea was cannulated. Anesthesia was then maintained by artificial ventilation with 2% isoflurane (Abbott) in oxygen. After the completion of surgery, anesthesia was replaced over ∼30 min. with urethane (Sigma) (1–1.5 g/kg iv). Rat were paralyzed with pancuronium bromide (AstraZeneca) (1.0–1.5 mg/kg iv) and mechanically ventilated with 100% oxygen. Deep anesthesia was established before paralysis, and the animal was allowed to recover between doses so that adequate anesthesia could be confirmed before paralysis was reestablished. Small supplemental doses of urethane (0.025–0.05 g) were given if necessary during the experiment to maintain anesthesia at a depth sufficient to abolish corneal and withdrawal reflexes.
The bladder was cannulated suprapubically and urine allowed to drain. The right femoral artery and vein were cannulated for measurement of arterial pressure and drug administration, respectively. The arterial line was filled with a solution of heparin sulfate (50 U/ml in 0.9% saline). Instantaneous arterial pressure and airway pressure were monitored continuously in all experiments with a transducer connected to a bridge amplifier (NL108, Digitimer). Expired CO2 concentration was monitored (FM1, Analytical Development) and end-expiratory levels maintained between 3.5% and 4.5% by adjusting ventilation volume.
The animal was mounted prone in a stereotaxic frame. Three thermocouples were glued with cyanoacrylate glue to different sites on the animal's shaved trunk skin, and their mean signal was read by a custom-built thermometer amplifier: this was taken as skin temperature. A fourth thermocouple was inserted 3 cm into the rectum to record core temperature. A hand-crafted water jacket was placed around the animal's shaved trunk (28, 30). Water from a warm (36–48°C) reservoir was circulated through the jacket at 120–200 ml/min by a pump. For cooling episodes, the pump inflow and outflow pipes were switched to a cold-water reservoir (4–10°C) for 1–4 min.
Sympathetic Nerve Fiber Recording
The tail nerve was dissected from next to the ventral tail artery, placed over a black Plexiglas platform, and desheathed. Under mineral oil, the central end of a cut nerve bundle was lifted onto one electrode of a pair of silver wires. A strand of connective tissue contacted the second electrode. Nerve bundles were dissected until clear, discriminable, few-fiber spike activity was evident in the recording (28). The right iBAT nerve was dissected from a dorsal approach, the interscapular BAT was divided near the midline, and the iBAT nerve was identified as a small bundle entering the tissue laterally from its underside (19). Under mineral oil, a small twig was cleaned and cut, and its central end was placed over a pair of silver wire electrodes. Nerve signals were recorded (NL104 amplifiers, Digitimer), amplified (×10,000) and filtered (band pass 1–1,000 Hz for iBAT; 100–600 Hz for tail SNA). A custom-built discriminator was used to detect spikes in the tail SNA signal. Raw nerve activity (digitized at 5 kHz), discriminated sympathetic unit counts, arterial pressure, airway pressure, and skin and rectal temperatures were recorded on computer using a CED 1401Plus interface and Spike2 software (Cambridge Electronic Design).
Discriminated tail sympathetic nerve spikes were counted and displayed in 15-s bins. iBAT sympathetic nerve activity was rectified and displayed in 15-s bins as the integrated activity above a threshold, which was set to eliminate baseline noise. Temperature thresholds and thermal sensitivities of both nerves were estimated using these 15-s activity measures. Linear regression lines relating nerve activity to skin or core temperature were calculated from appropriate sections of the data, following procedures described previously (28). The slope of the regression lines relating nerve activity vs. skin and core temperatures gave the thermal sensitivities of each nerve to skin cooling (K-skin), and core cooling (K-core). The relative sensitivity of each nerve to core cooling, compared with skin cooling, was given by the K-ratio; this was defined as K-core divided by K-skin (28).
To construct autospectra of nerve signals and coherence spectra for signals from two nerves, raw iBAT nerve activity was filtered (1–100 Hz) and digitized at 1 kHz. Discriminated tail sympathetic nerve spikes were converted to waveforms with a sinc function (1, 10) on Spike2 software, so that the autospectrum reflected the interspike interval rather than the shape of unit activity. Fast Fourier transforms of nerve activity were calculated on twenty-three 5.12-s Hanning windows with 50% overlap (61.5-s of data). The values of autospectra at each frequency were computed as the mean value of the powers at that frequency in the individual autospectra of these 23 segments, normalized by dividing the value at each frequency by the total mean power between 0 and 20 Hz. The coherence spectra were constructed to reveal any linear correlation between iBAT and tail nerve activity at each frequency. Partial coherence analysis (2, 11) was performed to examine the relationship between iBAT and tail nerve activity after mathematical removal of the influence of a third signal such as respiratory or cardiac modulation. The significance of coherence peaks was calculated as described by Rosenberg et al. (32): since spectra were averaged from 23 segments, a coherence peak was considered significant if it exceeded 0.13 (P < 0.05). Statistical analysis of the results was carried out using Student's t-test for paired data. The level of significance was taken as P < 0.05. Results were expressed as means ± SE.
Simultaneous recordings of ongoing SNA in iBAT and tail nerves were obtained from seven rats (Fig. 1). The activity in both nerves was similar to that described in previous publications (17, 24, 28). When activated, the discharge in the iBAT nerve consisted of irregular, high-amplitude bursts (17). Although previous reports found that respiratory modulation was only present after vagotomy (17), under our experimental conditions (with intact vagi), iBAT SNA showed an obvious relation to the ventilatory cycle (Fig. 1A). Tail SNA was recorded as few-fiber spike activity (Fig. 1B): in agreement with previous reports, we found that tail fiber activity was also modulated by the ventilatory cycle (9).
Response to Cooling Episodes
For baseline conditions, animals were kept as warm as necessary to minimize tail and iBAT nerve activity (rectal temperature = 38.5 ± 0.2°C, trunk skin temperature = 40.5 ± 0.3°C, n = 7). Following an established protocol (28), a sequence of 2- to 3-min cooling episodes was applied to the trunk skin by a water jacket. Each cooling episode lowered trunk skin temperature by 2–10°C; a sequence of 6–12 successive episodes caused rectal temperature to fall by 2–4°C (Fig. 2A). As described previously (24, 28), tail sympathetic nerve fibers were activated first by the fall in trunk skin temperature and then additionally by the subsequent fall in core temperature (Fig. 2A). The mean threshold skin temperature for activation by skin cooling was 40.1 ± 0.4°C (range 38.8–42.0°C, n = 7). After the early cooling episodes in four of seven rats, the skin rewarmed sufficiently before core temperature fell, such that tail SNA was silenced. The threshold skin temperature for this was 38.9 ± 0.9°C (n = 4). In the other three rats, tail SNA was reduced by skin rewarming. The mean threshold rectal temperature for the core temperature-related response of tail SNA was 38.3 ± 0.2°C (range 37.7–39.0°C, n = 7).
From a baseline during the warm control period, when there were no high-amplitude bursts, iBAT SNA increased robustly in response to skin cooling (Fig. 2A). The threshold skin temperature to activate iBAT SNA (37.3 ± 0.5°C; range 35.0–38.6°C; n = 7) was significantly lower than that to activate tail SNA (40.1 ± 0.4°C) (P < 0.01, n = 7). iBAT SNA returned to baseline each time the skin was rewarmed, at least until core temperature had fallen substantially (Fig. 2, A and B). The response showed hysteresis (Fig. 2B,b), with the threshold skin temperature for disappearance on rewarming being 36.3 ± 0.4°C (n = 7). Late in the cooling sequence, a modest level of iBAT SNA remained after the skin had rewarmed in five of the seven rats: this response component was attributed to the low core temperature (28). In two rats, no core temperature-related component of iBAT activity was detectable, even though core temperature fell below 36.5°C. In the five rats where it was present, the mean threshold rectal temperature for this iBAT SNA response was 36.9 ± 0.1°C (range 36.5–37.1°C), which was significantly lower than the corresponding value for tail SNA in the same animals (38.2 ± 0.2°C; P < 0.01, n = 5).
In each rat, we calculated for both tail and iBAT SNA, the relative sensitivity (K-ratio) to K-core compared with K-skin (28). The K-ratio (K-core/K-skin) was significantly lower for iBAT SNA (3.9 ± 1.1, n = 5) compared with tail SNA (13.9 ± 3.9; P < 0.01, n = 5).
All spectral analysis was performed on records taken during periods when both iBAT and tail nerves were active (i.e., during cooling). Representative data from one animal are shown in Fig. 3. The iBAT SNA autospectrum showed a sharp peak around 1 Hz and a broad peak centered around 5 Hz (2–8 Hz in different animals) (Fig. 3A,a). The tail SNA autospectrum also showed a sharp peak around 1 Hz, but above that frequency there was low power and no clear peak (cf. Ref. 37) (Fig. 3A,c). Because the ∼1-Hz peaks for both nerves corresponded to the frequency of artificial ventilation (reflected in the autospectrum of airway pressure, Fig. 3A,d), we used coherence analysis to test whether those 1-Hz peaks were attributable to modulation by the ventilatory cycle. This revealed that there was significant coherence between airway pressure, and both nerve activities in five of seven rats. In the five significant cases, the mean coherence with airway pressure at 1 Hz (1.0 ± 0.1 Hz) was 0.63 ± 0.15 for iBAT SNA and 0.47 ± 0.12 for tail SNA (Fig. 3B, c and d). Coherence analysis also revealed a significant relation of iBAT SNA to the cardiac cycle (7.6 ± 0.1 Hz) in six of seven rats (mean coherence, 0.33 ± 0.05, n = 6; e.g., Fig. 3B,a), confirming published findings (17), but this was the case for tail SNA in only two of seven rats (coherence 0.36 ± 0.10, n = 2).
To test whether common frequencies were present in iBAT and tail SNA, we performed coherence analysis on their spectra (Fig. 3C). Significant coherence was present in four of seven rats (mean value 0.45 ± 0.13), as a narrow peak at 1.0 ± 0.1 Hz. To test whether that coherence could be accounted for by a common input from the ventilator, we used partial coherence analysis. The coherence at that frequency between iBAT and tail SNA spectra was removed (from 0.45 ± 0.1 to 0.12 ± 0.06, n = 4) by partialization with respect to the airway pressure signal (Fig. 3D).
These first simultaneous recordings of activity from sympathetic nerves supplying iBAT and tail skin in rats revealed important differences in the ways the central nervous system controls these two neural outflows. The results from nerve responses to cold exposure demonstrate that skin and BAT sympathetic outflows have unequal sensitivities to skin and core temperatures and that their relative responsiveness to those two stimuli are quite disparate. Furthermore, spectral and partial coherence analysis showed that two quite distinct rhythmic activity patterns characterize the activity of the two neural outflows; all they appear to share is a degree of modulation by the ventilatory cycle. Both qualitative and quantitative differences thus exist between the activity patterns of tail and iBAT SNA, suggesting that they are controlled by independent central neural pathways.
The ventilatory modulation of tail SNA could have been due to a direct reflex action of pulmonary afferents, or secondary to entrainment by the ventilator of central respiratory drive (31), or of an intrinsically rhythmic firing pattern of tail vasoconstrictor neurons (9). Respiratory modulation of iBAT activity has been described by Morrison (17), although under his experimental conditions, this was only evident in vagotomized rats. In our experiments, however, rats with intact vagi showed ventilatory modulation of BAT SNA, presumably because of subtle differences in the experimental preparation. We cannot distinguish whether this was because of direct reflex modulation by lung inflation receptors or because of central respiratory drive, which is normally entrained to the ventilatory cycle in rats with intact vagi (31). The significance of finding respiratory-related modulation of both of these cold-defense neural pathways remains unclear.
In the present study, we showed that iBAT SNA responded much more readily to skin cooling than to core cooling. This broadly agrees with recent indirect observations by Osaka (27), who found that oxygen consumption in anesthetized rats increased in response to skin cooling but not core cooling. The skin cooling-induced increased metabolism was strongly attenuated by propranolol, indicating that it was mostly attributable to BAT activity. Our present data extend this observation by showing directly that the neural outflow to iBAT is indeed increased by skin cooling, but it can also be increased by core cooling if the temperature is allowed to fall sufficiently far. Osaka may not have seen the latter effect because of the confounding effect of cooling on overall body metabolism and/or because he did not cool the animal's core temperature below the threshold for iBAT activation. Another study by Morrison and colleagues (18) showed a strong relationship of iBAT activity with core temperature during cooling. In those experiments, however, the skin was also cooled, and experiments were not designed to separate the two effects. Overall, these studies indicate that skin temperature, not core temperature, is the dominant stimulus for BAT activation in the cold. The threshold measurements in the present study indicate that under natural conditions, with mean skin temperatures nearly always below 35°C, both the iBAT and tail sympathetic supplies would be tonically active.
Thermoregulatory responses to cold may be considered in three main categories: behavioral thermoregulation, cutaneous (including tail) vasoconstriction, and metabolic thermogenesis (shivering and nonshivering). Under general anesthesia, the latter two can occur, albeit less strongly than without anesthesia (35, 36), but only indirect data were available about the recruitment of sympathetic thermogenic and vasoconstrictor pathways. The present study shows clearly that, as skin and core temperatures fell, the neural drive for heat conservation (tail vasoconstriction) was recruited earlier than that for metabolic thermogenesis. Reducing heat loss before increasing heat production is a good way to economize on energy use. Indeed, when the energy supply is scarce (after starvation), rats may further lower their threshold temperature for thermogenesis without altering the tail vasomotor threshold (34). Such differential changes in thermogenic vs. vasoconstrictor thresholds are difficult to explain unless the neural pathways for these two functions are at least partly separate. Analogous considerations apply to the fact that iBAT and tail SNA participate in different physiological responses: tail SNA is influenced by cardiovascular factors, such as arterial baroreceptors (28), which do not affect iBAT SNA [at least the high-amplitude bursting activity that reflects the neural input to brown adipocytes (17)], whereas iBAT SNA may be selectively involved in metabolic responses to food intake (33).
There are, nevertheless, similarities in the descending central pathways controlling iBAT and tail SNA. Physiological studies have shown that local activation or disinhibition of raphé neurons activates both the iBAT and tail vasoconstrictor pathways (3, 17, 30, 41); moreover, inhibition of raphé neurons can block the excitatory responses of both pathways to cooling (15, 24, 25) and to experimental fever (12, 16). Anatomical studies show that medullary raphé neurons are amongst the first brain stem cell groups to be labeled after injections of transneuronal retrograde tracer into the tail (21, 38) or into the iBAT (7, 21), and in both cases, the labeled medullary raphé neurons express the same glutamate transporter (vesicular glutamate transporter 3) (21). Together, these findings indicate that premotor neurons for the sympathetic outflow to iBAT and those for the tail vessels are located in the medullary raphé. The question then arises as to whether the same raphé premotor neurons drive both these sympathetic outflows. The present study provides functional evidence to address that question.
Could the present and previous findings be reconciled with the unifying hypothesis that common neural pathways drive BAT and tail SNA, at least for thermoregulation? First, one would need to invoke additional, selectively stronger, connections from skin temperature to the BAT pathway and/or from core temperature to the tail pathway. Where might the skin and core temperature signals be integrated? A recent study by Osaka (26) shows that neurons in the preoptic area are critically involved in the stimulatory effect of skin cooling on nonshivering thermogenesis, because that response to skin cooling could be entirely blocked by microinjections the GABAA receptor antagonist bicuculline into the preoptic area. This same preoptic region contains the intrinsically temperature-sensitive neurons that are thought to drive thermoregulatory responses such as cutaneous vasoconstriction (5), and indeed, direct preoptic warming strongly inhibits tail SNA (28, 41). It is not yet known whether the effect of skin cooling on tail SNA depends on preoptic neurons.
The most likely interpretation of the above facts is that the pathways regulating iBAT and tail sympathetic outflows in the cold are independent of each other, from the preoptic area downward. We may infer that the preoptic neurons regulating tail SNA show greater intrinsic thermosensitivity than those regulating iBAT SNA.
Recent findings have changed our concept of how thermoregulation is organized in the brain. Specifically, the idea of a unitary controller in the anterior hypothalamus/preoptic area has been progressively replaced by evidence for independent thermosensitve pathways controlling different effector mechanisms (20). The present results support this model by providing evidence for the independent control of two further thermoregulatory mechanisms.
The medullary raphé contains not only the sympathetic premotor neurons for tail vasoconstriction and iBAT activation but also premotor neurons for the cardiac sympathetic supply (8, 43) and for the vagal control of gastrointestinal motility and secretion (39). These autonomic outflows all appear to be activated by cold exposure (3, 4, 17, 23–26, 28, 40). Our experience comparing the tail and iBAT outflows suggests that their pathways share no common drive beyond the ventilatory cycle, and thus they probably do not “crosstalk” in the raphé. A challenge for future research is to understand the significance of why these and other efferent pathways for cold defense [e.g., the fusimotor response to skin cooling (42)] should all have a critical synaptic relay in the same brain stem nucleus.
We thank the National Health and Medical Research Council of Australia for supporting this work. Dr. Ootsuka held an International Society for Hypertension Postdoctoral Fellowship from the Foundation for High Blood Pressure Research.
We are grateful to David Trevaks for his expert help with technical and computing aspects of this study.
Present address of Y. Ootsuka: Dept. of Human Physiology, School of Medicine, Flinders Univ., Bedford Park 5042, SA, Australia.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society