our understanding of how the brain regulates body temperature is currently at a fascinating stage, at which new findings about specific brain pathways are being integrated with long-established information on how the system behaves as a whole—filling in the details of the “black box.”
Neurons in the anterior hypothalamus/preoptic region (AH/PO) are considered to play a key role and to be the major (though not exclusive) site where body core temperature is sensed for thermoregulation (2, 19, 30). This view is based on classical experiments, which have shown that direct warming and cooling of the AH/PO cause compensatory thermoregulatory responses and that lesions of the AH/PO disable temperature regulatory mechanisms. Peripheral (skin) temperature signals also provide important inputs for thermoregulation, especially on the “cold” side (19, 30). Compared with core temperature, however, there has been much less agreement about where in the CNS skin temperature signals are organized and integrated; indeed, this topic has a checkered history. On the one hand, Boulant and others (2–4) have long maintained that peripheral temperature signals are integrated with those of intrinsically temperature-sensitive neurons in the AH/PO. This information is based on a number of studies performed in the 1960s and 1970s where single neuron activity was recorded from the AH/PO region in intact animals. Many neurons that responded to core body temperature also received sensory inputs from the skin: those sensory inputs included, but were not restricted to, cutaneous temperature (4). On the other hand, skeptics have pointed out that the experiments underpinning this view do not take account of a major confounding factor, which was discovered later: changing skin temperature is an “arousing” stimulus. In fact, it turns out that many central neurons, which apparently respond to changing skin temperature (initially established for the rat's scrotum, but later generalized to other skin areas), are, in fact, responding to changes in the animal's EEG state (arousal) rather than to skin temperature per se (12). Clearly, the evidence needed to be reassessed in the light of this new knowledge. When later investigations set out to do this, most central neurons, which initially appeared to be responding to skin temperature, failed to do so after anesthesia was manipulated to prevent transitions between EEG states (1, 11). [Thermoregulatory responses to skin temperature persist even when EEG state does not change, as do the responses of putative premotor neurons in the medullary raphé (28).] Indeed, the apparent absence of neurons in the AH/PO responding to skin temperature independently of EEG state famously led Berner and Heller (1) to doubt whether skin temperature signals actually reached that brain region.
Recently, experiments of a different type have helped to allay doubts about the AH/PO's central role in processing skin temperature signals, at least for the control of nonshivering thermogenesis by brown adipose tissue (BAT). Toshimasa Osaka showed that nonshivering thermogenesis (increased metabolic rate) could be induced by mild cooling of the flank skin in anesthetized rats (26) and then demonstrated that this metabolic surge in response to skin cooling could be entirely blocked by bilateral injections of the GABA antagonist, bicuculline, into a specific locus within the AH/PO, which he termed the dorsolateral preoptic area (25). (GABA agonists injected into the same region initiated thermogenesis). The cold-induced metabolic surge was also blocked by systemic administration of the nonselective β-adrenoreceptor antagonist, propranolol, indicating that it was due to BAT (5), although a metabolic contribution from the accompanying tachycardia (25), also blocked by propranolol, could not be excluded. These findings have now been confirmed and extended by an elegant study published in this issue of American Journal of Physiology—Regulatory, Integrative and Comparative Physiology.
Kazuhiro Nakamura and Shaun Morrison (20) first confirmed Osaka's important conclusion, by directly recording sympathetic thermogenic activity from the nerve supplying interscapular BAT (iBAT) sympathetic nerve activity (SNA). As predicted, injection of bicuculline into the dorsolateral preoptic area completely prevented the stimulant action of skin cooling on iBAT SNA (20). They then went on to determine the efferent pathway of this reflex as it travels caudally through the neuraxis.
It is well known that heat conservation responses in the cold are very similar to heat gain responses in experimental fever; indeed the “raised set point” explanation of fever postulates that they use the same neural circuitry (2, 15). It should be stated, however, that no critical evidence yet proves that this view is correct. Shaun Morrison and colleagues have previously delineated the efferent neural pathways to iBAT SNA in response to experimental fever after the injection of PGE2 into the AH/PO: critical synaptic relays in the excitatory drive pathway to iBAT have been identified in the dorsomedial hypothalamic area (DMH) area and the medullary raphé (16–18, 21). Now we can ask directly, how closely do the efferent pathways to iBAT for cold defense resemble those for fever? By microinjecting the GABAA agonist muscimol or other inhibitory agents, Nakamura and Morrison now confirm directly that the efferent pathway for iBAT activation by skin cooling also appears to involve critical synaptic relays in the DMH and the medullary raphé. Although this does not prove that the pathways for cold defense and fever are identical, any differences between them are too small to show up clearly with the methods used.
A less straightforward finding of Nakamura and Morrison's study, however, was the lack of any support for the involvement of the caudal midbrain periaqueductal gray (PAG) in the neural pathway from cool skin to iBAT. The results were very clear: inhibition of neither the dorsolateral nor ventrolateral regions of the caudal PAG had any effect on baseline parameters or on the iBAT SNA response to skin cooling. This was something of a surprise, because previous studies had shown that stimulation or disinhibition of caudal PAG neurons can cause BAT thermogenesis (7, 29, 32) and, critically, de Menezes and colleagues found that inhibition of the dorsolateral column of the caudal PAG (an apparently identical region) caused a fall in baseline body core temperature and an attenuation of the hyperthermic response to stimulation of the DMH in conscious rats (8). How may we resolve this discrepancy? Is the caudal PAG part of the descending neural pathway for the iBAT SNA response to a cool environment, or is it not? The reason for the discrepancy may lie in two factors: first, it seems likely that the relevant PAG neurons were tonically active in the conscious animals used by de Menezes et al. (8) (because neuronal inhibition alone caused baseline changes) but silent in the anesthetized animals used by Nakamura and Morrison (20) (in which neuronal inhibition did nothing). Second, there is no reason to expect that the interactions between the pathway from the caudal PAG and other excitatory pathways to iBAT SNA should be linear. Nonlinear interactions, such that the removal of one input lowers the gain of another input to the same final output pathway may be the norm for sympathetic outflows (e.g. 6, 24, 32). Although not being part of the AH/PO → DMH → raphé pathway, tonic drive from the caudal PAG could still interact with it further down the efferent pathway to iBAT (perhaps in the medullary raphé) in a nonlinear fashion; removal of tonic facilitatory drive from the PAG could then reduce the gain of the downstream sections of the pathway to iBAT. As a parenthetic point, we all need to be aware that such nonlinear interactions complicate the interpretation of experiments: if inhibition of a brain region blocks a reflex output, it does not necessarily mean that that brain region forms an integral stage of the reflex pathway, only that its participation is necessary for the process to work (e.g., 24).
Finally, do the lessons learned about the descending thermoregulatory pathway from AH/PO to iBAT tell us anything about the neural pathways of other cold defense mechanisms such as cutaneous vasoconstriction? We do not yet know. It is well established that the cutaneous vasomotor tone is held under tonic inhibition from warm-sensitive neurons in the AH/PO (10, 13, 27, 31) and that cutaneous vasoconstrictor pathways also appear to involve a critical synaptic relay in the medullary raphé (14, 22, 31). Despite their anatomical similarities, however, it is likely that the neural pathways regulating iBAT and those regulating cutaneous vessels are not the same. The control of cutaneous vasomotor outflows, such as the supply to the rat's tail is dominated by core temperature, although skin temperature provides a moderately strong additional input (27). By contrast, iBAT sympathetic activity is hardly interested in core temperature (20, 23, 26) but is strongly driven by cold signals from the skin (20, 23, 26). Moreover, the rhythmic activity patterns in the neural outflows to the tail and iBAT are also quite distinct, indicating that they are driven by different neural pathways (23). It will be interesting to see how future studies define these different thermoregulatory neural pathways. In particular, the involvement or not of critical synaptic relays in the DMH (9) and PAG needs to be established. The clear picture constructed from the work of Nakamura and Morrison (20) provides an excellent template with which to compare future investigations of other thermoregulatory mechanisms.
Work in the author's laboratory is supported by the National Health and Medical Research Council, the National Heart Foundation, and the Australian Research Council.
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 © 2007 the American Physiological Society