- Top of page
It is known that brain noradrenaline (norepinephrine) mediates fever, but the neuronal group involved is unknown. We studied the role of the major noradrenergic nucleus, the locus coeruleus (LC), in lipopolysaccharide (LPS)-induced fever. Male Wistar rats had their LC completely ablated electrolytically or their catecholaminergic LC neurones selectively lesioned by microinjection of 6-hydroxydopamine; the controls were sham-operated. Both lesions resulted in a marked attenuation of LPS (1 or 10 μg kg−1, i.v.) fever at a subneutral (23°C) ambient temperature (Ta). Because electrolytic and chemical lesions produced similar effects, the role of the LC in fever was further investigated using electrolytic lesions only. The levels of prostaglandin (PG) E2, the terminal mediator of fever, were equally raised in the anteroventral third ventricular region of LC-lesioned and sham-operated rats during the course of LPS fever, indicating that LC neurones are not involved in febrigenic signalling to the brain. To investigate the potential involvement of the LC in an efferent thermoregulatory neuronal pathway, the thermoregulatory response to PGE2 (25 ng, i.c.v.) was studied at a subneutral (23°C, when fever is brought about by thermogenesis) or neutral (28°C, when fever is brought about by tail skin vasoconstriction) Ta. The PGE2-induced increases in metabolic rate (an index of thermogenesis) and fever were attenuated in LC-lesioned rats at 23°C, whereas PGE2-induced skin vasoconstriction and fever normally developed in LC-lesioned rats at 28°C. The LC-lesioned rats had attenuated PGE2 thermogenesis despite the fact that they were fully capable of activating thermogenesis in response to noradrenaline and cold exposure. It is concluded that LC neurones are part of a neuronal network that is specifically activated by PGE2 to increase thermogenesis and produce fever.
The brain plays a central role in coordinating several responses to systemic inflammation, including fever. It transduces the febrigenic signals from the periphery into appropriate adjustments of thermoeffector activity to ultimately increase body core temperature (Tc). Signals from the periphery may gain access to the brain by three routes: (1) via afferent fibres that travel mostly through the vagus nerve and make their first synapse in the nucleus of the solitary tract (Blatteis et al. 2000; Romanovsky et al. 2000); (2) via circumventricular organs, such as the organum vasculosum laminae terminalis and the subfornical organ, which lack a blood–brain barrier (Takahashi et al. 1997; Blatteis et al. 2000); and (3) via interaction with cells located in the blood–brain interface, i.e. endothelial (Tilders et al. 1994; Cao et al. 1996) and perivascular (Elmquist et al. 1997; Schiltz & Sawchenko, 2003) cells. Activation of these afferent pathways ultimately increases the level of PGE2 in the brain (Blatteis & Sehic, 1997b). By interacting with EP3 receptors (Ushikubi et al. 1998; Oka et al. 2003) and consequently decreasing the intracellular level of cyclic AMP in preoptic neurones (Steiner et al. 2002; Steiner & Branco, 2003), PGE2 triggers an appropriate thermoeffector response to increase Tc. The preoptic region sends both inhibitory and excitatory efferent projections to nuclei located in the hypothalamus, midbrain, and brainstem that are involved in the control of thermoeffector activity (Nagashima et al. 2000).
There is strong evidence supporting an involvement of brain noradrenaline in the genesis of fever. First, systemic administration of exogenous and endogenous pyrogens causes a rapid activation of noradrenergic terminals in the preoptic region (Dunn & Wang, 1995; Linthorst et al. 1995). Second, both in vitro (Malik & Sehic, 1990) and in vivo (Sehic et al. 1996) studies have shown that noradrenaline triggers the release of the terminal mediator of fever, PGE2 (Blatteis & Sehic, 1997a). Third, and most important, depletion of brain catecholamines using the neurotoxin 6-hydroxydopamine abolishes the fever induced by interleukin-1 in rats (Ovadia et al. 1989). However, the noradrenergic neuronal group involved in fever remains unknown.
The locus coeruleus (LC), a well-delineated cluster of noradrenaline-containing neurones located adjacent to the fourth ventricle in the pontine brainstem (Berridge & Waterhouse, 2003), is the major noradrenergic nucleus in the brain. It is estimated that ∼50% of all of the noradrenergic projections in the central nervous system originate in the LC (Aston-Jones et al. 1995; Berridge & Waterhouse, 2003). Hence, it is reasonable to hypothesize that LC neurones play a major role in the development of fever. Because of its strategic location, the LC may convey febrigenic messages to the preoptic region and/or control the activity of thermoeffector neuronal pathways. For example, by receiving inputs from the nucleus of the solitary tract (Van Bockstaele et al. 1999) and projecting to the preoptic region (Aston-Jones et al. 1995), LC neurones may be involved in febrigenic signalling via the vagus nerve. Yet, the LC neurones may be involved in the control of thermoeffector (thermogenesis and/or heat loss) activity because they receive inputs from the preoptic region (Steininger et al. 2001) and project to brain regions involved in the control of thermoeffector activity (Jones & Yang, 1985; Grzanna & Fritschy, 1991; Morrison et al. 1999; Uno & Shibata, 2001). To date, however, the involvement of the LC in the febrile response has not been addressed. In the present study, we addressed the role of the LC in the fever induced by bacterial lipopolysaccharide (LPS) in rats.
- Top of page
In the present study, we have used both electrolytic ablation of the LC and selective chemical lesion of LC catecholaminergic neurones to verify the possible involvement of this nucleus in the fever induced by LPS. Our data show that lesions of the LC resulted in an attenuation of LPS-induced fever at a subneutral Ta, even when catecholaminergic neurones, but not fibres of passage, were selectively lesioned by microinjection of 6-OHDA. This finding strongly supports a role of catecholaminergic LC neurones in the genesis of fever. Next, we investigated such a role in more detail. To this end, we used the practical and reproducible electrolytic lesioning approach, since both electrolytic and chemical lesions produced very similar – if not identical – effects on LPS fever.
Neuroanatomical evidence suggests that febrigenic signals from vagal afferents arriving in the nucleus of the tractus solitarius (NTS) may reach the preoptic region through a relay in the LC (Aston Jones et al. 1995; Van Bockstaele et al. 1999). Hence, it is possible that the LC is involved in febrigenic signalling via the vagus nerve. This possibility is, however, disproved in the present study by the following observations. First, lesions of the LC attenuated not only the monophasic (vagus-dependent; Romanovsky et al. 1997b) fever induced by a low dose of LPS, but also the long-lasting (vagus-independent; Romanovsky et al. 1997b) fever induced by a moderate dose of LPS. Second, lesion of the LC did not affect the increased production of PGE2 (the terminal mediator of fever) in the AV3V region during the course of LPS-induced fever, irrespectively of the LPS dose.
Another possibility is that the LC is part of a thermoeffector neuronal pathway activated by pyrogens. To address this possibility, we verified whether PGE2 (which acts directly on the preoptic region to activate febrigenic thermoeffector pathways; Blatteis & Sehic, 1997b) was able to produce fever in LC-lesioned rats. Remarkably, we found that PGE2 fever is substantially attenuated in LC-lesioned rats at a subneutral but not at a neutral Ta. In agreement with previous studies (Stitt, 1973; Crawshaw & Stitt, 1975), we observed that PGE2 fever was brought about by enhanced metabolic heat production (increase in oxygen consumption) at a subneutral Ta and by skin vasoconstriction (decrease in HLI) at a neutral Ta. At a subneutral Ta, the attenuated fevers of LC-lesioned rats were associated with a strong blockade of the PGE2-induced activation of metabolic heat production. At thermoneutrality, on the other hand, LC-lesioned rats showed marked tail skin vasoconstriction and were able to mount a normal febrile response to PGE2.
These results suggest the involvement of the LC in a thermogenic neuronal pathway, but an alternative explanation is also possible. The impaired capacity of LC-lesioned rats to develop increased thermogenesis and fever in response to LPS and PGE2 may be due to an atrophy of thermogenic tissues. Indeed, an impaired thermogenesis induced by malnutrition has been quoted as an important ‘side-effect’ of lesion (Shido et al. 1989) and denervation (Romanovsky et al. 1997a) experiments. To assess the thermogenic capacity of the rats, we verified the effects of i.v. noradrenaline on the metabolic rate (assessed by the rate of oxygen consumption) and Tc. This method is based on the fact that noradrenaline directly activates thermogenesis in the brown adipose tissue, the main thermogenic organ in rats (Cannon & Nedergaard, 1998). It was successfully employed in previous studies (Hayashi & Nagasaka, 1983; Gordon, 2000). That LC-lesioned rats responded to noradrenaline with increases in metabolic rate and Tc identical to those of sham-operated rats indicates that the thermogenic reserves of LC-lesioned rats were intact. Consequently, impaired thermogenic and febrile responses of LC-lesioned rats to LPS and PGE2 in a subthermoneutral environment can be ascribed to the interruption of a specific fever-inducing thermogenic neuronal pathway rather than atrophy of thermogenic tissues. Consistent with such a role of the LC are the studies showing that the number of neurones expressing c-fos (a marker of neuronal activation) increase in the LC after LPS administration (Hare et al. 1995; Xu et al. 2003). Moreover, neurochemical and electrical activation of LC neurones occur when febrigenic substances, i.e. LPS (Molina-Holgado & Guaza, 1996) and corticotropin-releasing factor (Valentino et al. 1983; Emoto et al. 1993), are administered at regular laboratory temperatures, which are subneutral for rats.
Although lesion of the LC markedly attenuated LPS- and PGE2-induced thermogenesis and fever, it had no effect on cold-induced thermogenesis. In agreement, studies by Cano et al. (2003) and Bachtell et al. (2003) revealed no activation of the LC (assessed by c-fos expression) during exposure to cold. Only one study (Kiyohara et al. 1995) showed cold-induced c-fos expression in the LC, but this effect may be unrelated to the activation of thermogenesis; it may be related to the activation of the stress response to a new environment. Indeed, a recent study showed that cold-induced c-fos expression in LC neurones disappears in rats systematically adapted to transfer to a cold environment, despite the fact that adapted rats still respond to cold with enhanced thermogenesis (Bachtell et al. 2003). It thus seems that catecholaminergic LC neurones are specifically involved in the activation of thermogenesis by pyrogens.
Further investigation is needed to understand the specific functional connections of the LC with the neuronal circuitry involved in the control of thermogenesis, but speculative scenarios can be proposed. Based on the results of their recent study using pseudorabies virus (a retrograde transynaptic tracer), Cano et al. (2003) have proposed that the LC is multisynaptically connected to the brown adipose tissue. This finding agrees with previous studies showing that there is no direct projection from the LC to the intermediolateral cell column of the spinal cord (Grzanna & Fritschy, 1991; Proudfit & Clark, 1991), where preganglionic sympathetic neurones involved in brown adipose tissue activation are located (Zhang et al. 2000; Cano et al. 2003). It also agrees with the facts that LC neurones receive projections from the preoptic region, either directly (Steininger et al. 2001) or via relays in the paraventricular hypothalamus (Aston-Jones et al. 1991), and project to several brainstem structures involved in the control of brown adipose tissue thermogenesis, e.g. the inferior olive (Grzanna & Fritschy, 1991; Uno & Shibata, 2001) and the raphe pallidus (Jones & Yang, 1985; Morrison et al. 1999).
In summary, the present study shows that catecholaminergic LC neurones are part of a thermoeffector neuronal pathway that is specifically activated by pyrogens (e.g. PGE2) to activate thermogenesis and produce fever in a subthermoneutral environment.