How circulating cytokines trigger the neural circuits that control the hypothalamic–pituitary–adrenal axis

Psychoneuroendocrinology 26 (2001) 761–788 www.elsevier.com/locate/psyneuen

2001 Curt P. Richter award

How circulating cytokines trigger the neural circuits that control the hypothalamic–pituitary– adrenal axis Serge Rivest* Laboratory of Molecular Endocrinology, CHUL Research Center and Laval University, 2705, boul. Laurier, Que´bec, Canada, G1V 4G2 Received 2 July 2001

Abstract It is now no secret that the brain plays a crucial role in organizing, adapting and restraining the systemic inflammatory response via a complex cascade of mechanisms involving proteins of the innate immune system, molecules of the proinflammatory signal transduction pathways, prostaglandins (PGs) and specific populations of neurons. These neuronal circuits, in particular those controlling autonomic functions, are all together involved in engaging the physiological responses that may help eliminating the foreign material and adjust the inflammatory events to prevent detrimental consequences. For instance, elevation in plasma glucocorticoid levels is one of the most powerful endogenous and well-controlled feedback on the pro-inflammatory signal transduction machinery taking place across the organisms. The main Center that controls this neuroendocrine system is the paraventricular nucleus of the hypothalamus (PVN) that receives neuronal projections from numerous hypothalamic and extra-hypothalamic nuclei and areas. There is now compelling evidence that molecules produced by cells of the blood–brain barrier (BBB) may bind to their cognate receptors expressed at the surface of neurons that are responsible to trigger the hypothalamic–pituitary adrenal (HPA) axis. This review presents the new molecular insights regarding the pro-inflammatory signal transduction pathways that occur in these cells and how they are related to the neuroendocrine circuits mediating the increase in plasma glucocorticoid levels during systemic and localized immunogenic insults.  2001 Elsevier Science Ltd. All rights reserved. Keywords: Systemic inflammation; Cytokines; Prostaglandins; Signal transduction; Nuclear factor kappa

* Tel.: +1-418-654-2296; fax: +1-418-654-2761. E-mail address: [email protected] (S. Rivest). 0306-4530/01/$ - see front matter  2001 Elsevier Science Ltd. All rights reserved. PII: S 0 3 0 6 - 4 5 3 0 ( 0 1 ) 0 0 0 6 4 - 6

762

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

B; Mitogen-activated proteins; Endothelial cells; Blood–brain barrier; Glucocorticoids; Transcription factors; Corticotropin-releasing factor; Neuroendocrinology

1. Introduction The rapid production of cytokines, chemokines and prostaglandins is an essential feature of the innate immune response. These proteins belong to the superfamily of pro-inflammatory molecules that are largely responsible for most of the autonomic/neuroendocrine changes that occur during the acute-phase response of all types of aggression. Tumour necrosis factor alpha (TNF-α), interleukin-1 (IL-1α and ß), IL-6, IL-10, IL-12, IL-15, IL-18 and type I interferons (IFN-α and IFN-ß) are cytokines that belong to the innate immune system, although they are not all proinflammatory (IL-10 is anti-inflammatory) and only a few of them are generally recognized to be potent modulators of the neuroendocrine system. Among them, TNF-α, IL-1ß and IL-6 are rapidly induced in response to foreign material and they can circulate into the bloodstream to act on distant organs.

2. TNF-a TNF-α is a pleiotropic cytokine originally recognized for its anti-tumour activity (Carswell et al., 1975), but now referred to as a pro-inflammatory factor that plays a central role in initiating the cascade of other cytokines all together involved in a timely controlled immune response to infection. This cytokine is produced by a variety of cell types including neutrophils, activated monocyte/macrophages (Miller et al., 1993), T and B lymphocytes (Goldfeld et al., 1991), NK cells, astrocytes (Chung and Benveniste, 1990), mast cells, endothelial cells (Wilkinson and Edwards, 1991), smooth muscle cells, synovial cells (Brennan et al., 1992), brain ependymal cells (Liu et al., 1996) and microglia (Nadeau and Rivest 1999a, 2000). Macrophages and NK cells remain nevertheless the major cell source, at least in the periphery. The endotoxin lipopolysaccharide (LPS) and other inflammatory agents cause a rapid production of TNF-α both from in vitro and in vivo phagocytic cells and a subsequent increase of other pro-inflammatory cytokines, such as IL-1ß and IL-6. TNF plays important roles in a wide variety of events, namely septic shock, cell proliferation and apoptosis that can be mediated by either one of the TNF receptors (p55 and p75), both of which belong to the TNF receptor superfamily. The binding of TNF-α to its cognate receptors leads to the formation of the TNFR1-associated death domain (TRADD)/TNF receptor-associated factor 2 (TRAF2) complex, which activates the nuclear factor kappa B (NF-␬B) signalling events (Fig. 1). TNF-α is actually one of the most potent effectors of NF-␬B activity through the 55 kd TNF type I receptor in most types of cells in the periphery as well as in the CNS (Laflamme and Rivest, 1999). FAS-associated death domain protein (FADD)/MORT1, TRAF2 and the dead domain kinase receptor interacting protein

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

763

Fig. 1. Proximal components of the TNF receptor type I signal transduction pathways leading to apoptosis, activation of the nuclear factor kappa B (NF-␬B) cascade or induction of the mitogen-activated protein (MAP) kinases. See text for details and abbreviations.

(RIP) are recruited and may also interact directly with TRADD (see Fig. 1). While FADD/MORT1 is essential for TNF-induced apoptosis, RIP and TRAF2 seem to be the key molecules for activating both NF-␬B and c-Jun N-terminal kinase (JNK). This recruitment leads to activation of the protein kinase IKK that is composed of two catalytic subunits, IKKα and IKKß, and one regulatory subunit, IKKγ/NEMO (Chu et al., 1999). IKKγ/NEMO is required for the activation of NF-␬B by many stimuli, including TNF, although only the IKKß catalytic subunit is essential to trigger NF-␬B in response to the cytokine. Indeed, IKKß was recently found to be the target for upstream signals generated by pro-inflammatory stimuli resulting in I␬B phosphorylation at serine 32/36 and degradation by proteasome after polyubiquitination (Delhase et al., 1999; Hu et al., 1999). This frees NF-␬B and allows its translocation into the nucleus and the subsequent activation of target genes (Fig. 2). One of the best known immune function of TNF is to stimulate the recruitment of neutrophils and monocytes to the site of infection and to activate these cells in a paracrine manner to eliminate the foreign material. This mechanism evolves members of the chemokine family that play a primary role in the recruitment of inflammatory cells in different tissues during acute and chronic inflammation. The monocyte chemoattractant protein-1 (MCP-1) is one of the principal CC chemokines that is rapidly synthesized by endothelial cells in response to leukocyte-derived TNF and IL-1ß for allowing the monocyte recruitment and rolling prior to transmigration and arrival at the site of inflammation. This transcriptional process is under the direct control on the previously described cascade of events, i.e. TRADD/TRAF2/RIP/IKK/I␬Bα/NF-␬B. During severe infections, TNF is produced in large concentrations and was found to be involved in systemic clinical and patho-

764

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

Fig. 2. The pro-inflammatory signal transduction pathways evolving the nuclear factor kappa B (NF␬B). p50 and p65 are the two most common DNA-binding subunits of the NF-␬B dimer that have the ability to trigger the transcription of numerous target genes. See text for details and abbreviations.

logic abnormalities, such as fall in blood pressure and shock. The septic shock is a complication of severe Gram-negative bacterial sepsis associated with vascular colapse, disseminated intravascular coagulation and profound metabolic disturbances. Although TNF was believed to be responsible for LPS-induced septic shock, numerous experiments failed to provide the evidence that this syndrome was entirely attributed to the high circulating levels of this pro-inflammatory cytokine. However, the circulating levels of TNF may be predictive of the outcomes of severe Gram-negative infection. At lower levels, the cytokine was found to be involved in various autonomic and endocrine changes, including fever and increase in the HPA axis (see below).

3. Interleukin-1 This pro-inflammatory cytokine (especially the ß form) is probably the most important molecule capable of modulating cerebral functions during systemic and localized inflammatory insults. There are two forms of IL-1, IL-1α and IL-1ß, that share less than 30% homology and bind to the same receptors. This explains the original thought that both molecules had similar biological activities, although the alpha form is no longer believed to be a potent player in the neural-immune interac-

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

765

tion. Actually, the IL-1 family consists of three different genes located in human on the long arm chromosome 2 (Webb et al., 1986), which encode three distinct proteins with structural homologies. IL-1α and IL-1ß act as agonist molecules as opposed to the third member of the family, IL-1 receptor antagonist (IL-1ra) that binds to the same receptors as IL-1α and IL-1ß but does not induce any intracellular signal and therefore acts as an endogenous inhibitor of the IL-1 activity. The physiological relevance of the IL-1ra in modulating the biological activity of IL-1 was confirmed in transgenic mice that produce either no IL-1ra or that overproduce the protein (Hirsch et al., 1996). Mice lacking IL-1ra have a decreased body mass compared with wild-type controls. They were more susceptible than controls to lethal endotoxemia but were less susceptible to infection with Listeria monocytogenes. Conversely, IL-1ra overproducers were protected from the lethal effects of endotoxin but were more susceptible to listeriosis. IL-1α and IL-1ß have different mechanisms of expression, synthesis and secretion. IL-1α gene does not contain sequences corresponding to the classical transcription initiation motif known as “TATA box”, whereas this motif is found in the IL-1ß gene (Shirakawa et al., 1993). The transcription of the IL-1ß gene seems to be controlled in a more complex way and the promoter region contains cAMP-response element (CRE), an element responding to LPS called NF-IL6, site analogous to the NF-␬B transcription factor and numerous other binding sites (Roux-Lombard, 1998). These observations are quite important for explaining the ability of peripheral blood monocytes to synthesize IL-1ß in response to very low concentrations of LPS, although the IL-1ß gene is not constitutively expressed in monocytes as well as other cell types. The regulation of that gene is under a sophisticated transcriptional control that is triggered by a number of immune insults, a phenomenon that is generally observed for most NF-␬B-responsive genes. Following its cleavage by a member of the caspase family called caspase-1 or IL-1ß-converting enzyme (ICE), IL-1ß has the ability to circulate in the bloodstream and act on distant organs to trigger different physiological responses in binding to its type I receptor. IL-1α, IL-1ß and IL-1ra share the same receptors, called the IL-1 receptor type I (IL-1RI), IL-1RII and IL-1 receptor accessory protein (IL-1R-AcP). IL-1RI is a glycoprotein expressed at the surface of numerous types of cells, especially on endothelial cells where it mediates the rapid induction of the chemokine MCP-1 to engage the emigration process at the site of infection. The binding of IL-1ß to its cognate type I receptor leads to the formation of the IL-1 receptor-associated kinases (IRAK)/TNF receptor-associated factor 6 (TRAF6) complex, which recruits the general adaptor protein MyD88 and activates NIK/IKK kinases involved in the phosphorylation and degradation of I␬Bα (Fig. 2). NF-␬B is then translocated into the nucleus and may bind to its ␬B consensus sequence on target genes (Baeuerle and Baltimore, 1996). 4. Interleukin-6 and gp130-related cytokines Interleukin-6 (IL-6) is one of the most pleiotropic cytokines known that is involved in regulating a wide variety of immune functions, such as B- and cytotoxic T-cell

766

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

differentiation, induction of IL-2 production and IL-2 receptor expression in T cells, T cell growth as well as the acute-phase reactions and hematopoiesis (Taga and Kishimoto, 1997; Hirano, 1998). A critical role of the pro-inflammatory cytokine in the acute-phase response was indeed reported in IL-6-deficient mice that exhibit a lower acute-phase protein (APP) synthesis than wild-type mice during systemic inflammatory insults (Kopf et al., 1994). Although the biosynthesis of the APP by hepatocytes is regulated by several factors including TNF and IL-1, it was found that IL-6 can function as the key hepatocyte stimulating factor to induce, at least in rodents, fibrinogen, cysteine proteinase inhibitor, α-2-macroglobulin and α-1-acid glycoprotein (Hirano, 1998). In addition to be produced by both systemic lymphoid and nonlymphoid cells, IL-6 protein and mRNA have also been found in specific populations of cells in the CNS during different experimental conditions. One of these stimuli is the intraperitoneal (IP) or intravenous (IV) administration of the endotoxin LPS that caused a profound transcriptional activation of the gene encoding the cytokine in the choroid plexus (chp) and the circumventricular organs (CVOs), structures devoid of blood–brain barrier (Vallie`res and Rivest, 1997). This phenomenon is of particular interest, as it provides the evidence that IL-6 may be secreted in the cerebrovascular spinal fluid (CSF) and reach its receptor subunits that are widely distributed throughout the neural tissue (Vallie`res and Rivest, 1997) to influence different neurophysiological functions (see below). The first step in the induction of the transduction signals by IL-6 is the binding of the ligand to its IL-6 receptor subunit (IL-6R), which is either located at the cell surface or present as soluble form in the liquids of the organism. The association of these two molecules with the membrane subunit gp130 forms a high affinity complex that triggers specific transduction signals (Fig. 3). The gp130 protein does not serve as signal transducer only for IL-6, but also for the ciliary neurotrophic factor (CNTF), leukemia-inhibitory factor (LIF), oncostatin M (OSM), CT-1 and IL-11 (Taga, 1997; Taga and Kishimoto, 1997). However, the actions of these pro-inflammatory cytokines are limited by the mechanisms that control their synthesis, as they are produced in a tissue-specific manner in response to different immunogenic stimuli. Three members of the janus kinase family, JAK1, JAK2 and TYK2, are closely related to gp130 and rapidly activated in presence of IL-6 (Lu¨ tticken et al., 1994; Narazaki et al., 1994; Stahl et al., 1995). These kinases phosphorylate the tyrosine residues of the gp130 cytoplasmic domain, which allows the recruitment and phosphorylation of at least two transcription factors of the signal transducers and activators of transcription family (STAT1, STAT3) and one tyrosine phosphatase (SHP-2). Once activated, the STAT proteins may activate different genes in combining their SH2 domains and forming homodimers (Leaman et al., 1996; Watanabe and Arai, 1996). Like the NF-␬B signalling that is inhibited by I␬Bα, the JAK/STAT pathway is inhibited by at least two intracellular systems to avoid exaggerated responses. The first is the internalization of the IL-6/IL-6R/gp130 complex and its degradation by specific enzymes (Rose-John et al., 1991; Nesbitt and Fuller, 1992; Zohlnhofer et al., 1992). This process does not require activation of the JAK kinases and is transduced by a degradation of the receptor at the level of the cell surface (Thiel et al., 1998). The second involves activation of the transduction signals and the de novo

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

767

Fig. 3. Interleukin 6-induced transduction signals leading to the transcription of acute-phase protein (APP). The first step in the induction of the transduction signals by IL-6 is the binding of the ligand to its IL-6 receptor subunit (IL-6R), which is either located at the cell surface or present as soluble form in the liquids of the organism. The association of these two molecules with the membrane subunit gp130 forms a high affinity complex that triggers specific transduction signals. Three members of the janus kinase family, JAK1, JAK2 and TYK2, are closely related to gp130 and rapidly activated in presence of IL-6. These kinases phosphorylate the tyrosine residues of the gp130 cytoplasmic domain, which allows the recruitment and phosphorylation of at least two transcription factors of the signal transducers and activators of transcription family (STAT1, STAT3) and one tyrosine phosphatase (SHP-2). Once activated, the STAT proteins may activate different genes in combining their SH2 domains and forming homodimers. Moreover, SHP-2 is able to activate the membrane protein Ras, which leads to the induction of the MAP kinase ERK-1 and ERK-2. Two pathways exist relating SHP-2 and Ras, one using the adapter protein Gab1 that is associated with phosphatidylinositol-3 kinase and another via the Grb2-Sos complex. The activated MAP kinases may in turn induce nuclear proteins, such as NF-IL6.

768

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

production of inhibitory proteins that prevent phosphorylation of the transcription factor STAT and activation of the MAP kinases by interacting with the catalytic domain of the JAK kinases (Aman and Leonard, 1997). These cytokine-inducible proteins named as suppressors of cytokine signalling (SOCS-1 to 7), JAK binding protein (JAB), STAT-induced STAT inhibitors (SSI) and cytokine-inducible SH2 protein (CIS) are rapidly induced by IL-6 and other members of the gp130 cytokine family (Hilton et al., 1998; Nicholson and Hilton, 1998). All these different names are now being grouped under SOCS-1 to 7 and CIS, although there are still some controversies for this unique terminology. The SOCS proteins are characterized by a highly conserved carboxyl-terminal SOCS box motif that is preceded by an SH2 domain (Hilton et al., 1998). Although SOCS-1 and SOCS-3 have potent activity for the inhibition of IL-6 signalling, the other members of the SOCS family and CIS have little or no impact (for a review, see Alexander et al., 1999). Both SOCS-1 and SOCS-3 act by preventing cytokine-dependent activation of the JAK/STAT pathway, but the intracellular mechanisms involved in these effects are quite different. SOCS-1 inhibits the intrinsic kinase activity by interacting with the catalytic domain of the JAK kinases (especially the JAK2 kinase), whereas SOCS-3 prevents IL-6 signalling at steps distal to JAK activation, i.e. recruitment of STAT factors and/or via binding to the tyrosine-phosphorylated receptor (Hansen et al., 1999). Indeed, recent studies have shown that SOCS-3 is unable to inhibit directly either JAK1 or JAK2 kinase activity (Nicholson et al., 1999).

5. How these circulating molecules talk with the brain parenchyma The previous sections described three key players that are involved in the neuralimmune interface and while both the systemic and cerebral cells can produce immune molecules, their contributions and mechanisms of action in modulating the neuroendocrine system are quite distinct. It is indeed very difficult to compare studies in which cytokines were injected systemically or centrally, because the receptors involved and the pathways are obviously very different despite a general outcome that may be similar, i.e. increase in body temperature, cardiovascular changes and stimulation of the HPA axis. The main reason for this is that cytokines that are produced by the systemic immune system are unlikely to diffuse in concentrations high enough into the cerebral tissue and they have to target cells that can be reachable from the systemic circulation. Cells forming the blood–brain barrier (BBB) are in privileged position to transfer the information from the circulation to the brain parenchyma and there are exciting new developments regarding the molecular events taking place in the endothelium of the cerebral arterioles, small capillaries and venules during systemic immune challenges. Cytokines, when secreted by cells of myeloid origin in the circulation, trigger a series of events in cascade leading to the MAP kinases/NF-␬B or the JAK/STAT transduction pathways in vascular-associated cells of the CNS. The brain blood vessels exhibit both constitutive and induced expression of receptors for TNF-α (Nadeau and Rivest, 1999b), IL-1ß (Ericsson et al., 1995; Herkenham et al., 1998) and IL-6

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

769

(Vallie`res and Rivest, 1997) that can stimulate these signalling molecules. Depending on the challenges and the cytokines involved, the transduction signal(s) solicited in cells of the BBB may orient the neuronal activity in a very specific manner in activating the transcription and production of soluble factors, such as prostaglandins (PGs).

6. PG synthesis by the cerebral microvasculature The formation of PGs is initiated by the action of cyclooxygenase (COX, also known as prostaglandin endoperoxide H-synthase or PGHS), which catalyses two separate reactions; the first being the oxygenation of arachidonic acid to the unstable PGG2 by a cyclooxygenase function and the second, the subsequent reduction of PGG2 leading to a more stable PGH2 by peroxidase reaction. Although constitutive expression of the isoform COX-1 was found in various cell types, mRNA and protein levels remain somewhat unchanged during inflammatory challenges and therefore believed to play housekeeping roles (DeWitt and Smith, 1988; DeWitt et al., 1990). In contrast, COX-2 isoform is undetectable in most tissues under basal conditions, but marked transcriptional activation can be observed in macrophages and other cell types by the endotoxin LPS and pro-inflammatory cytokines (Lee et al., 1992; Jones et al., 1993; Hempel et al., 1994). Interestingly, systemic inflammatory insults cause a robust transcriptional activation of COX-2 along cells of the mouse and rat microvasculature, although the intensity of the signal and the pattern of expression depend on the challenge and the dose of the inflammatory agents (Cao et al., 1995; Lacroix and Rivest, 1998; Quan et al., 1998; Laflamme et al., 1999; Nadeau and Rivest, 1999b). Intravenous and intraperitoneal injections of low to moderate doses of the endotoxin LPS generally induce COX-2 mRNA and protein along blood vessels of the entire brain microvasculature, choroid plexus and leptomeninges. The stimulation of PGs by cells of the BBB is not uniquely observed during endotoxemia, but in other experimental models of systemic inflammation (Fig. 4). Indeed, intramuscular (IM) turpentine insult stimulates transcription of the COX-2 gene within the microvessels, chp and leptomeninges and the signal of this transcript paralleled the inflammation of the left hind limb (Lacroix and Rivest, 1998). This experimental model induces a local tissue damage provoking sterile inflammation, i.e. an inflammatory response that develops in the absence of any microbial stimulus (Fantuzzi and Dinarello, 1996). A robust COX-2 mRNA signal is also rapidly detected in the cerebral microvessels in response to IV IL-1ß and TNF-α injection, but not following high doses of IL-6 (Lacroix and Rivest, 1998). The inability of IL-6 to stimulate COX-2 raises serious concerns for the potential role of this cytokine in mediating fever during the acute-phase response. The large majority of COX-2-expressing cells are positive for a marker of the cerebral endothelium, i.e. Von Willebrand Factor (vWF). Fig. 5 (top and middle panels) shows representative examples of such dual labelling in a long blood vessel of the caudal medulla 45 min after IV injection of recombinant rat TNF-α. Numerous other reports have reported similar data that the endothelium of the cerebral blood vessels was the main source of PG production in response systemic LPS and IL-1

770

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

Fig. 4. Representative example of the distribution of cyclooxygenase-2 (COX-2) transcript in the rat brain after sterile local inflammation induced by intramuscular (IM) administration of turpentine (50 µl/100 g bw). Animals were sacrificed 6 h after IM injection of turpentine or the vehicle solution in the left hind limb. These rostro–caudal coronal sections (30 µm) exhibit positive signal on x-ray films (Biomax) for COX-2 mRNA throughout the brain microvasculature and other non-parenchymal structures, including the choroid plexus and the leptomeninges of turpentine-treated rat. 4V, fourth ventricle; AQ, aqueduct; BLA, basolateral nucleus of the amygdala; bv, blood vessels; Cer, cerebellum; CP, caudate putamen; DG, dentate gyrus; DVC, dorsovagal complex; Hip, hippocampus; IPN, interpeduncular nucleus; LGc, lateral geniculate complex; LV, lateral ventricle; MGv, medio–ventral geniculate nucleus; PB, parabrachial nucleus; PG, pontine gray; Pir, piriform cortex; VLM, ventrolateral medulla. Taken from Lacroix and Rivest (1998).

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

771

Fig. 5. Phenotype of COX-2- and I␬Bα-expressing cells in a rat cerebral blood vessel (bv) during systemic inflammatory insults. Endothelial cells were labelled by immunohistochemistry using antisera against the von Willebrand Factor (vWF-ir), whereas a COX-2 antibody (COX-2-ir) was used to perform the dual labelling with I␬Bα-expressing cells (bottom panels). COX-2 or I␬Bα mRNA was thereafter hybridized on the same sections by means of a radioactive in situ hybridization technique (silver grains). Arrowheads, dual-labelled cells (endothelial/COX-2 mRNA, top and middle panels; COX-2-ir/I␬Bα mRNA, bottom panels). Taken from Rivest (1999).

treatments (Cao et al. 1996, 1999; Quan et al., 1998; Matsumura et al., 1998a,b,c). However, other studies found COX-2-immunoreactive (ir) cells within perivascular microglia and meningeal macrophages (Elmquist et al., 1997; Schiltz and Sawchenko, 1998). The presence of the so-called perivascular microglia may actually be circulating monocytes that are attached to the cerebral endothelium in response to systemic immune challenges. There is a robust transcriptional activation of the gene encoding the MCP-1 in microvascular-associated elements of LPS-, IL-1ß- and TNFα-challenged animals (Thibeault et al., 2001). These regions would therefore have the ability to rapidly attract cells that bear the MCP-1 receptor CCR2, such as circulating monocytes. This chemoattraction is likely to explain the cells of monocytic lineage that are attached to the cerebral endothelium in animals that received systemic injec-

772

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

tion with recombinant IL-1ß or LPS. In contrast to the endothelium, these attached cells are not constitutive elements of the BBB. 7. NF-␬B and COX-2 in the cerebral endothelium A robust activation of NF-␬B is also detected in the endothelium of the brain capillaries in response to different systemic inflammatory stimuli (Laflamme et al., 1999; Laflamme and Rivest, 1999). Systemic LPS, IL-1ß and TNF-α injections provoke a rapid de novo expression of I␬Bα mRNA (index of NF-␬B activity) in the endothelium of the brain blood vessels and parenchymal microglia (Laflamme et al., 1999; Laflamme and Rivest, 1999). The effects of IV IL-1ß and TNF-α take place rapidly (within 30–60 min) and vanish 3 h after the injection (Laflamme and Rivest, 1999 and Fig. 6). On the other hand, a selective expression of I␬Bα was detected along the cerebral endothelium of IM turpentine-injected rats (Laflamme and Rivest, 1999). Fig. 7 (bottom right) shows that the mRNA encoding I␬Bα is essentially expressed within endothelial cells of the brain irrigating system. The rapid and transient induction of I␬Bα indicates a strong NF-␬B activity in cells of the BBB by circulating proinflammatory cytokines, which may lead to the transcription of target genes. One potential candidate is the gene encoding COX-2; the nuclear factor binding to the COX-2 promoter is able to influence the enzyme transcription in response to different immunogenic ligands, including IL-1ß and TNF (Crofford et al., 1997; Sorli et al., 1998). Two putative NF-␬B motifs from the COX-2 promoter were found to bind p50/p65 NF-␬B heterodimers in an IL-1ß-dependent manner and the two NF-␬B subunits synergistically activate a ⫺917/+49 COX-2 promoter construct (Crofford et al., 1997; Sorli et al., 1998). The presence of IL-1ß and TNF type I receptors in the endothelium of the brain capillaries provides the anatomical evidence that, once into the bloodstream, both cytokines have the ability to trigger these events within the cells that form the BBB. Activation of the cerebral endothelium also occurs in response to circulating LPS (Lacroix and Rivest, 1998; Laflamme et al., 1999; Laflamme and Rivest, 1999), but endothelial cells do not express membrane CD14. However, these cells respond to LPS/LBP in a soluble (s)CD14-dependent manner in stimulating the tyrosine phosphorylation of MAP kinases (Arditi et al., 1995), which engages the NF-␬B signalling and COX-2 transcription. These molecular events occurring in the cerebral microvasculature lead to the formation of specific PGs that have the ability to talk with neurons in cognate with their transmembrane receptors.

8. PGE2 is the key endogenous ligand within the CNS We propose here that the microvasculature is the source of PG formation into the brain during systemic inflammatory challenges (LPS, IL-1 and TNF, but not IL-6) that trigger the NF-␬B signalling pathways and COX-2 transcription within the cer-

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

773

Fig. 6. Representative examples of the influence of proinflammatory cytokines on the expression of mRNA encoding I␬Bα in the rat brain. Animals were killed 1 h after intravenous (IV) injection of TNFα (4.1 µg/kg of bw), IL-1β (1.8 µg/kg of bw) and IL-6 (12 µg/kg of bw). These darkfield photomicrographs show in situ hybridization signals for I␬Bα mRNA through the entire brain microvasculature and within small cells of the brain parenchyma in both TNF-α- and IL-1β-treated rats, but not following IL6 injection. b.v., blood vessels; chp, choroid plexus; OVLT, organum vasculosum of the lamina terminalis; SFO, subfornical organ. Magnifications: X25 and X50 (bottom panels). Taken from Laflamme and Rivest (1999).

774

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

Fig. 7. Activation of the nuclear factor kappa B (NF-␬B) and cyclooxygenase-2 (COX-2) genes in the cerebral blood vessels in response to a systemic inflammation. Top panels depict COX-2 gene expression in a large blood vessel (bv) of a rat that received the endotoxin lipopolysaccharide (LPS) into the peritoneal cavity. The bottom panels are examples of I␬Bα mRNA signal within the endothelium of brain microvasculature. Left panels are darkfield photomicrographs, whereas right panels are the same samples but taken under brightfield illumination. Endothelial cells were labelled by immunohistochemistry using antisera against the von Willebrand Factor (vWF-ir), whereas I␬Bα mRNA was thereafter hybridized on the same sections by means of a radioactive in situ hybridization technique (silver grains). COX-2 mRNA was visualized by single radioactive in situ hybridization histochemistry (top panels, silver grains). Taken from Rivest (1999).

ebral endothelium. It is interesting to note that inflammation-induced NF-␬B activity and COX-2 transcription is rather unspecific across the cerebral blood vessels and small capillaries, while the neuronal activity is limited to selective nuclei. PG synthesis via COX-2 activity along with site-specific expression of PG receptors within parenchymal cells adjacent to the site of production determine the action of these molecules in the brain. The exact PG subtype(s) and the site(s) of action within the brain involved in these effects have recently been unravelled and compelling evidence points in direction to the PG of E2 type. IL-1 increases the release of PGE2 from rat hypothalamic explants in vitro (Navarra et al., 1992), medial preoptic area (MPOA)/OVLT, PVN, dorsal hippocampus, lateral ventricle in vivo (Komaki et al., 1992), rat astrocyte cultures (Katsuura et al., 1989), isolated pancreatic islets (Hughes et al., 1989), and papillary collecting duct (Kohan, 1989). Central treatment with PGE2 is not only associated with an increase in the activity of the HPA axis, but is also known to produce many other physiological responses such as the alteration of the cardiovascular and

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

775

sympathetic nervous system functions (Feuerstein et al., 1982; Ando et al., 1995) and hyperthermia in rats (Oka and Hori, 1994). The most pyrogenic preoptic sites are clustered along the ventromedial aspect of the POA, surrounding and just anterior to the OVLT (Scammell et al., 1996). When injected into the brain, PGE2 causes a selective cellular activation as indicated by the rapid and transient expression of the immediate-early gene (IEG) c-fos mRNA and protein within multiple regions of the brain recognized to be activated during the acute-phase response of an immune challenge (Lacroix et al., 1996; Scammell et al., 1996). In a similar manner, the PG triggered transcription of CRF and its type 1 receptor essentially in the hypothalamic PVN (Lacroix et al., 1996). Local production of PGE2 is therefore likely to be a crucial step within the CNS to mediate the effects of circulating cytokines and bacterial cell wall components on the neuronal circuitry mediating the activation of fever, HPA axis and other autonomic functions. Once produced by the endothelium of the brain capillaries, PGE2 has the ability to diffuse through the brain parenchyma and activate different populations of cells in interacting with their transmembrane receptors.

9. PGE2 receptors Classic PG receptors comprise a family of eight genes encoding transmembrane G-protein-coupled receptors. These receptors are classified on the basis of selective affinities for naturally occurring prostanoids. There are distinct receptors for thromboxane A2, prostacyclins (PGI2), PGF2α, PGD2 and four different receptors for PGE2; EP1, EP2, EP3 and EP4 (Coleman et al., 1994; Narumiya et al., 1999). Multiple alternatively spliced isoforms exist for one of the PGE receptors (EP3) (Namba et al., 1993). Each receptor is associated with a unique G-protein and consequently a unique second messenger system, namely elevation of intracellular Ca2+ (EP1) and stimulation (EP2, EP4) or inhibition (EP3) of adenylate cyclase. We (Zhang and Rivest 1999, 2000) and other (Oka et al., 2000) have recently reported an interesting distribution pattern of EP2 and EP4 in the rat brain and proposed the EP4 subtype as the possible functional receptor for mediating the action of PGE2 on specific groups of neurons in response to circulating pro-inflammatory cytokines, such as IL1ß. Indeed, EP4-expressing neurons of the PVN, nucleus of the solitary tract (NTS) and the caudal ventrolateral medulla (cVLM) are activated by circulating IL-1ß and LPS (Zhang and Rivest, 1999) and it is possible that such activation depends on the local synthesis of PGE2 by cells of the microvasculature penetrating these regions. It is interesting to note that EP4 receptors were located in regions that are likely involved in the control of neuroendocrine and autonomic activities (Fig. 8). These include the PVN, SON, parabrachial nucleus (PB), NTS and cVLM and change in expression levels during systemic immunogenic challenges indicates that specific compartments of these nuclei may participate in the circuits triggered by the PG. The most dramatic change was the profound transcriptional activation of the gene encoding the EP4 receptor over the parvocellular CRF neurons of the hypothalamic PVN in response to different experimental models of systemic inflammation. The

776

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

Fig. 8. Distribution of the mRNA encoding PGE2 EP4 receptor in the rat brain under basal conditions (vehicle), 3 h after IV injection of LPS (10 µg/100 g bw), 6 h after IM injection of turpentine (50 µl/100 g bw) and 3 h after IV injection of recombinant rat IL-1β (187 ng/100 g bw). These rostrocaudal coronal sections (30 µm) display a positive signal on x-ray film (Biomax, Rochester, NY) for the EP4 transcript in various structures throughout the brain. BNST, bed nucleus of stria terminalis; Cer, cerebellar cortex; cVLM, caudal ventrolateral medulla; ep, ependymal lining cells of the lateral ventricle; LC, locus coeruleus; MeA, medial nucleus of the amygdala; NTS, nucleus of the solitary tract; PB, parabrachial nucleus; PMv, paramammillary nucleus, ventral part; PVN, paraventricular nucleus of the hypothalamus; SH, septohippocampal nucleus; SFO, subfornical organ; SON, supraoptic nucleus of the hypothalamus; VSA, ventroseptal area. Taken from Zhang and Rivest (1999).

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

777

presence of Fos-ir proteins in EP4-expressing neurons of the endocrine hypothalamus also confirmed that these cells were activated during the immunogenic insults (Fig. 9). Another interesting result is the expression of the EP4 receptor in the A2/C2 and A1 cell groups that were positive for Fos-ir labelling 3 h after IV IL-1ß injection (Zhang and Rivest 1999, 2000). The pro-inflammatory cytokine also stimulated the EP4 biosynthesis within the A1 noradrenergic cells, whereas the A2/C2 group remained without significant effects (Zhang and Rivest 1999, 2000). These studies provided the anatomical evidence that both EP2 and EP4 subtypes are widely distributed across different neuronal populations of the brain and that systemic immunogenic stimuli modulate expression of these transcripts in a site- and cell-dependent manner (Zhang and Rivest 1999, 2000). The activation of CRF and A1 groups of neurons along with the robust hybridization signal for the EP4 receptor clearly indicate that PGE2 may have a direct action onto these cells, which may be leading to the appropriate physiological responses necessary to preserve the integrity of the organism when challenged by foreign materials (Fig. 10). The positive EP4 signal over the neuroendocrine population of CRF neurons is indeed quite fascinating, as it provides direct anatomical evidence that the release of PGE2 within the PVN is likely to have direct access to the neurons controlling the corticotroph axis (Zhang and Rivest, 1999). A strong hybridization signal for COX-2 mRNA along the microvessels of the parvocellular PVN occurs in response to systemic inflammatory insults (Lacroix and Rivest, 1998) and systemic injection of IL-1 provokes sharply increased levels of PGE2 in the PVN (Watanobe and Takebe, 1994). Central PGE2 injection also induces c-fos mRNA in CRF neurons, triggers the transcription of the neuropeptide specifically within the parvocellular PVN (Lacroix et al., 1996) and increases plasma corticosterone levels (Rassnick et al., 1995). Like the EP2 subtype, PGE2 EP4 receptors activate Gs proteins and adenylate cyclase (Fedyk et al., 1996), which is the most effective second messenger systemtaking place in activated CRF neurons and activates gene transcription through cAMP-responsive element located on the CRF promoter (Guardiola-Diaz et al., 1994). Together these data suggest that PGE2 produced by the PVN blood vessels directly targets the EP4 receptor onto neuroendocrine CRF cells, which leads to cAMP signalling and infundibular release as well as transcription of the stress-related neuropeptide during systemic inflammatory processes (see Fig. 10).

10. The IL-6 story Although IL-6 can stimulate the HPA axis directly at the level of both pituitary and adrenal glands, the cytokine is believed to trigger infundibular CRF secretion originating from parvocellular PVN neurons. It is therefore surprising that a single injection of IL-6 is insufficient to induce hypothalamic transcription of c-fos and CRF (Vallie`res et al., 1997), a phenomenon contrasting with the profound stimulation of these genes in the PVN of IL-1ß and LPS-injected rats (Rivest and Rivier 1994, 1995; Rivest, 1995; Rivest and Laflamme, 1995; Rivest et al. 1995, 2000). This lack of effect of IL-6 could be explained by the absence of IL-6 receptor (IL-6R) in the

Fig. 9. Demonstration of EP4 mRNA activated neurons (using c-fos-immunoreactive protein as an index of cellular activation) in response to intravenous IL-1β administration in regions of the PVN, NTS and cVLM. Animals were sacrificed 3 h after the IV injection of the pro-inflammatory cytokine. Immunocytochemistry (Fos protein, stained nucleus) was performed on the same brain sections (30 µm) prior to in situ hybridization histochemistry (EP4 mRNA, silver grains). High-power brightfield photomicrographs of right column were taken respectively from the sections depicted by the middle column. Solid arrows show EP4 cells expressing the IEG Fos-ir in their nuclei. Magnification of left and middle column, ×25; right column, ×250. Taken from Zhang and Rivest (1999).

778 S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

779

Fig. 10. Intracellular mechanisms mediating the influence of circulating interleukin-1ß (IL-1ß) on the transcription of the cyclooxygenase-2 (COX-2) within an endothelial cell of the blood brain barrier. Although simplistic, both the MAP kinases and NF-␬B pathways may be transduction/transcription signals in these processes. The production of the prostaglandin of E2 type is believed to be a key mediator to diffuse through the parenchymal brain and the neurons that control fever and the hypothalamic–pituitary– adrenal axis. The subsequent release of glucocorticoids is determinant for the immunosuppression of the systemic inflammation and the down regulation COX-2 transcription. Glucocorticoids may increase I␬Bα transcription and/or interfere with NF-␬B binding ability on COX-2 promoter in cerebral vascular cells.

PVN neurons under basal conditions. Induction of the acute-phase response stimulates the expression of IL-6R mRNA in this hypothalamic nucleus and this event either permits or potentiates the action of IL-6 on neuroendocrine CRF cells and the HPA axis. However, the fact that blood vessels express the receptor for IL-6 in response to LPS and IL-1ß suggests that such phenomenon may occur in a nonselective manner through brain microvasculature. The physiological relevance of the IL-6R expression along blood vessels of LPS- and IL-1ß-injected rats remains to be clearly established, although it is highly possible that endothelial are target cells of circulating IL-6, which in turn may transduce specific signals to parenchymal cells (including neurons) during a complex systemic immune challenge. As for IL-1ß and TNF, it is conceivable to believe that circulating IL-6 stimulates the cerebral endothelium to release paracrine factors, such as PGs. Unfortunately, depending on the experimental procedure and the tissue studied, it can be found in the literature as many papers that prove their involvement in mediating the central

780

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

effects of IL-6 as there are papers that are in disagreement with it. For examples, it has been demonstrated that PGs mediate IL-6-induced fever (Dinarello et al., 1991; Xin and Blatteis, 1992; Fernandez-Alonso et al., 1996) and HPA axis stimulation (Lyson and McCann, 1992), but IL-6 is unable to activate PG formation in cerebral microvessels (Bishai and Coceani, 1996) and to induce COX-2 mRNA synthesis in rat brains (Lacroix and Rivest, 1998) or in cultured microglial cells (Bauer et al., 1997). It is clear, however, that IL-6 does not stimulate the production of PGs in peripheral organs and that, conversely, its own synthesis is induced by them (Hinson et al., 1996; Williams and Shacter, 1997). These observations suggest that IL-6 does not stimulate COX-2 gene transcription, but the possibility that IL-6 may influence PG synthesis at post-transcriptional levels or cooperate with them to activate CRF neurons cannot be ruled out. Recent data suggest that IL-6 signalling is enhanced during endotoxemia and IL6 modulates PVN functions only after pre-induction of its receptor in immune-challenged animals (Vallie`res and Rivest, 1999). Using a dual labelling technique, we also found that some CRF neurons express IL-6R suggesting that IL-6 may directly target these cells to trigger neuronal activation and CRF secretion (Vallie`res and Rivest, 1999). In this regard, systemic LPS activates the transcription of the gene encoding IL-6 selectively in the CVOs and the choroid plexus (Vallie`res and Rivest, 1997), which provides evidence that the cytokine is produced in the brain and may act on neurons or other parenchymal elements of the CNS. In addition, our latest data obtained from the experiment performed in wild-type and IL-6-deficient mice support the concept that IL-6, although not involved during the initial phases of endotoxemia, is necessary during the later phases for maintaining the stimulation of CRF neurons controlling the HPA axis and prolonging the activation of neural cells throughout the brain (Vallie`res and Rivest, 1999). The increase in circulating corticosterone levels is also lower in IL-6⫺/⫺ than IL-6+/+ mice in response to IP LPS injection, but not during restrain stress (Ruzek et al., 1997). This suggests that the involvement of IL-6 in the control of the HPA axis is quite specific to the immune stimuli and not neurogenic stresses.

11. IL-6 signalling cascade in the CNS The bacterial endotoxin triggers transcription of the gene encoding the suppressor of cytokine signalling 3 (SOCS-3) in all the CVOs and their adjacent structures, chp, the ependymal lining cells of the cerebroventricular system and along the endothelium of the cerebral capillaries (Lebel et al., 2000). This indicates that IL-6 can trigger the JAK/STAT signalling in these groups of cells that may allow the transcription of the target genes involved in the neural-immune interface. A single systemic injection of LPS caused a rapid increase of TNF-α and IL-1ß in the bloodstream, which is followed by a gradual elevation of plasma IL-6 (Givalois et al., 1994). Despite the possible overlap with the JAK/STAT/SOCS pathways, TNF and IL1 are well-recognized NF-␬B-inducible cytokines (Baeuerle and Baltimore, 1996; Baeuerle, 1998) and their contribution may be minimal in regulating SOCS-3. Sup-

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

781

porting this idea is the fact that the inhibitory factor (SOCS-3) was no longer expressed in the cerebral microvasculature and the ventricle lining walls of IL-6deficient mice during endotoxemia (Lebel et al., 2000). Such phenomenon is however not generalized to all responsive structures, as the CVOs and choroid plexus of these animals still exhibit a positive signal for SOCS-3 mRNA following the LPS challenge. This result suggests that IL-6 is a key ligand for activating SOCS-3 transcription only in barrier-associated groups of cells, but not in the CVOs. Despite the fact that SOCS-3 transcription is dependent on the endogenous release of IL-6 in cells lining the BBB and the ventricles, a single administration of the cytokine failed to activate the signalling events leading to SOCS-3 production in vivo (Lebel et al., 2000). This result may be explained by the lack of constitutive expression of both IL-6 receptors, at least in the endothelium of the brain microvessels (Vallie`res and Rivest 1997, 1999). As mentioned, systemic LPS challenge induces both IL-6R and gp130 transcripts in endothelial cells and this phenomenon is a prerequisite for the IL-6 signalling during the acute-phase response (Vallie`res and Rivest 1997, 1999). Whether IL-6R and gp130 expression in the brain microvasculature depends on the release of IL-1ß and/or TNF-α or is a direct action of the endotoxin onto the cerebral endothelium is still an open question. Besides, IL-1ß is capable of activating IL-6 receptors in cells that can be reached from the systemic circulation (Vallie`res and Rivest, 1997) and IL-1ß precedes the release of IL-6 during systemic inflammation. This potential mechanism is quite interesting as it follows the time-related events that occur during the acute-phase reaction of an immune challenge. Further supporting this concept is the late expression of SOCS-3 in the brain microvasculature of animals injected with turpentine into the left hind paw (Lebel et al., 2000). It is therefore tempting to propose the following sequence of events: 1) IL-1ß and TNF-α are released early in the bloodstream during the acute-phase reaction; 2) these circulating cytokines have the ability to reach the large arterioles, small capillaries and venules, which trigger NF-␬B nuclear translocation and stimulate IL-6R and gp130 transcription in endothelial cells; 3) these receptors may then respond when the ligand becomes available in the circulation that activates selective JAK/STAT molecules; 4) phosphorylated homodimers of STATs (most likely STAT1 and/or STAT3) target SOCS-3 promoter and stimulate its transcription; and 5) increase in SOCS-3 protein inhibits this cytokine signalling and the proinflammatory signal transduction pathways. Similar events may also take place directly in the brain with the choroid plexus resident macrophages as being the cellular source of IL-1ß, TNFα and IL-6 and the ependymal as the cells responding to these ligands of central origin. Altogether these data suggest that the pre-induction of the IL-6 receptors is a prerequisite allowing IL-6 to trigger the transducing events and then SOCS-3 production in the cerebral endothelium and a specific group of supportive cells. The cytokine, when present in the circulation, may then act as the subsequent step to maintain the neuronal activity involved in the adequate control of the homeostatic balance during systemic inflammation. Moreover, IL-6-induced SOCS-3 is likely to be part of the anti-inflammatory mechanisms that take place in a very organized manner within the CNS. The molecules mediating the action of circulating IL-6 that

782

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

are targeted by the JAK/STAT signalling and diffuse across the brain parenchyma to stimulate neurons involved in the autonomic functions have yet to be determined. As mentioned, the JAK/STAT transduction pathways do not activate COX-2 transcription and is unlikely to increase PG synthesis, which are quite selectively induced by cytokines and agents that activate the NF-␬B signalling cascade. Therefore, although the molecular events mediating the action of LPS, IL-1 and TNF are well characterized (see above), those completing the IL-6 transduction pathways in the cerebral tissue still await to be fully described.

12. Proposed cascade of events and concluding remarks

1. Circulating molecules produced by systemic inflamed sites target cells of the BBB to release intermediates in the brain parenchyma. 2. IL-1ß is the most likely candidate to mediate the early effects during systemic and localized insults, but not during endotoxemia (Laflamme et al., 1999). 3. The binding of IL-1ß to its type I receptor and IL-1R-AcP engages the IRAK/TRAF6/MyD88/NIK/IKK pathway responsible to trigger COX-2 gene transcription in the cerebral endothelium. 4. Activation of the COX-2 enzyme leads to PGE2 synthesis. 5. PGE2 may therefore diffuse through the parenchymal elements of the brain and bind to its EP4 receptor expressed at the surface of CRF dendrites, which induces the cAMP/CREB transduction pathway. 6. This PG may also cognate to its receptor expressed in key innervating neurons of the VmPO and medulla to induce/maintain fever and the corticotroph axis, respectively. 7. All together these circuits are responsible for engaging the autonomic/neuroendocrine functions during the acute-phase response of an immune-challenge. 8. IL-6, in activating the JAK/STAT/SOCS signalling, takes the relay to prolong the neuronal activity and the autonomic functions. 9. Glucocorticoids are the essential feedback to inhibit the pro-inflammatory signal transduction pathways and transcriptional activation of target genes. We know a little more on the site(s) of action and the neuronal pathways participating in the influence of circulating cytokines on the brain functions fundamental to restoring homeostasis during immunogenic stimuli. Obviously, investigating the transduction signals involved in these events is somewhat limited in vivo, though the time-related induction of specific signalling molecules within the same population(s) of cells has generated crucial information for future directions. The participation of the endothelium, microglia or other cell types lining the BBB may depend on numerous factors, including the severity and duration of the inflammatory insults as well as the model of the systemic immunogenic challenges. Although the

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

783

intracellular events may largely differ among stimuli and endogenous ligands (LPS, TNF, IL-1 Vs IL-6), the resulting production of PGE2 is likely to be the critical link between the circulating immune molecules and parenchymal elements of the brain to activate the neuronal circuits controlling different neurophysiological and neuroendocrine functions. The HPA axis is one of these functions having a profound impact in regulating the inflammation (see Fig. 10). Inappropriate plasma levels of glucocorticoids may play a crucial role as contributing to deviant regulation of the immune response, indicating the importance to identify and characterize the mechanisms through which inflammatory molecules interact with, and depend on, the neuroendocrine system. Disorders of this fine interplay may indeed contribute to the onset and progression of various pathological states characterized by an exaggerated inflammatory response in the periphery as well as in the CNS. The recent evidence that corticoids may inhibit pro-inflammatory molecules via a direct genomic effect in stimulating the transcription of the inhibitory factor I␬Bα (Auphan et al., 1995; Scheinman et al., 1995) or in interfering with the transactivation potential of the NF-␬B p65 subunit (De Bosscher et al., 1997; Wissink et al., 1998) are exciting new developments and we believe that such events are essential inputs to the brain. Indeed, the negative feedback of the glucocorticoids on the NF-␬B signalling, COX-2 transcription and then PGE2 production is likely to be the mechanism by which corticoids suppress the previously engaged responses, namely fever, HPA axis as well as other autonomic functions.

Acknowledgements This work is supported by the Canadian Institutes of Health Research (CIHR; the former Medical Research Council of Canada (MRCC)). The author is a MRCC Scientist and holds a Canadian Research Chair in Neuroimmunology.

References Alexander, W.S., Starr, R., Metcalf, D., Nicholson, S.E., Farley, A., Elefanty, A.G., Brysha, M., Kile, B.T., Richardson, R., Baca, M., Zhang, J.G., Willson, T.A., Viney, E.M., Sprigg, N.S., Rakar, S., Corbin, J., Mifsud, S., DiRago, L., Cary, D., Nicola, N.A., Hilton, D.J., 1999. Suppressors of cytokine signaling (SOCS): negative regulators of signal transduction. J. Leukoc. Biol. 66, 588–592. Aman, M.J., Leonard, W.J., 1997. Cytokine signaling: cytokine-inducible signaling inhibitors. Curr. Biol. 7, R784–R788. Ando, T., Ichijo, T., Katafuchi, T., Hori, T., 1995. Intracerebroventricular injection of prostaglandin E2 increases splenic sympathetic nerve activity in rats. Am. J. Physiol. 269, R662–R668. Arditi, M., Zhou, J., Torres, M., Durden, D.L., Stins, M., Kim, K.S., 1995. Lipopolysaccharide stimulates the tyrosine phosphorylation of mitogen-activated protein kinases p44, p42, and p41 in vascular endothelial cells in a soluble CD14-dependent manner. Role of protein tyrosine phosphorylation in lipopolysaccharide-induced stimulation of endothelial cells. J. Immunol. 155, 3994–4003. Auphan, N., DiDonato, J.A., Rosette, C., Helmberg, A., Karin, M., 1995. Immunosuppression by glucocorticoids: inhibition of NF-kappa B activity through induction of I kappa B synthesis. Science 270, 286–290.

784

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

Baeuerle, P.A., 1998. Pro-inflammatory signaling: last pieces in the NF-kappa B puzzle. Curr. Biol. 8, R19–R22. Baeuerle, P.A., Baltimore, D., 1996. NF kappa B: ten years after. Cell 87, 13–20. Bauer, M.K., Lieb, K., Schulze-Osthoff, K., Berger, M., Gebicke-Haerter, P.J., Bauer, J., Fiebich, B.L., 1997. Expression and regulation of cyclooxygenase-2 in rat microglia. Eur. J. Biochem. 243, 726–731. Bishai, I., Coceani, F., 1996. Differential effects of endotoxin and cytokines on prostaglandin E2 formation in cerebral microvessels and brain parenchyma: implications for the pathogenesis of fever. Cytokine 8, 371–376. Brennan, F.M., Maini, R.N., Feldmann, M., 1992. TNF alpha: a pivotal role in rheumatoid arthritis? Br. J. Rheumatol. 31, 293–298. Cao, C., Matsumura, K., Yamagata, K., Watanabe, Y., 1995. Induction by lipopolysaccharide of cyclooxygenase-2 mRNA in the rat brain; its possible role in the febrile response. Brain Res. 697, 187–196. Cao, C.Y., Matsumura, K., Yamagata, K., Watanabe, Y., 1996. Endothelial-cells of the rat-brain vasculature express cyclooxygenase-2 messenger-RNA in response to systemic interleukin-1: a possible site of prostaglandin synthesis responsible for fever. Brain Res. 733, 263–272. Cao, C., Matsumura, K., Ozaki, M., Watanabe, Y., 1999. Lipopolysaccharide injected into the cerebral ventricle evokes fever through induction of cyclooxygenase-2 in brain endothelial cells. J. Neurosci. 19, 716–725. Carswell, E.A., Old, L.J., Kassel, R.L., Green, S., Fiore, N., Williamson, B., 1975. An endotoxin-induced serum factor that causes necrosis of tumors. Proc. Natl. Acad. Sci. USA 72, 3666–3670. Chu, W.M., Ostertag, D., Li, Z.W., Chang, L., Chen, Y., Hu, Y., Williams, B., Perrault, J., Karin, M., 1999. JNK2 and IKKbeta are required for activating the innate response to viral infection. Immunity 11, 721–731. Chung, I.Y., Benveniste, E.N., 1990. Tumor necrosis factor-alpha production by astrocytes. Induction by lipopolysaccharide, IFN-gamma, and IL-1beta. J. Immunol. 144, 2999–3007. Coleman, R.A., Smith, W.L., Narumiya, S., 1994. International Union of Pharmacology classification of prostanoid receptors: properties, distribution, and structure of the receptors and their subtypes. Pharmacol. Rev. 46, 205–229. Crofford, L.J., Tan, B., McCarthy, C.J., Hla, T., 1997. Involvement of nuclear factor kappa B in the regulation of cyclooxygenase-2 expression by interleukin-1 in rheumatoid synoviocytes. Arthritis Rheum. 40, 226–236. De Bosscher, K., Schmitz, M.L., Vanden Berghe, W., Plaisance, S., Fiers, W., Haegeman, G., 1997. Glucocorticoid-mediated repression of nuclear factor-kappaB-dependent transcription involves direct interference with transactivation. Proc. Natl. Acad. Sci. USA 94, 13504–13509. Delhase, M., Hayakawa, M., Chen, Y., Karin, M., 1999. Positive and negative regulation of IkappaB kinase activity through IKKbeta subunit phosphorylation. Science 284, 309–313. DeWitt, D.L., Smith, W.L., 1988. Primary structure of prostaglandin G/H synthase from sheep vesicular gland determined from the complementary DNA sequence. Proc. Natl. Acad. Sci. USA 85, 1412–1416. DeWitt, D.L., El-Harith, E.A., Kraemer, S.A., Andrews, M.J., Yao, E.F., Armstrong, R.L., Smith, W.L., 1990. The aspirin and heme-binding sites of ovine and murine prostaglandin endoperoxide synthases. J. Biol. Chem. 265, 5192–5198. Dinarello, C.A., Cannon, J.G., Mancilla, J., Bishai, I., Lees, J., Coceani, F., 1991. Interleukin-6 as an endogenous pyrogen: induction of prostaglandin E2 in brain but not in peripheral blood mononuclear cells. Brain Res. 562, 199–206. Elmquist, J.K., Breder, C.D., Sherin, J.E., Scammell, T.E., Hickey, W.F., Dewitt, D., Saper, C.B., 1997. Intravenous lipopolysaccharide induces cyclooxygenase 2-like immunoreactivity in rat brain perivascular microglia and meningeal macrophages. J. Comp. Neurol. 381, 119–129. Ericsson, A., Liu, C., Hart, R.P., Sawchenko, P.E., 1995. Type-1 interleukin-1 receptor in the rat brain: distribution, regulation, and relationship to sites of IL-1-induced cellular activation. J. Comp. Neurol. 361, 681–698. Fantuzzi, G., Dinarello, C.A., 1996. The inflammatory response in interleukin-1ß-deficient mice: comparison with other cytokine-related knock-out mice. J. Leukoc. Biol. 59, 489–493. Fedyk, E.R., Ripper, J.M., Brown, D.M., Phipps, R.P., 1996. A molecular analysis of PGE receptor (EP)

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

785

expression on normal and transformed B lymphocytes: coexpression of EP1, EP2, EP3beta and EP4. Mol. Immunol. 33, 33–45. Fernandez-Alonso, A., Benamar, K., Sancibrian, M., Lopez-Valpuesta, F.J., Minano, F.J., 1996. Role of interleukin-1 beta, interleukin-6 and macrophage inflammatory protein-1 beta in prostaglandin-E2induced hyperthermia in rats. Life Sci. 59, L185–L190. Feuerstein, G., Adelberg, S.A., Kopin, I.J., Jacubowitz, D.M., 1982. Hypothalamic sites for cardiovascular and sympathetic modulation by prostaglandin E2. Brain Res. 231, 335–342. Givalois, L., Dornand, J., Mekaouche, M., Solier, M.D., Bristow, A.F., Ixart, G., Siaud, P., Assenmacher, I., Barbanel, G., 1994. The temporal cascade of plasma level surges in ACTH, corticosterone and cytokines in endotoxin-challenged rats. Am. J. Physiol. 266, R164–R170. Goldfeld, A.E., Strominger, J.L., Doyle, C., 1991. Human tumor necrosis factor alpha gene regulation in phorbol ester stimulated T and B cell lines. J. Exp. Med. 174, 73–81. Guardiola-Diaz, H.M., Boswell, C., Seasholtz, A.F., 1994. The cAMP-responsive element in the corticotropin-releasing hormone gene mediates transcriptional regulation by depolarization. J. Biol. Chem. 269, 14784–14791. Hansen, J.A., Lindberg, K., Hilton, D.J., Nielsen, J.H., Billestrup, N., 1999. Mechanism of inhibition of growth hormone receptor signaling by suppressor of cytokine signaling proteins. Mol. Endocrinol. 13, 1832–1843. Hempel, S.L., Monick, M.M., Hunninghake, G.W., 1994. Lipopolysaccharide induces prostaglandin H synthase-2 protein and mRNA in human alveolar macrophages and blood monocytes. J. Clin. Invest. 93, 391–396. Herkenham, M., Lee, H.Y., Baker, R., 1998. Temporal and spatial patterns of c-fos mRNA induced by intravenous interleukin-1: a cascade of non-neuronal cellular activation at the blood–brain barrier. J. Comp. Neurol. 400, 175–196. Hilton, D.J., Richardson, R.T., Alexander, W.S., Viney, E.M., Willson, T.A., Sprigg, N.S., Starr, R., Nicholson, S.E., Metcalf, D., Nicola, N.A., 1998. Twenty proteins containing a C-terminal SOCS box form five structural classes. Proc. Natl. Acad. Sci. USA 95, 114–119. Hinson, R.M., Williams, J.A., Shacter, E., 1996. Elevated interleukin 6 is induced by prostaglandin E2 in a murine model of inflammation: possible role of cyclooxygenase-2. Proc. Natl. Acad. Sci. USA 93, 4885–4890. Hirano, T., 1998. Interleukin 6 and its receptor: ten years later. Int. Rev. Immunol. 16, 249–284. Hirsch, E., Irikura, V.M., Paul, S.M., Hirsh, D., 1996. Functions of interleukin 1 receptor antagonist in gene knockout and overproducing mice. Proc. Natl. Acad. Sci. USA 93, 11008–11013. Hu, Y., Baud, V., Delhase, M., Zhang, P., Deerinck, T., Ellisman, M., Johnson, R., Karin, M., 1999. Abnormal morphogenesis but intact IKK activation in mice lacking the IKKalpha subunit of IkappaB kinase. Science 284, 316–320. Hughes, J.H., Easom, R.A., Wolf, B.A., Turk, J., McDaniel, M.L., 1989. Interleukin-1-induced prostaglandin E2 accumulation by isolated pancreatic islets. Diabetes 38, 1251–1257. Jones, D.A., Carlton, D.P., McIntyre, T.M., Zimmerman, G.A., Prescott, S.M., 1993. Molecular cloning of human prostaglandin endoperoxide synthase type II and demonstration of expression in response to cytokines. J. Biol. Chem. 268, 9049–9054. Katsuura, G., Gottschall, P.E., Dahl, R.R., Arimura, A., 1989. Interleukin-1ß increases prostaglandin E2 in rat astrocyte cultures: modulating effect of neuropeptides. Endocrinology 124, 3125–3127. Kohan, D.E., 1989. Interleukin-1 regulation of prostaglandin E2 synthesis by the papillary collecting duct. J. Lab. Clin. Med. 114, 717–723. Komaki, G., Arimura, A., Koves, K., 1992. Effect of intravenous injection of IL-1ß on PGE2 levels in several brain areas as determined by microdialysis. Am. J. Physiol. 262 (Endocrinol. Metab. 25), E246–E251. Kopf, M., Baumann, H., Freer, G., Freudenberg, M., Lamers, M., Kishimoto, T., Zinkernagel, R., Bluethmann, H., Ko¨ hler, G., 1994. Impaired immune and acute-phase responses in interleukin-6deficient mice. Nature 368, 339–342. Lacroix, S., Rivest, S., 1998. Effect of acute systemic inflammatory response and cytokines on the transcription of the genes encoding cyclooxygenase enzymes (COX-1 and COX-2) in the rat brain. J. Neurochem. 70, 452–466.

786

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

Lacroix, S., Vallie`res, L., Rivest, S., 1996. C-fos mRNA pattern and CRF neuronal activity throughout the brain of rats injected centrally with a prostaglandin of E2 type. J. Neuroimmunol. 70, 163–179. Laflamme, N., Rivest, S., 1999. Effects of systemic immunogenic insults and circulating proinflammatory cytokines on the transcription of the inhibitory factor kappa B alpha within specific cellular populations of the rat brain. J. Neurochem. 73, 309–321. Laflamme, N., Lacroix, S., Rivest, S., 1999. An essential role of interleukin-1ß in mediating NF-␬B activity and COX-2 transcription in cells of the blood–brain barrier in response to systemic and localized inflammation, but not during endotoxemia. J. Neurosci. 19, 10923–10930. Leaman, D.W., Leung, S., Li, X., Stark, G.R., 1996. Regulation of STAT-dependent pathways by growth factors and cytokines. FASEB J. 10, 1578–1588. Lebel, E., Vallieres, L., Rivest, S., 2000. Selective involvement of interleukin-6 in the transcriptional activation of the suppressor of cytokine signaling-3 in the brain during systemic immune challenges. Endocrinology 141, 3749–3763. Lee, S.H., Soyoola, E., Chanmugam, P., Hart, S., Sun, W., Zhong, H., Liou, S., Simmons, D., Hwang, D., 1992. Selective expression of mitogen-inducible cyclooxygenase in macrophages stimulated with lipopolysaccharide. J. Biol. Chem. 267, 25934–25938. Liu, L., Kita, T., Tanaka, N., Kinoshita, Y., 1996. The expression of tumour necrosis factor in the hypothalamus after treatment with lipopolysaccharide. Int. J. Exp. Pathol. 77, 37–44. Lu¨ tticken, C., Wegenka, U.M., Yuan, J., Buschmann, J., Schindler, C., Ziemiecki, A., Harpur, A.G., Wilks, A.F., Yasukawa, K., Taga, T., et al. 1994. Association of transcription factor APRF and protein kinase Jak1 with the interleukin-6 signal transducer gp130. Science 263, 89–92. Lyson, K., McCann, S.M., 1992. Involvement of arachidonic acid cascade pathways in interleukin-6stimulated corticotropin-releasing factor release in vitro. Neuroendocrinology 55, 708–713. Matsumura, K., Cao, C., Ozaki, M., Morii, H., Nakadate, K., Watanabe, Y., 1998a. Brain endothelial cells express cyclooxygenase-2 during lipopolysaccharide-induced fever: light and electron microscopic immunocytochemical studies. J. Neurosci. 18, 6279–6289. Matsumura, K., Cao, C., Ozaki, M., Morii, H., Nakadate, K., Watanabe, Y., 1998b. Electron microscopic evidence for induction of cyclooxygenase-2 in brain endothelial cells. Ann. NY Acad. Sci. 856, 278–280. Matsumura, K., Cao, C., Watanabe, Y., 1998c. Prostaglandin system in the brain: sites of biosynthesis and sites of action under normal and hyperthermic states. Prog. Brain. Res. 115, 275–295. Miller, V.E., Rogers, K., Muirden, K.D., 1993. Detection of tumour necrosis factor alpha and interleukin1 beta in the rheumatoid osteoarthritic cartilage–pannus junction by immunohistochemical methods. Rheumatol. Int. 13, 77–82. Nadeau, S., Rivest, S., 1999a. Regulation of the gene encoding tumor necrosis factor alpha in the rat brain and pituitary in response to different models of systemic immune challenge. J. Neuropathol. Exp. Neurol. 58, 61–77. Nadeau, S., Rivest, S., 1999b. Effects of circulating tumor necrosis factor (TNF) on the neuronal activity and expression of the genes encoding the TNF receptors (p55 and p75) in the rat brain: a view from the blood–brain barrier. Neuroscience 93, 1449–1464. Nadeau, S., Rivest, S., 2000. Role of microglial-derived tumor necrosis factor in mediating CD14 transcription and NF-␬B activity in the brain during endotoxemia. J. Neurosci. 20, 3456–3468. Namba, T., Sugimoto, Y., Negishi, M., Irie, A., Ushikubi, F., Kakizuka, A., Ito, S., Ichikawa, A., Narumiya, S., 1993. Alternate splicing of C-terminal tail of prostaglandin E receptor subtype EP3 determine G-protein specificity. Nature 365, 166–170. Narazaki, M., Witthuhn, B.A., Yoshida, K., Silvennoinen, O., Yasukawa, K., Ihle, J.N., Kishimoto, T., Taga, T., 1994. Activation of JAK2 kinase mediated by the interleukin 6 signal transducer gp130. Proc. Natl. Acad. Sci. USA 91, 2285–2289. Narumiya, S., Sugimoto, Y., Ushikubi, F., 1999. Prostanoid receptors: structures, properties, and functions. Physiol. Rev. 79, 1193–1226. Navarra, P., Pozzoli, G., Brunetti, L., Ragazzoni, E., Besser, M., Grossman, A., 1992. Interleukin-1ß and interleukin-6 specifically increase the release of prostaglandin E2 from rat hypothalamic explants in vitro. Neuroendocrinology 56, 61–68.

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

787

Nesbitt, J.E., Fuller, G.M., 1992. Dynamics of interleukin-6 internalization and degradation in rat hepatocytes. J. Biol. Chem. 267, 5739–5742. Nicholson, S.E., Hilton, D.J., 1998. The SOCS proteins: a new family of negative regulators of signal transduction. J. Leukoc. Biol. 63, 665–668. Nicholson, S.E., Willson, T.A., Farley, A., Starr, R., Zhang, J.G., Baca, M., Alexander, W.S., Metcalf, D., Hilton, D.J., Nicola, N.A., 1999. Mutational analyses of the SOCS proteins suggest a dual domain requirement but distinct mechanisms for inhibition of LIF and IL-6 signal transduction. Embo J. 18, 375–385. Oka, T., Hori, T., 1994. EP1-receptor mediation of prostaglandin E2-induced hyperthermia in rats. Am. J. Physiol. 267, R289–R294. Oka, T., Oka, K., Scammell, T.E., Lee, C., Kelly, J.F., Nantel, F., Elmquist, J.K., Saper, C.B., 2000. Relationship of EP(1–4) prostaglandin receptors with rat hypothalamic cell groups involved in lipopolysaccharide fever responses. J. Comp. Neurol. 428, 20–32. Quan, N., Whiteside, M., Herkenham, M., 1998. Cyclooxygenase 2 mRNA expression in rat brain after peripheral injection of lipopolysaccharide. Brain Res. 802, 189–197. Rassnick, S., Zhou, D.H., Rabin, B.S., 1995. Central administration of prostaglandin E (2) suppresses in vitro cellular immune responses. Am. J. Physiol. 269, R92–R97. Rivest, S., 1995. Molecular mechanisms and neural pathways mediating the influence of interleukin-1 on the activity of neuroendocrine CRF motoneurons in the rat. Int. J. Dev. Neurosci. 13, 135–146. Rivest, S., 1999. What is the cellular source of prostaglandins in the brain in response to systemic inflammation? facts and controversies. Mol. Psychiatry. 4, 501–507. Rivest, S., Laflamme, N., 1995. Neuronal activity and neuropeptide gene transcription in the brain of immune-challenged rats. J. Neuroendocrinol. 7, 501–525. Rivest, S., Rivier, C., 1994. Stress and interleukin-1ß-induced activation of c-fos, NGFI-B and CRF gene expression in the hypothalamic PVN: Comparison between Sprague–Dawley, Fisher-344 and Lewis rats. J. Neuroendocrinol. 6, 101–117. Rivest, S., Rivier, C., 1995. The role of CRF and interleukin-1 in the regulation of neurons controlling reproductive functions. Endocr. Rev. 16, 177–199. Rivest, S., Laflamme, N., Nappi, R.E., 1995. Immune challenge and immobilization stress induce transcription of the gene encoding the CRF receptor in selective nuclei of the rat hypothalamus. J. Neurosci. 15, 2680–2695. Rivest, S., Lacroix, S., Vallie`res, L., Nadeau, S., Zhang, J., Laflamme, N., 2000. How the blood talks to the brain parenchyma and the paraventricular nucleus of the hypothalamus during systemic inflammatory and infectious stimuli. Proc. Soc. Exp. Biol. Med. 223, 22–38. Rose-John, S., Hipp, E., Lenz, D., Legres, L.G., Korr, H., Hirano, T., Kishimoto, T., Heinrich, P.C., 1991. Structural and functional studies on the human interleukin-6 receptor. Binding, cross-linking, internalization, and degradation of interleukin-6 by fibroblasts transfected with human interleukin-6receptor cDNA. J. Biol. Chem. 266, 3841–3846. Roux-Lombard, P., 1998. The interleukin-1 family. Eur. Cytokine Netw. 9, 565–576. Ruzek, M.C., Miller, A.H., Opal, S.M., Pearce, B.D., Biron, C.A., 1997. Characterization of early cytokine responses and an interleukin (IL)-6- dependent pathway of endogenous glucocorticoid induction during murine cytomegalovirus infection. J. Exp. Med. 185, 1185–1192. Scammell, T.E., Elmquist, J.K., Griffin, J.D., Saper, C.B., 1996. Ventromedial preoptic prostaglandin E2 activates fever-producing autonomic pathways. J. Neurosci. 16, 6246–6254. Scheinman, R.I., Cogswell, P.C., Lofquist, A.K., Baldwin, A.S. Jr., 1995. Role of transcriptional activation of I kappa B alpha in mediation of immunosuppression by glucocorticoids. Science 270, 283–286. Schiltz, J.C., Sawchenko, P.E., 1998. Interleukin-1 induction of cyclooxygenase-2 expression in ED2positive perivascular microglia. Soc. Neurosci. Abstr. 24, 1615. Shirakawa, F., Saito, K., Bonagura, C.A., Galson, D.L., Fenton, M.J., Webb, A.C., Auron, P.E., 1993. The human prointerleukin 1 beta gene requires DNA sequences both proximal and distal to the transcription start site for tissue-specific induction. Mol. Cell. Biol. 13, 1332–1344. Sorli, C.H., Zhang, H.J., Armstrong, M.B., Rajotte, R.V., Maclouf, J., Robertson, R.P., 1998. Basal expression of cyclooxygenase-2 and nuclear factor-interleukin 6 are dominant and coordinately regulated by interleukin 1 in the pancreatic islet. Proc. Natl. Acad. Sci. USA 95, 1788–1793.

788

S. Rivest / Psychoneuroendocrinology 26 (2001) 761–788

Stahl, N., Farruggella, T.J., Boulton, T.G., Zhong, Z., Darnell, J.E. Jr., Yancopoulos, G.D., 1995. Choice of STATs and other substrates specified by modular tyrosine-based motifs in cytokine receptors. Science 267, 1349–1353. Taga, T., 1997. The signal transducer gp130 is shared by interleukin-6 family of haematopoietic and neurotrophic cytokines. Ann. Med. 29, 63–72. Taga, T., Kishimoto, T., 1997. Gp130 and the interleukin-6 family of cytokines. Annu. Rev. Immunol. 15, 797–819. Thibeault, I., Laflamme, N., Rivest, S., 2001. Regulation of the gene encoding the monocyte chemoattractant protein 1 (MCP-1) in the rat brain in response to circulating LPS and proinflammatory cytokines. J. Comp. Neurol. 434, 461–477. Thiel, S., Behrmann, I., Dittrich, E., Muys, L., Tavernier, J., Wijdenes, J., Heinrich, P.C., Graeve, L., 1998. Internalization of the interleukin 6 signal transducer gp130 does not require activation of the Jak/STAT pathway. Biochem. J. 330, 47–54. Vallie`res, L., Lacroix, S., Rivest, S., 1997. Influence of interleukin-6 on neural activity and transcription of the gene encoding corticotropin-releasing factor in the rat brain: an effect depending upon the route of administration. Eur. J. Neurosci. 9, 1461–1472. Vallie`res, L., Rivest, S., 1997. Regulation of the genes encoding interleukin-6, its receptor, and gp130 in the rat brain in response to the immune activator lipopolysaccharide and the proinflammatory cytokine interleukin-1ß. J. Neurochem. 69, 1668–1683. Vallie`res, L., Rivest, S., 1999. Interleukin-6 is a needed proinflammatory cytokine in the prolonged neural activity and transcriptional activation of corticotropin-releasing factor during endotoxemia. Endocrinology 140, 3890–3903. Watanabe, S., Arai, K., 1996. Roles of the JAK–STAT system in signal transduction via cytokine receptors. Curr. Opin. Genet. Dev. 6, 587–596. Watanobe, H., Takebe, K., 1994. Effects of intravenous administration of interleukin-1-beta on the release of prostaglandin E2, corticotropin-releasing factor, and arginine vasopressin in several hypothalamic areas of freely moving rats: estimation by push–pull perfusion. Neuroendocrinology 60, 8–15. Webb, A.C., Collins, K.L., Auron, P.E., Eddy, R.L., Nakai, H., Byers, M.G., Haley, L.L., Henry, W.M., Shows, T.B., 1986. Interleukin-1 gene (IL1) assigned to long arm of human chromosome 2. Lymphokine Res. 5, 77–85. Wilkinson, L.S., Edwards, J.C., 1991. Binding of antibodies raised against tumour necrosis factor alpha (TNF alpha) to blood vessels and macrophages in inflamed synovial tissue. Rheumatol. Int. 11, 19–25. Williams, J.A., Shacter, E., 1997. Regulation of macrophage cytokine production by prostaglandin E2. Distinct roles of cyclooxygenase-1 and -2. J. Biol. Chem. 272, 25693–25699. Wissink, S., van Heerde, E.C., van der Burg, B., van der Saag, P.T., 1998. A dual mechanism mediates repression of NF-kappaB activity by glucocorticoids. Mol. Endocrinol. 12, 355–363. Xin, L., Blatteis, C.M., 1992. Hypothalamic neuronal responses to interleukin-6 in tissue slices: effects of indomethacin and naloxone. Brain Res. Bull. 29, 27–35. Zhang, J., Rivest, S., 1999. Distribution, regulation and colocalization of the genes encoding the EP2and EP4-PGE2 receptors in the rat brain and neuronal reponses to systemic inflammation. Eur. J. Neurosci. 11, 2651–2668. Zhang, J., Rivest, S., 2000. A functional analysis of EP4 receptor-expressing neurons in mediating the action of PGE2 within specific nuclei of the brain in response to circulating interleukin-1ß. J. Neurochem. 74, 2134–2145. Zohlnhofer, D., Graeve, L., Rose-John, S., Schooltink, H., Dittrich, E., Heinrich, P.C., 1992. The hepatic interleukin-6 receptor. Down-regulation of the interleukin-6 binding subunit (gp80) by its ligand. FEBS Lett. 306, 219–222.