- Top of page
The signalling pathways that mediate early central effects of interleukin-1 (IL-1) during the acute phase reaction have been poorly elucidated. Interaction of IL-1β to its specific receptor interleukin-1 receptor type I (IL-1RI) leads to nuclear factor kappa B (ΝFκB) nuclear translocation and a robust transcriptional activation of inhibitor of kappa B alpha (IκBα) within the rat brain. Indeed, we demonstrated that IL-1RI expressed in blood brain barrier (BBB) cells and in circumventricular organs (CVOs) is crucial for p65-NFκB translocation induced by peripheral injection of IL-1β. Moreover, it has been previously shown that monitoring IκBα mRNA synthesis is an effective tool to investigate the activity of the transcription factor NFκB into the CNS. However in the present study we observed time-related and cell-type differences between IκBα mRNA synthesis and p65-NFκB translocation. This indicates that the expression of IκBα mRNA does not strictly parallel p65-NFκB nuclear translocation, suggesting that these markers are not interchangeable to investigate NFκB activity but must be studied together. Thus, we hypothesize that IL-1β reached the brain across the CVOs that lack a BBB and endothelial cells all over the brain and interacted with its receptors to induce NFκB translocation. The study of the consequences of the impairment of NFκB pathway activation in in vivo experimentation should bring important clues about the precise role of this transcription factor.
Infectious and inflammatory processes evoke a broad spectrum of centrally mediated responses, including fever, somnolence, anorexia and altered metabolic activities (Konsman et al. 2002). The psychological and behavioural components of sickness represent a highly organized strategy of the organism to fight infection referred to as ‘sickness behaviour’. Interleukin-1β (IL-1β), produced by activated neutrophils and monocytes in contact with invading microorganisms, is an important cytokine for the induction of sickness behaviour. Peripheral and central administration of IL-1β induces all the central components of the acute phase reaction and inhibition of its action by central administration of the IL-1 receptor antagonist attenuates cytokine-induced sickness behaviour (Dantzer 2001a).
Although the recent elucidation of the signal transduction pathways activated by IL-1 receptors (Fitzgerald and O'Neill 2000) has prompted a surge of interest for the receptor-mediated mechanisms of IL-1β actions in the brain, the signalling pathways that mediate the behavioural effects of IL-1β have not yet been elucidated. IL-1β has been proposed to induce the synthesis of small lipophylic mediators such as prostaglandins or nitric oxide at the blood brain barrier (BBB) (Pahl 1999), which in turn are thought to diffuse into the brain parenchyma and activate key brain structures mediating behavioural and physiological responses (Dantzer 2001b).
Central action of IL-1β is mediated by brain IL-1 receptors that are similar to that of the family cloned from blood cell lineages. An overview of the distribution of interleukin-1 receptor type I (IL-1RI) was reported in the rat brain using in situ hybridization that identified labelling predominantly over brain barrier-related cells, including the leptomeninges, non-tanycytic portions of the ependyma, the choroid plexus, and vascular endothelium. Low to moderate levels of IL-1RI mRNA were detected in just a few neuronal cell groups, including the basolateral nucleus of the amygdala, the arcuate nucleus of the hypothalamus, the trigeminal and hypoglossal motor nuclei and the area postrema (Ericsson et al. 1995).
Binding of IL-1β drives dimerization of IL-1RI with its accessory protein followed by recruitment and phosphorylation of the receptor-associated kinases (IRAK1 and IRAK2) via the docking molecule MyD88. IRAK1 subsequently interacts with TNF receptor-associated factor 6 (TRAF6). This is followed by phosphorylation of nuclear factor κB (NFκB)-inducing kinase (NIK), which phosphorylates the IκB kinases, resulting in the release of IκB and translocation of NFκB to the nucleus. The DNA binding nuclear form of NFκB is usually an heterodimer which typically includes one 50 kDa (p50) and one 65 kDa (p65) polypeptide (Beg and Baldwin 1993; Baeuerle and Henkel 1994; Moynagh et al. 1994). In the cytosol NFκB dimers occur in complex with isoforms of the inhibitory protein IκB. Multiple IκB proteins exist, among them IκBα, which associates either with p65/p50 NFκB complexes or with homodimer of p65 (Whiteside et al. 1997). Once within the nucleus, NFκB binds to its consensus sequence on target genes that promote transcription of a variety of genes like IκBα, immunoreceptors, cytokines, chemokines and a few neuromediators (Miyamoto and Verma 1995).
In previous studies, Quan et al. (1997) and Laflamme and Rivest (1999) observed a robust transcriptional activation of IκBα within the rat brain in response to different models of systemic immunogenic stimuli. The bacterial endotoxin (lipopolysaccharide, LPS) or IL-1β injected by intraperitoneal route (i.p.) caused a rapid and intense expression of IκBα mRNA in the endothelium of the brain capillaries, parenchymal astrocytes and microglia. These authors proposed that monitoring IκBα mRNA synthesis is an effective tool to investigate the activity of the transcription factor NFκB into the CNS because IκBα is ubiquitous, parallels NFκB activity duration of the activating extracellular stimulation and responds to most form of extracellular stimulation. However, a number of recent studies demonstrated that nuclear NFκB translocation and IκBα mRNA synthesis are uncoupled due to NFκB-independent IκBα mRNA regulation, as in the case of glucocorticoids or norepinephrine (NE) (Farmer and Pugin 2000; Gavrilyuk et al. 2002), two factors induced after an immune challenge (Kabiersch et al. 1988; Besedovsky et al. 1991). Consequently, in order to obtain clear-cut responses concerning NFκB pathway activation after i.p. and intracerebroventricular (i.c.v.) injection of IL-1β, it is important to measure both IκΒα mRNA synthesis and NFκB p65 nuclear translocation, as hallmarks of inflammation and/or IL-1R activation. By studying, in vivo, the activation of NFκB at the brain level, we sought to detect the very early and primary response of the brain to circulating or intracerebrally injected IL-1β. To investigate the temporal, spatial and cellular distribution of NFκB in rat brain sections by immunohistochemistry, we used a specific anti-p65 antibody as previously described (Parnet et al. 2003). A double-immunohistochemical technique was used to identify the activated cells stimulated by IL-1β.
So far, previous studies have been focused on in situ hybridization localization of IL-1RI but never paralleled the presence of IL-1RI protein and molecular markers of its activity in the CNS. Using a specific sheep anti-rat IL-1RI antibody (Konsman et al. 2000b) we revealed the presence of functional receptors in the CNS of IL-1β-treated rat by colocalization with nuclear translocated p65-NFκB. The importance of IL-1RI in transducing this response was confirmed by the use of IL-1RI deficient mice.
Mice deficient for the expression of IL-1RI were generated by the process of gene targeting in murine embryonic stem cells (Glaccum et al. 1997) and have been shown to no longer express non-specific symptoms of disease after i.p. or i.c.v. injection of IL-1β (Bluthe et al. 2000). Adult male IL-1RI–/– mice used for these studies were random C57Bl/6 × 129/SVJ hybrids obtained from Immunex (Seattle, WA, USA).
C57Bl/6 × 129/SVJ F2,6−8 weeks of age and weighing 30–35 g were used as wild type (WT) controls for all experiments and purchased from CDTA-CNRS, Orléans, France.
Mice were housed in groups of 10 in 42 × 22 × 17 cm polypropylene cages and maintained in standard conditions with a reverse dark–light cycle (lights off from 6.00 am to 6.00 pm; temperature 23 ± 1°C). Food (Extralabo, Provins, France) and water were available ad libitum.
Recombinant rat IL-1β (rrIL-1β, biological activity: 317 IU/mg, NIBSC, Potters Bar, UK) was dissolved in sterile saline containing 0.1% endotoxin-free bovine serum albumin (Sigma-Aldrich Corporation, St Louis, MI, USA). The dose of rrIL-1β (20 µg/rat and 3 µg/mouse, i.p.; 70 ng/rat, i.c.v.) was selected on the basis of previous experiments on rat (Anforth et al. 1998) or mouse (Cremona et al. 1998) as demonstrated by its ability to induce the whole spectrum of clinical signs of sickness. Control mice and rats were injected with physiological saline.
Rats or mice that had received i.p. injection of rrIL-1β or saline were killed by decapitation 0, 15 min, 30 min, 1 h, 2 h or 6 h post-treatment. Rats with i.c.v. injection of rrIL-1β (70 ng/rat) or saline were killed by decapitation 0, 15 min, 30 min, 1 h, 2 h or 4 h post-treatment.
A polyclonal antibody generated to NFκB p65 subunit (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) (SC 372-G, lot#E132 for rats; C-20, lot#G016 for mice) was used in this study at a final dilution of 1 : 500. This antibody recognized both the inactive form of p65 subunit, bound to p50 and IκB in the cytoplasm, and the active monomeric form in the nucleus. Immunohistochemical processing was performed on free-floating sections using the streptavidin–biotin–immunoperoxidase technique. After washing off the cryoprotectant solution, sections were incubated for 2 min in phosphate-buffered saline containing 0.3% Triton X-100. Endogenous peroxidase activity was quenched by treating the sections in 3% H2O2 to phosphate-buffered saline for 30 min. After 30 min in blocking buffer (phosphate-buffered saline containing 0.2% gelatin and 5% non-fat milk), the primary antibody was added for 48 h at 4°C (0.1 mg/mL). After three washes in phosphate-buffered saline, sections were incubated for 2 h at room temperature with biotinylated donkey anti-goat IgG (Amersham Biosciences, Buckinghamshire, UK for rats; Biosys, France for mice) at 1 : 800 in phosphate-buffered saline containing 3% bovine serum albumin. After rinses, avidin-biotin peroxidase (Vectastain ABC kit), diluted in phosphate-buffered saline according to the manufacturer's instructions (Vector laboratories, Burlingame, CA, USA), was added for 2 h at room temperature. The peroxidase reaction product was developed using diaminobenzidine and the nickel-enhanced glucose oxidase method (Shu et al. 1988). Sections were then mounted onto gelatin-coated glass, dried, dehydrated through alcohol to xylene and coverslipped for light microscopic analysis.
To check the specificity of polyclonal anti-p65 antibody, immunohistochemistry was performed on IL-1-stimulated brain sections with polyclonal anti-p65 antibody diluted at 1 : 300 and incubated for 30 min at 37°C in the presence of 100-fold excess of the specific blocking peptide (C-20, 200 µg/mL; Santa Cruz Biotechnology) (data not shown).
Sections were incubated with the same anti-NFκB antibody as described above, in the presence of neuron marker (NeuN) or microglial and phagocytic cell markers (Ox42, ED1) for 48 h at 4°C in phosphate-buffered saline 1×, 0.25% bovine serum albumin and 0.1% Triton X-100. The antibodies dilutions were the following: a mouse polyclonal anti-NeuN antibody (Chemicon International, Euromedex, Souffelweyersheim, France) was diluted at 1 : 1000, mouse polyclonal anti-ED1 antibody and mouse polyclonal anti-Ox42 antibody were diluted at 1 : 200 (Serotec, Raleigh, NC, USA). After rinses, sections were incubated with a biotinylated donkey anti-goat secondary antibody (1 : 800; Amersham Biosciences) followed by incubation in the presence of streptavidin Alexa 488-conjugated (1 : 1000) for 2 h at room temperature to reveal NFκB. Simultaneously, anti-NeuN, Ox42 and ED1 were revealed with an anti-mouse antibody Alexa 594-coupled (1 : 1000) (Molecular Probes, Interchim, Eugene, OR, USA). The astrocytic marker (glial fibrillary acidic protein, GFAP) was revealed by a ‘two-step’ method: NFκB detection was performed as described above prior to incubation with a rabbit polyclonal anti-GFAP antibody diluted at 1 : 5000 (Gautron et al. 2002) and incubated for 24 h at 4°C. GFAP antibody was then revealed with an Alexa 594-conjugated anti-rabbit antibody (1 : 1000) (Molecular Probes).
Western blot analysis
- Top of page
Mice were killed by decapitation 1 h after injection and the brain removed. Brains were homogenized on ice in 5 mL of buffer A (Hepes 10 mm pH 7.9, MgCl2 1.5 mm, KCl 10 mm, dithiothreitol 1 mm, aprotinin 1 mg/mL, benzamidine 1 mm, Na orthovanadate 1 mm, NaF 1 mm, 4-(2-aminoethyl)-benzene-sulfonyl fluoride 1 mm, leupeptin 1 mm, pepstatin 1 mm). After 10 min of centrifugation at 850 g at 4°C, the pellets were resuspended in 4 mL of buffer A with 0.1% Triton X-100. After 10 min of incubation on ice and 10 min of centrifugation at 850 g at 4°C, the pellets enriched in nuclear fraction were resuspended in 2 mL of buffer B (Hepes 20 mm pH 7.9, glycerol 25%, NaCl 420 mm, MgCl2 1.5 mm, EDTA 0.2 mm, supplemented with antiproteases cocktail), incubated for 30 min on ice and centrifuged for 15 min at 20 000 g. The final pellet of nuclei was resuspended in buffer B and then sonicated. Protein contents were determined by Bio-Rad protein assay according to the manufacturer's protocol (Bio-Rad, Paris, France) and heated to 100°C for 5 min in Laemmli sample buffer (2% sodium dodecyl sulfate and 5% dithiothreitol) to entirely denature the proteins. Subsequently, equal quantities of proteins (40 µg/lane) were electrophoresed onto an 8% polyacrylamide gel with a 5% stacking gel. In parallel, proteins were detected by Coomassie Blue R-250 staining. Proteins were blotted on PVDF membranes (Immobilon p, Millipore, Paris, France) in Mini-protean II (Bio-Rad). Membranes were first saturated by incubation with 5% (wt/vol) milk in Tris-buffered saline-Tween (Tris-HCl pH 7.5/NaCl 100 mm/Tween-20 0.1%) for 1 h and incubated overnight with the primary goat polyclonal anti-p65 NFκB antibody (Santa Cruz Biotechnology; 1 : 3000). After washings in Tris-buffered saline-Tween 0.1%, membranes were incubated with the secondary Ab-HRP (Southern Biotechnology Associates, Birmingham, AL, USA) diluted in Tris-buffered saline-Tween 0.1% supplemented with 3% milk, for 2 h at room temperature, then washed in Tris-buffered saline-Tween 0.1% for 15 min, then Tris-buffered saline-Tween 0.05% twice for 5 min and finally in Tris-buffered saline alone. The complex was detected with ECL (Enhanced ChemiLuminescence) reagents and exposed on Hyperfilm ECL (Amersham, Orsay, France). Optical density analysis of the obtained signal has been performed with the GS-800 calibrated densitometer (Bio-Rad).
- Top of page
Total cellular RNA extraction was performed by using a RNA Now kit (Biogentex, Seabrook, TX, USA) and following the manufacturer's instructions. After RNA extraction, 1 µg of purified total RNA was used for reverse transcription in a final volume of 20 µL. Primers sequences were: IκBα 5′-CCAGCTGGGCAGGGCCAG, 3′-TAACGTCAGACGCTGGCC; β2mgl (β2microglobulin) 5′-ATCTTTCTGGTGCTTGTCTC, 3′-AGTGTGAGCCAGGATGTAGT. Quantities of cDNA as well as the number of cycles of amplification were determined in previous experiments so as to be in the exponential phase of amplification for each sample. PCR amplification was carried out on 1 µL of the cDNA of each sample and 30 cycles of amplification were used to amplify β2mgl and IκBα (5 min 94°C, 1 min 94°C, 1 min Tm, 1 min 72°C where Tm = 55°C for β2mgl, Tm = 58°C for IκB) in the presence of (α-32P)dCTP (Amersham, 3000 Ci/mmol). PCR products of IκBα and β2mgl did not reach a plateau in the conditions used. Twenty µL of PCR products were size-separated by PAGE. Determination of incorporated radioactivity was performed by using a phosphorimager (Molecular Dynamics, Sunnyvale, CA, USA). Results obtained as arbitrary units of amplification were expressed as the ratio of IκBα/(β2mgl) × 100.
Time course and distribution of activation of NFκB pathway in rat brain after an intraperitoneal injection of IL-1β
- Top of page
No NFκB translocation was observed in saline-treated animals (Figs 1a–e and Figs 2a–c). However, cytoplasmic p65 labelling could be detected in choroid plexus (Fig. 1e), in a few cells of the area postrema (Fig. 1c) and in cells lining the ventricles (Fig. 2a) in control animals.
Figure 1. Distribution of the nuclear translocation of p65-NFκB subunit in the CVOs of the rat brain in response to an i.p. injection of rrIL-1β. Intraperitoneal injection of saline did not induce NFκB translocation. No nuclear staining was seen in SFO (a), OVLT (b), area postrema (c) or median eminence (d). Choroid plexus presents a cytoplasmic signal only (e). On the contrary i.p. treatment with rrIL-1β (2 h) was responsible for NFκB translocation in various structures: SFO (f), OVLT (g), area postrema (h) or median eminence (i), choroid plexus (j). Arrowheads show NFκB staining in the parenchyma of the CVOs and in the NTS, and arrows point out endothelial-like staining. NTS: nucleus of the tractus solitari; 3V: third ventricle. Scale bar = 100 µm (a–c, f–h); Scale bar = 50 µm (d, e, i, j).
Download figure to PowerPoint
Figure 2. Distribution of the nuclear translocation of p65-NFκB subunit in endothelial cells, meninges and ependymal cell layer of ventricles of the rat brain in response to an i.p. injection of rrIL-1β. Intraperitoneal treatment with rrIL-1β (2 h) induced NFκB translocation in the ependymal cell (d), the blood vessels (e1, e2, e3) and the meninges (f), whereas no staining was seen after saline injection in the blood vessels (b1, b2, b3) or the meninges (c) and only a cytoplasmic staining in the ependymal cell (a). b1 and e1 represent a longitudinally sectioned blood vessel that is magnified on b2 and e2. b3 and e3 correspond to a transversally sectioned vessel. 3V: third ventricle; arrowheads: nuclear staining; arrow: cytoplasmic staining. Scale bar = 100 µm (b1 and e1); Scale bar = 50 µm (a and d, b2, b3, e2 and e3, c and f).
Download figure to PowerPoint
On sections of brain from IL-1β-treated rats, p65-NFκB immunoreactivity was observed in the nuclear compartment of cells present in defined structures. Table 1 summarizes the time course of induction of nuclear NFκB detected by immunolabelling in the brain of IL-1β-treated rats. A rapid NFκB translocation after 15 min of stimulation by IL-1β was detected in most of the brain vasculature (Fig. 2e1–e3, Table 1). After 30 min of treatment, NFκB translocation was detected in a few circumventricular organs (CVOs) such as organ vascularis of the lamina terminalis (OVLT) (Fig. 1g), area postrema (Fig. 1h), and median eminence (Fig. 1i). The subfornical organ (SFO) (Fig. 1f), choroid plexus (Fig. 1j), third ventricle (Fig. 2d) and meninges (Fig. 2f) are positive at 1 h of treatment, when nuclear NFκB immunolabelling in the previously cited structures peaked. Additionally, translocated p65 was observed in a structure adjacent to the area postrema, the nucleus of the tractus solitari (NTS) (Fig. 1h). After 6 h of IL-1β treatment, no NFκB translocation could be detected in any structure.
Table 1. Time course of the influence of an intraperitoneal injection of IL-1β on the translocation of NFκB-p65 subunit
| ||IL-1β i.p. 0 min|| 15 min|| 30 min|| 1 h|| 2 h|| 6 h|
|Blood vessels||–||+||+||+ +||+||–|
|Choroid plexus||–||–||–||+ + +||+ + +||–|
|Ventricles||–||–||–||+ + +||+ +||–|
|SFO||–||–||–||+ +||+ +||–|
|Median eminence||–||–||+||+ +||+ +||–|
|Area postrema||–||–||+||+ +||+ +||–|
Time course and distribution of NFκB in rat brain after an intracerebroventricular injection of IL-1β
- Top of page
Following i.c.v. injection of saline in rat brain, some NFκB translocation was observed in ependymal cells of the cerebral ventricles, in choroid plexus of the lateral ventricle and in the meninges due to cannula implantation as compared to non-cannulated animals, which did not display any NFκB translocation in these same structures as illustrated on Fig. 3 (a1–3). An i.c.v. injection of IL-1β induced translocation of p65-NFκΒ in meninges (Fig. 3b5), ependyma (Fig. 3b6), choroid plexus (Fig. 3b7), but also throughout the brain's vasculature (Fig. 3b8i–iii), and in the CVOs (Fig. 3b1–4) (i.e. SFO (Fig. 3b1), area postrema (Fig. 3b2), OVLT (Fig. 3b3), and median eminence (Fig. 3b4). Evidence of NFκB translocation was seen until 4 h post-treatment in the CVOs (Table 2).
Figure 3. (a) Basal nuclear translocation of p65-NFκB subunit under cannulation. Few p65-NFκB translocated cells were detected in meninges (a1), ependymal cells lining the brain ventricles (a2) and in choroid plexus (a3). Arrowheads show NFκB nuclear translocation. Arrows show cytoplasmic NFκB. Scale bar = 50 µm. (b) Distribution of the nuclear translocation of p65-NFκB subunit in the rat brain in response to an i.c.v. injection of rrIL-1β. Intracerebroventricular treatment with rrIL-1β (2 h) was responsible for NFκB translocation in various structures as SFO (1), area postrema (2), OVLT (3), median eminence (4), meninges (5), ependymal cells lining the brain ventricles (6), choroid plexus (7). Arrowheads show NFκB staining in the parenchyma of the CVOs, and arrows point out endothelial-like staining around blood vessels (8i–iii). LV: lateral ventricle; 3V: third ventricle. Scale bar = 100 µm (1–4 and 8i); scale bar = 50 µm (5–7, 8ii and 8iii).
Table 2. Time course of the influence of an intracerebroventricular injection of IL-1β on the translocation of NFκB-p65 subunit
| ||IL-1β i.c.v. 0 min|| 15 min|| 30 min|| 1 h|| 2 h|| 4 h|
|Blood vessels||+||+ +||+||+||+||+|
|Choroid plexus||+||+ +||+||+||+||+|
|Ventricles||+||+ +||+ +||+ +||+ +||+|
|Area postrema||–||+ +||+ +||+||+||+ +|
IL-1β-mediated NFκB nuclear translocation in endothelial cells and astrocytes
- Top of page
After immunoperoxidase staining, p65-NFκB was present in the nuclei of endothelial cells of most of the cerebral blood vessels (Fig. 2e1–3) as well as epithelial cells of the choroid plexus (Fig. 1j) and ependymal cells of the ventricles (Fig. 2d).
The phenotype of cells displaying nuclear NFκB was further characterized by double labelling using two distinct fluorochromes. Most of the cells displaying NFκB nuclear immunoreactivity were GFAP positive all over the brain sections, in the CVOs and around the blood vessels (Fig. 4). NeuN, Ox42 and ED-1 positive cells did not demonstrate nuclear but only cytoplasmic NFκB localization (Fig. 5).
Figure 4. (a–f) Double-labelling fluorescent staining using first anti-p65 NFκB antiserum then anti-GFAP antibody on representative rat brain sections after i.p. injection of IL-1β. Transversally sectioned vessels stained for p65-NFkB (green round staining) and GFAP (red staining) are visible on the first lane. The second lane represents p65-NFkB and GFAP staining in longitudinally sectioned capillary. The left column shows NFkB labelling (green), the middle column shows GFAP labelling (red) and the right column is the merge of the two types of labelling. In each case, some p65-NFkB positive cells are colocalized with GFAP. p65-NFkB staining that is not colocalized with GFAP staining was morphologically representative of endothelial cells (c and f). Scale bar = 50 µm.
Download figure to PowerPoint
Figure 5. Double-labelling fluorescent staining using simultaneously anti-p65 NFκB antiserum (b, c, e, f, h, i) and anti-ED1 antibody (a, c), anti-OX42 antibody (d, f),or anti-NeuN antibody (g, i), on rat brain sections after i.p. injection of IL-1β. The left column shows NFκB labelling (green), the middle column shows cell markers staining (red) and the right column is the merge of the two labelling. No colocalization is seen for p65-NFκB and Ox42 (f). NeuN (i) and ED1 (c) colocalized only with cytoplasmic NFκB. Scale bar = 50 µm.
Download figure to PowerPoint
IL-1RI and nuclear NFκB are both present on brain capillaries
- Top of page
Immunohistochemistry of IL-1RI revealed the presence of the receptor on brain capillaries of the CVOs and of several regions of the parenchyma. One hour after injection of IL-1β, the receptor was detected on the capillaries of the CVOs and in non-vascular elements of these structures as shown in the OVLT (Figs 6a and b). When compared to NFκB staining, both the transcription factor and IL-1RI were present on and around the capillaries (Figs 6c and d). Similar findings were obtained in the area postrema, median eminence and SFO.
Figure 6. Comparison of IL-1RI and p65-NFκB immunodetection on rat brain sections after i.p. injection of IL-1β. One hour after an i.p. injection of rrIL-1β, IL-1RI (a) and NFκB nuclear staining (b) were apparent in CVOs like OVLT. Higher magnification revealed their presence on cells lining the blood vessels (c: IL-1RI staining; d: NFκB staining). Double-labelling immunoperoxidase staining revealed colocalization of NFκB and IL-1RI in a representative cell lining a blood vessel (e). Arrowheads: IL-1RI staining; arrow: NFκB nuclear staining. Scale bar = 100 µm (a and b); scale bar = 50 µm (c, d, e).
Download figure to PowerPoint
In the rest of the brain, IL-1RI was only present in endothelial cells of blood vessels of several central nuclei: BLA, SON, median preoptic nucleus, medium septum, PVN and NTS (data not shown). On Fig. 6(c), a higher magnification illustrates IL-1RI immunostaining on blood vessels of the parenchyma. This labelling is correlated to NFκB immunostaining (Fig. 6d). Finally, Fig. 6(e) shows colocalization of NFκB with IL-1RI in these vessels.
NFκB translocation depends on activation of IL-1RI after intraperitoneal injection of rrIL-1β
- Top of page
To investigate the role of IL-1RI in transducing IL-1β response in the brain, IL-1RI knockout (KO) mice were used to perform p65-NFκB detection by western blot and immunohistochemistry. A representative western blot on nuclear extracts from WT and KO mice revealed the presence of a 65-kDa protein induced by IL-1β treatment and detected by the anti-p65 NFκB antibody. The induction was low in nuclear extracts from saline-treated WT mice, saline-treated KO mice and IL-1β-treated KO mice compared to IL-1β-treated WT mice (Fig. 7a).
Figure 7. (a) p65-NFκB subunit detected by western blot in brain nuclear extracts from WT mice or IL-1RI KO mice. Translocated p65-NFκB was more abundant in WT mice compared to IL-1RI KO mice after treatment with IL-1β (1 h), as shown by a representative western blot. Histogram represents the relative optical density (OD) value of each extract obtained by densitometer measurement. p65-NFκB immunodetection in area postrema of WT (b, c) and IL-1RI KO (d, e) mice after i.p. injection of rrIL-1b. A strong nuclear staining of p65-NFκB was revealed in WT mice brain (b) and at higher magnification (c). On the contrary, KO mice did not display p65-NFκB translocation in the area postrema (d). The cells stained in that case showed a cytoplasmic labelling.
Download figure to PowerPoint
Immunoreactivity for p65-NFκB was detectable by immunohistochemistry in the brain from IL-1β-treated WT mice (Fig. 7b) but totally absent in brain from IL-1RI KO mice (Fig. 7d). The cells that remained stained in that case show a cytoplasmic staining (Fig. 7e) whereas it was clearly nuclear for WT mice (Fig. 7c). In IL-1β-treated WT mice, the time course of translocation and the cellular localization of NFκB were the same as detected in rat brain sections. A strong activation of NFκB appeared from 15 min and peaked at 1 h in CVOs, endothelial cells, meninges and ependymal cells layer (data not shown).
- Top of page
In the present work, a detailed time course and cellular location study of NFκB activity in the rat brain was performed in response to i.p. and i.c.v. IL-1β injection. The results of immunohistochemical localization of p65-NFκB show that the NFκB pathway was activated by both treatments but restricted to specific cell populations and structures, namely the CVOs, cells forming the BBB, or at the interface between the cerebrospinal fluid (CSF) and the brain parenchyma. Furthermore, this study is the first to correlate, in vivo, the presence of the protein IL-1RI and evidence of NFκB pathway activation in the same cells in response to IL-1β treatment. The essential role of IL-1RI in NFκB pathway activation is also demonstrated using IL-1RI KO mice. Original findings emerge from the present results that will be extensively discussed in comparison to previous experiments performed by several investigators (Quan et al. 1997; Laflamme and Rivest 1999; Proescholdt et al. 2002).
This is the first report demonstrating NFκB pathway activation by the visualization of p65-NFκB nuclear translocation in cells of the rat brain after an i.p. or i.c.v. administration of IL-1β. As summarized in Tables 1 and 2, p65 immunoreactivity was detected 15 min after IL-1β treatment in most of the brain microvasculature. Then, the NFκB signal propagated to areas devoided of BBB and accessible to circulating cytokine (when injected i.p.) or to cytokine present in the CSF (when injected i.c.v.). After 1 h of treatment, meninges, ependymal layer of the cerebral ventricles, choroid plexus and CVOs became labelled. NFκB activation was extinguished 6 h after injection (when injected i.p.). The time course and intensity of immunostaining differed slightly between the two routes of treatment: NFκB activation occurred earlier after i.c.v. treatment and part of it was due to surgery, which induced a slight NFκB activation at the site of the cannulation but also in the ventricular ependyma near the injection site, in choroid plexus and basal meninges, as shown earlier by Proescholdt et al. (2002).
The effect of i.p. or i.v. (intravenous) IL-1β treatment on the activation of NFκB pathway along blood vessels has been largely discussed in previous work (Laflamme and Rivest 1999) and involved the presence of cell surface IL-1R that are reachable for circulating cytokine. To assess whether p65-NFκB nuclear translocation in response to IL-1 injection could be explained by the presence of IL-1RI on the activated cells, we examined the immunoreactivity for IL-1RI. We first used a specific antibody raised against rat IL-1RI (Konsman et al. 2000b) and demonstrated that IL-1RI was discreetly distributed in barrier cells, including blood vessels, meninges, epithelial cells of the choroid plexus and in CVOs. Labelling was also present in ependymal cells of the brain ventricles. Inside the CVOs, positive cells were located in blood vessels of the OVLT, SFO and area postrema and in undetermined cells of the parenchyma of these structures. The cellular localization of IL-1RI is in good agreement with IL-1RI mRNA detection (Ericsson et al. 1995; Herkenham et al. 1998) or IL-1 binding sites (Van Dam et al. 1996). Furthermore, after 1 h of stimulation by IL-1β, p65-NFκB labelling was colocalized with IL-1RI, suggesting a causal link between the receptor and the signalling pathway. However, it is known that cytokines act as a network fashion. In some previous works, we demonstrated that LPS injected i.p. induced synthesis of brain IL-1β, tumour necrosis factor alpha (TNFα) and IL-6 transcripts (Laye et al. 1994). Intracerebroventricular cotreatment with IL-1 receptor antagonist totally blocked the LPS-induced enhanced expression of proinflammatory cytokine mRNA measured in the hypothalamus 1 h after treatment, suggesting that IL-1β was responsible for TNFα expression in the brain (Laye et al. 2000). By consequence, as TNFα can activate NFκB pathway after interaction with its receptor (Traenckner et al. 1995), we cannot rule out the possibility of TNFα involvement in NFκB translocation after peripheral or central injection of IL-1β. This is reinforced by the fact that TNFα replaced IL-1β to mediate behavioural responses when activity of this last cytokine was deficient (Bluthe et al. 2000).
To test the requirement of IL-1 receptors in the observed responses, we used mice lacking IL-1RI and found that IL-1β treatment did not induce p65-NFκB translocation in endothelial cells or parenchymal cells of the CVOs. IL-1RI KO mice have been previously used to study the activation of NFκB at the pituitary level and confirm that the presence of this receptor was indispensable to allow IL-1 stimulation in somatotroph cells, GFAP positive cells and macrophage-like cells (Parnet et al. 2003).
Activation of cells in perivascular spaces after i.c.v. injection of IL-1 remains intriguing. However, recent data obtained by Mercier et al. (2003) point out the evidence of a possible link between the CSF compartment and perivascular cells via an extracellular matrix network. Therefore, IL-1β injected i.c.v. could travel through this extracellular matrix and consequently reach perivascular cells. Furthermore, Konsman et al. (2000a) demonstrated that IL-1β (injected i.c.v.) followed CSF flow in the ventricles and subarachnoid spaces of the major cisternae, entered perivascular spaces and diffused slowly into the brain along white matter and blood vessels. That result was correlated with NFκB translocation into ependymal cells of the ventricles, meninges, perivascular spaces and CVOs but not in brain parenchyma. A similar pattern of NFκB translocation was obtained in the present work after i.c.v. injection of IL-1β.
Taken together, these results suggest that exogenous IL-1β reaches the brain across the CVOs to interact with its receptors in those structures, but also that IL-1β targets endothelial cells all over the brain vasculature. The binding of IL-1β to its receptor induces NFκB translocation in all cells expressing IL-1RI.
To test whether LPS or IL-1β act on the rat brain and trigger a specific response, a number of studies detected de novo IκBα mRNA synthesis by in situ hybridization based on the concept that activation of transcription by NFκB induces the synthesis of IκBα (Quan et al. 1997; Laflamme and Rivest 1999; Proescholdt et al. 2002). Although most of the observations are similar to our findings, some discrepancies emerge.
One of the principal discrepancies between our data and these different studies concerned the type of cells involved in the response. If there is a general agreement about labelling of endothelial cells, epithelial cells of the choroid plexus, ependymal cells of the ventricles, additional labelled cells present in the brain parenchyma are questionable. Laflamme and Rivest (1999) and Quan et al. (2000) brought evidence that IκBα mRNA was colocalized with endothelial cells but also in scattered small cells throughout the brain parenchyma described as Ox42-positive microglia. Quan et al. (1997) also demonstrated that astrocytes synthesized IκBα mRNA in vivo after a peripheral immune challenge. In the present work a colocalization of p65-NFκB and GFAP immunoreactivity was demonstrated but not with ED1, Ox42 or NeuN, markers used to label brain macrophages, microglia (Milligan et al. 1991) and neurones, respectively. This means that astrocytes of the CVOs, endothelial cells of the brain vasculature, epithelial cells of the choroid plexus and ependymal cells of the ventricles are the privileged targets of exogen IL-1β or of newly synthesized proinflammatory cytokines like IL-1β itself or TNFα.
Additionally, although we visualized a single wave of p65-NFκB translocation after IL-1β challenge, we found two waves of IκBα mRNA synthesis after i.p. injection of IL-1β. The first wave of induction of IκBα mRNA occurred 2 h after stimulation and a second peaked at 6 h. On the contrary, i.c.v. injection of IL-1β induced a single wave of transcription that started 30 min after stimulation and peaked at 2 h. Two waves of IκBα mRNA induction following an i.p. injection of LPS have been previously reported. The first wave was restricted to the level of the BBB and CVOs and the second one reached parenchymal cells (Quan et al. 1997). For Laflamme and Rivest (1999), the most intense signal was detected at 1 h after i.v. injection of IL-1β, returned to basal level after 3 h but was not investigated later. In our study, the second wave of IκBα mRNA observed after an i.p. IL-1β injection was not associated with p65-NFκB translocation (see Table 1). Therefore, it is evident that the expression of IκBα mRNA does not strictly parallel p65-NFκB nuclear translocation and cannot be strictly used as an index of NFκB pathway activation. Molecular mechanisms and cell types responsible for the second wave of expression of IκBα mRNA may differ from those ones responsible for the induction of proinflammatory molecules. Indeed, a number of molecular characteristics of the NFκB pathway need to be taken in consideration when studying its activation. Recent advances in understanding the mechanism of NFκB activation demonstrate that pro-inflammatory genes expression and IκBα synthesis were associated with p65 homodimers, whereas the resolution of inflammation involved the p50–p50 homodimers to trigger expression of anti-inflammatory genes and IκBα (Lawrence et al. 2001). Therefore, studying p50 translocation in the CNS would be useful to characterize late IκBα mRNA distribution following IL-1β injection.
Second, a series of recent studies reported uncoupling between NFκB translocation and IκBα mRNA synthesis. Adrenal glucocorticoids (GCs) and NE are potent stimulators of IκBα mRNA expression (Farmer and Pugin 2000; Almawi and Melemedjian 2002; Gavrilyuk et al. 2002). NE induces an increase of IκBα, on primary culture of astrocytes, at both transcriptional and post-transcriptional levels, independently of p65-NFκB nuclear translocation (Gavrilyuk et al. 2002). GCs can induce IκBα expression in vitro or in vivo and antagonize NFκB activity. Alternatively, GCs may act through a direct protein–protein interaction involving complexing of GCs/GC receptor and NFκB binding to DNA (Auphan et al. 1995; Scheinman et al. 1995; Quan et al. 2000; Almawi and Melemedjian 2002). Accordingly, Quan et al. (2000) demonstrated that the rise in GCs observed after a peripheral injection of LPS selectively induced cerebral IκBα mRNA expression which serves as a negative feedback mechanism of LPS-induced synthesis of proinflammatory cytokines. As a consequence, one needs to compare p65-NFκB nuclear translocation and IκBα mRNA synthesis in order to separate the primary pro-inflammatory effects of IL-1 from the anti-inflammatory effects, which both lead to IκBα mRNA synthesis.
Pharmacological studies correlated to behavioural analysis have been used to determinate accurate doses of IL-1β injected i.p. or i.c.v. to induce, in the rat, physiological modifications such as fever, anorexia, reduction of activities similar to those experienced by patients during an episode of viral or bacterial infection (Anforth et al. 1998). Brain response lasts 12 h, is reversible and is responsible for the reorganization of the perception and actions of the organism to deal with infection in the most efficient way (for review, see Konsman et al. 2002). The context chosen to perform NFκB activity detection in the brain is important to consider when comparing works with different time course analysis or different immune challenges. The study of the consequences of the impairment of NFκB pathway activation in in vivo experimentation should bring important clues about the precise role of this transcription factor: is its role limited to the communication between the peripheral immune system and the CNS or does it extend to a protective role in resolving brain inflammation (Lawrence et al. 2001)?