The present study was designed to determine the role of endogenous brain interleukin (IL)-1 in the anorexic response to lipopolysaccharide (LPS). Intraperitoneal administration of LPS (5–10 μg/mouse) induced a dramatic, but transient, decrease in food intake, associated with an enhanced expression of proinflammatory cytokine mRNA (IL-1β, IL-6, and tumor necrosis factor-α) in the hypothalamus. This dose of LPS also increased plasma levels of IL-1β. Intracerebroventricular pretreatment with IL-1 receptor antagonist (4 μg/mouse) attenuated LPS-induced depression of food intake and totally blocked the LPS-induced enhanced expression of proinflammatory cytokine mRNA measured in the hypothalamus 1 h after treatment. In contrast, LPS-induced increases in plasma levels of IL-1β were not altered. These findings indicate that endogenous brain IL-1 plays a pivotal role in the development of the hypothalamic cytokine response to a systemic inflammatory stimulus.
- receptor antagonist
- food intake
proinflammatory cytokines mediate the local and systemic components of the acute-phase response. Interleukin (IL)-1 is one of the key cytokines for the development of the centrally mediated signs of sickness, including depression of food intake and food-motivated behavior (8, 28). Peripheral and central administration of recombinant IL-1β decreases food intake in rats and mice (21, 26). This effect is abrogated by pretreatment with the IL-1 receptor antagonist (IL-1ra) (19, 25, 27). IL-1β is synthesized in the hypothalamus, together with other proinflammatory cytokines including tumor necrosis factor-α (TNF-α) and IL-6, in response to peripheral inflammatory stimuli (11, 23). Proinflammatory cytokines act in a network fashion at the periphery, meaning that each cytokine induces its own synthesis and the synthesis of other cytokines that potentiate or oppose its effects. The possibility that a similar network exists in the brain has been suggested by time-course studies of the expression of transcripts of proinflammatory cytokines in the brain in response to peripheral administration of the cytokine inducer lipopolysaccharide (LPS). Intraperitoneal administration of a behaviorally active dose of LPS increases the expression of TNF-α and IL-1β mRNA in the brain within 1 h after the treatment, and 3 and 6 h later IL-6 and IL-1ra mRNA are expressed (11, 23).
To determine whether endogenous brain IL-1 plays a pivotal role in LPS-induced behavioral depression, we measured food intake and proinflammatory cytokine mRNA expression in the hypothalamus of mice that were treated by a peripheral injection of LPS and a central injection of IL-1ra. To ensure that centrally injected IL-1ra blocked exclusively brain IL-1β, we measured circulating levels of IL-1β.
We confirmed that LPS induced a decrease in food intake that was accompanied by an increase in proinflammatory cytokine synthesis in the hypothalamus. LPS-induced anorexia was partially blocked by central administration of IL-1ra. LPS-induced cytokine mRNA expression in the hypothalamus was fully blocked by IL-1ra, whereas increased plasma levels of IL-1β were not affected.
MATERIALS AND METHODS
Animals and materials.
Seven-week-old male mice of the CD-1 (ICR) BR strain (Charles River), weighing 30–35 g at the start of the experiment, were housed in polypropylene cages in a room that was maintained at 23 ± 1°C. Lights were turned on from 8 PM to 8 AM, and food and water were available ad libitum except for food intake experiments. LPS (Escherichia coli, serotype 0127:B8) was obtained from Sigma Chemical (St. Louis, MO). This serotype was used because it had been demonstrated to reliably increase body temperature and metabolic rate in the rat, despite the fact that the rat is relatively insensitive to endotoxin (15). Recombinant mouse IL-1β and sheep antibodies to mouse IL-1β were obtained from National Institute for Biological Standards and Control (Potters Bar, UK). IL-1ra was obtained from Synergen (Boulder, CO). pMus3 was a kind gift from D. Shire (Sanofi, Labèges, France). Avidin-horseradish peroxidase was purchased from Dako and RPMI and calf serum from GIBCO-BRL (Cergy Pontoise, France).
Surgery and treatments.
The experiments were conducted according to Authorization for Practicals on Animals ethical standards for humane treatment of animals and the French legislation concerning animal experimentation.
For intracerebroventricular injections, a 23-gauge stainless steel guide cannula was inserted stereotaxically over the lateral cerebral ventricle 10 days before the experiment in mice anesthetized with a mixture of ketamine (1.8 mg/kg) and xylazine (12.2 mg/kg) at 10 ml/kg body wt ip, as previously described (6, 7). Coordinates for the guide cannula were 0.6 mm posterior to bregma, 1.6 mm lateral, and 2 mm below the skull surface at the point of entry. Cannulas were secured with two stainless steel screws and dental cement. Mice were allowed to recover for 1 wk before experiments were performed. Intracerebroventricular injections (2 μl) were made by gravity, after a 30-gauge needle was lowered in the guide cannula (6, 7). Mice were given an intracerebroventricular injection of saline or IL-1ra (4 μg/mouse) immediately followed by an intraperitoneal injection of saline or LPS (5 or 10 μg/mouse). Mice were killed by decapitation 1 h after the injection (n = 5 for each experimental group), and the hypothalamus was rapidly dissected and frozen at −80°C until use for RT-PCR. Blood was collected on ice in 10 U/ml endotoxin-free heparin, and plasma was stored at −80°C until determination of IL-1β concentrations by ELISA. The 1-h delay was selected on the basis of the time-course experiment showing the induction of IL-1β, TNF-α, and IL-6 mRNAs in the hypothalamus of mice injected intraperitoneally with LPS (22). Doses of LPS and IL-1ra were selected on the basis of previous experiments showing that 5 and 10 μg of LPS reliably induce sickness behavior and proinflammatory cytokine mRNA expression in mice (3, 22, 23) and 4 μg of IL-1ra fully block the behavioral effects of IL-1β (unpublished data).
Experimental design and food consumption measurement.
Mice were randomly allocated to two experimental groups receiving intracerebroventricular IL-1ra (4 μg/mouse, n = 7) or saline (n = 5) treatments. In each experimental group, the animals were used as their own control and intraperitoneally injected with LPS (5 μg/mouse) or saline in a randomized order immediately after the intracerebroventricular treatment. For the measurement of food intake, mice were provided with a food pellet placed on the floor of their home cage. Cumulative food intake was measured by weighing food (precision 0.1 g) before the injection and at 2, 4, 8, 12, and 24 h after injection.
Semiquantitative RT-PCR analysis of hypothalamic cytokine transcripts.
Total cellular RNA extraction was performed with a kit (RNA Now, Biogentex, Seabrook) following the manufacturer's instructions. After RNA extraction, 2 μg of purified total RNA were used for reverse transcription in 20 μl. For PCR amplification, 0.5 μl (β2-microglobulin) and 3 μl (cytokines; for sequences of the primers see Ref. 22) of the cDNA of each sample were used. Thirty cycles of amplification were used to amplify β2-microglobulin, IL-1β, IL-6, and TNF-α (1 min at 94°C, 1 min at 60°C, 1 min at 72°C) in the presence of [α-32P]dCTP (Amersham, 3,000 Ci/mmol) and an internal control (pMus3). pMus3 (10 pg/amplification) was calibrated to be used in the same amount for the coamplification of cytokines and β2-microglobulin performed in the same conditions (data not shown). PCR products of proinflammatory cytokines and β2-microglobulin did not reach a plateau in the conditions used (30 cycles, pMus 10 pg/amplification). Twenty microliters of cDNA products were separated according to size. The amount of radioactivity incorporated was quantified with a PhosphorImager (Molecular Dynamics). To compare the expression of cytokine mRNAs in the different experimental groups, results were expressed as the ratio of cytokines per insert to β2-microglobulin per insert × 100.
Plasma IL-1β measurement by ELISA.
Plasma levels of IL-1β were determined by ELISA, as previously described (22). Briefly, blood was collected on ice in 10 U/ml endotoxin-free heparin (Sigma Chemical). Microtiter plates were coated with sheep antibody to mouse IL-1β (2 μg/ml in PBS, pH 7.2, at 4°C). After addition of samples and standards (recombinant mouse IL-1β, 1.9–1,000 pg/ml) diluted in PBS with 0.5% BSA, plates were incubated for 24 h at 4°C, and the secondary biotinylated antibody to mouse IL-1β (1:1,000 in 2% ovalbumin) was added. After incubation, avidin-horseradish peroxidase (1:5,000) was added and the colorimetric reaction was carried out using 0.4 mg/ml orthophenylene diamine (Sigma Chemical) and 0.01% H2O2 in substrate buffer (34.7 mM citric acid, 66.7 mM Na2HPO4, pH 5.0). Absorbance was read at 490 nm by microplate reader spectrophotometer (Dynatech). The detection limit was 1.9 pg/ml mouse IL-1β. Data are expressed as picograms per milliliter of plasma.
Food consumption results were analyzed using a three-way ANOVA with saline vs. IL-1ra as between-group factor and saline vs. LPS and time as within-group factors. When the interaction was significant, post hoc analyses were performed using Newman-Keuls tests. Significance was set at P < 0.05. Results from RT-PCR and ELISA are expressed as means ± SE. Data were analyzed using ANOVA (ANOVA, treatment factor) followed by post hoc tests with the Newman-Keuls test.
Baseline food intake did not differ between treatments (P > 0.05; Fig. 1). LPS significantly depressed food intake for the duration of measurement (P < 0.001). Intracerebroventricular administration of IL-1ra significantly attenuated the depression in food intake induced by LPS at 4, 6, 8, and 10 h after treatment (P < 0.05). Mice treated with IL-1ra intracerebroventricularly and LPS intraperitoneally ate significantly less than mice treated with saline intracerebroventricularly and intraperitoneally when food intake was measured 8 and 10 h after treatment (P < 0.05). These results show that intracerebroventricular administration of IL-1ra attenuated the decrease in food intake induced by LPS.
To assess the role of IL-1 in the induction of the expression of proinflammatory cytokines in the hypothalamus in response to LPS, mice were pretreated intracerebroventricularly with IL-1ra or saline just before they were injected intraperitoneally with LPS (10 μg/mouse) or saline. The effect of the different treatments under study was assessed by semiquantitative RT-PCR with 10 pg of insert per sample. Figure 2 illustrates a typical RT-PCR gel showing a weak basal expression of IL-1β, IL-6, and TNF-α mRNAs and a marked induction of the expression of cytokine mRNAs in the hypothalamus in response to LPS. pMus amplification was the same in all experimental groups. After intracerebroventricular IL-1ra treatment, LPS-induced cytokine mRNA expression was attenuated (Fig. 2). In all cases, the expression of β2-microglobulin, a housekeeping gene, was not affected by the treatments. Figure 3 illustrates in a quantitative manner the levels of IL-1β, IL-6, and TNF-α mRNA expressed as percentage of β2-microglobulin mRNA. A one-way ANOVA revealed a significant effect of the treatments on the expression of the different cytokines under study: IL-1β (P< 0.05), IL-6 (P < 0.01), and TNF-α (P< 0.01). When administrated alone, LPS significantly induced IL-1β (P < 0.05), IL-6 (P < 0.05), and TNF-α (P < 0.01) mRNA expression in the hypothalamus of mice compared with saline injection. IL-1ra injected intracerebroventricularly abrogated LPS-induced IL-1β, IL-6 mRNA, and TNF-α mRNA expression in the hypothalamus, as evidenced by the comparison of the intraperitoneal LPS-intracerebroventricular saline and intraperitoneal LPS-intracerebroventricular IL-1ra groups [IL-1β (P < 0.05), IL-6 (P < 0.01), and TNF-α (P < 0.01)].
To ensure that the effect of IL-1ra was specific to the brain, IL-1β levels were measured in the plasma of IL-1ra-treated and nontreated mice. IL-1β levels were very low in the plasma of saline-treated mice (∼8 pg/ml). LPS treatment dramatically increased plasma levels of IL-1β (saline ip-saline icv vs. LPS ip-saline icv,P < 0.01; saline ip-IL-1ra icv vs. LPS ip-IL-1ra icv,P < 0.001), and this increase was not significantly altered by intracerebral administration of IL-1ra (Fig.4).
The findings of the present study show that blockade of brain IL-1 receptor activation by intracerebroventricular IL-1ra attenuates the depressing effects of systemic LPS on food intake and abrogates LPS-induced cytokine expression in the hypothalamus.
IL-1ra is a specific antagonist of IL-1 receptors that has no intrinsic activity. It has already been shown to abrogate all known effects of IL-1 on its target cells, and this applies to brain cells as well (2,13). Administration of IL-1ra in the lateral ventricle of the brain blocks the anorexic effects of IL-1β administered via the same route (19, 27). However, when IL-1β is injected at the periphery, rather than centrally, its depressing effects on food intake are only partially abrogated by central IL-1ra, suggesting that the effects of IL-1β on food intake are mediated at the periphery and in the brain (20). The present findings that intracerebroventricular IL-1ra only partially blocked the depressing effects of intraperitoneal LPS on food intake are in accordance with this interpretation. An alternative hypothesis is that intracerebroventricular IL-1ra did not fully abrogate the development of the central cytokine response to systemic LPS by leaving intact, for example, the induction of brain interferons, which have been shown to be involved in LPS-induced anorexia (for a recent review see Ref. 16). This possibility clearly deserves further investigation.
The observation that systemic LPS induced the expression of brain IL-1β at the mRNA level is in line with previous findings (11, 17,22). In addition to inducing the expression of brain IL-1β mRNA, systemic LPS has been shown to induce the expression of brain IL-1ra mRNA (11). This induction is delayed compared with that of IL-1β (unpublished data), which is consistent with the concept that endogenous IL-1ra acts as a modulator of the effects of IL-1β on its target cells. However, changes at the mRNA level do not necessarily recapitulate changes at the protein level. In apparent accordance with this possible discrepancy, contradictory results concerning brain IL-1β synthesis at the protein level have been reported. IL-1β immunoreactivity was detected in microglial cells and meningeal and perivascular macrophages in the brain of LPS-treated rats and was accompanied by the release of bioactive IL-1β in the extracellular fluid (29, 32). However, no IL-1-like activity in the cerebrospinal fluid of LPS-treated rats was detected by other authors (4, 31). The sensitivity of the technique used for measuring IL-1β is likely to be critical, since the levels of IL-1β that occur in the brain in response to peripheral LPS treatment are very low.
In the present study we used semiquantitative RT-PCR to accurately determine relative amounts of transcripts. The validity of this technique was first checked to confirm the lack of a plateau amplification of the internal standard (pMus3) and cDNA and to ensure that β2-microglobulin mRNA levels remained stable in hypothalamus of mice injected systematically with LPS. Because this was the case, the results were expressed as the ratio of cytokine amplification to β2-microglobulin × 100. As previously demonstrated, intraperitoneal administration of LPS induced expression of cytokine mRNA in the hypothalamus of mice 1 h after treatment (22). The basal expression of proinflammatory cytokines in the hypothalamus was relatively high in the present study compared with our observations in previous studies (22, 23). This difference could be due to a propagation of the cytokine signal induced by implantation of the cannula into the brain (18, 35). However, this possibility still needs to be checked by a direct comparison of cytokine expression in the brain of implanted and nonimplanted mice. Despite this increase in basal levels of cytokines, exogenous injection of IL-1ra in the lateral ventricle of the brain was able to fully block the LPS-induced expression of IL-1β, TNF-α, and IL-6 mRNAs. Contradictory data have been reported concerning the effect of intracerebroventricular IL-1ra on LPS-induced IL-6 protein (1, 24). In the rat, Luheshi et al. (24) could not detect any effect of intracerebroventricularly injected IL-1ra on LPS-induced IL-6 in the cerebrospinal fluid, whereas it was blocked in the cat (1). In both studies, IL-6 was measured by ELISA. Inasmuch as more cerebrospinal fluid can be collected from the brain of a cat than from the brain of a rat, it might be easier to pick up significant differences due to IL-1ra in the cat than in the rat.
In terms of mechanisms of action, IL-1ra binds to both types of IL-1 receptors: type I (IL-1RI), which is the active receptor, and type II (IL-1RII), which acts as a decoy target for IL-1 (2, 5). IL-1ra abrogates the effect of IL-1 on its receptors by preventing the formation of the complex IL-1RI–IL-1–IL-1R accessory protein, which is thought to be the transducing receptor complex (14, 34). Although these data have been obtained in peripheral immune and nonimmune cells, there is evidence that brain cells also express IL-1RI, IL-1RII, and IL-1RacP that do not differ from peripheral IL-1 receptor subtypes (10, 12, 17) and that the in vivo effects of IL-1 in the brain are mediated by IL-1RI (6, 30) and its accessory protein (unpublished observation), whereas the brain IL-1RII downregulates the effects of IL-1 (7).
The importance of hypothalamic IL-1 in the induction of proinflammatory cytokines differs from what occurs at the periphery, where TNF-α is usually considered to play a central role in inducing the release of IL-1 and IL-6 during bacterial infections (9). Accordingly, time-course studies confirmed that, in response to LPS, the release of IL-1 and IL-6 was maximal when the levels of TNF-α were declining (33). The present findings, therefore, reinforce the concept that the brain differs from the periphery in the way the cytokine network is activated (22).
In summary, the results obtained in the present study demonstrate that endogenous hypothalamic IL-1 plays a pivotal role in the organization of the neural components of the host response to infection.
Many data on the expression of proinflammatory cytokines in the brain in response to peripheral inflammatory stimuli have been collected during the last decade. However, because of the redundant properties of these cytokines, little was known about their relative importance in the organization of the brain response to systemic insults. The present findings are the first to demonstrate the key role of brain IL-1 in the organization of the hypothalamic cytokine network. These results are important, since they indicate that IL-1 is certainly a better target molecule for controlling inflammation in the brain than at the periphery.
We thank Viviane Tridon for skillful assistance with intracerebral implantation of the guide cannula in mice.
This work was supported by Institut National de la Santé et de la Recherche Médicale, Institut National de la Recherche Agronomique, Direction des Recherches Études et Techniques, and European Community BIOMED 2 Grant CT97-2492. S. Layé was supported by a Biotech training grant.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Address for reprint requests and other correspondence: S. Layé, Neuroendocrinologie Fonctionelle, Université Bordeaux 1, Ave. des Facultés, 33405 Talence, France (E-mail:).
- Copyright © 2000 the American Physiological Society