Interleukin-1 mediates endothelin-1-induced fever and prostaglandin production in the preoptic area of rats

Aline S. C. Fabricio, Giuseppe Tringali, Giacomo Pozzoli, Miriam C. Melo, Juliana A. Vercesi, Gloria E. P. Souza, Pierluigi Navarra


The intracerebroventricular injection of endothelin-1 (ET-1) induces fever and increases PG levels in the cerebrospinal fluid of rats. Likewise, the injection of IL-1 into the preoptic area (POA) of the rat hypothalamus causes both fever and increased PG production. In this study, we conducted in vivo and in vitro experiments in the rat to investigate 1) the hypothalamic region involved in ET-1-induced fever and PG biosynthesis and 2) whether hypothalamic IL-1 plays a role as a mediator of the above ET-1 activities. One hundred femtomoles of ET-1 increased body temperature when injected in the POA of conscious Wistar rats; this effect was significantly counteracted by the coinjection of 600 pmol IL-1 receptor antagonist (IL-1ra). In experiments on rat hypothalamic explants, 100 nM ET-1 caused a significant increase in PGE2 production and release from the whole hypothalamus and from the isolated POA, but not from the retrochiasmatic region, in 1-h incubations. Six nanomoles of IL-1ra or 10 nM of a cell-permeable interleukin-1 converting enzyme inhibitor completely counteracted the effect of ET-1 on PGE2 release from the POA. One hundred nanomoles ET-1 also caused a significant increase in IL-1β immunoreactivity released into the bath solution of hypothalamic explants after 1 h of incubation, although during such time ET-1 failed to modify the gene expression of IL-1β and other pyrogenic cytokines within the hypothalamus. In conclusion, our results show that ET-1 increases IL-1 production in the POA, and this effect appears to be correlated to ET-1-induced fever in vivo, as well as to PG production in vitro.

  • prostaglandin E2
  • interleukin-1 receptor antagonist
  • interleukin-1-converting enzyme inhibitor
  • posterior hypothalamus
  • hypothalamic explants

endothelins (ETs) are a family of peptides endowed with pleiotropic biological actions, including vasoconstrictor, immunological, endocrinological, and neurological effects; as a member of the ET peptide family (along with ET-2 and ET-3), ET-1 exerts its effects via the activation of two specific G protein-coupled receptors, named ETA and ETB receptors (30, 40). ET-1 is generated by a two-step process involving the intracellular cleavage of a specific pre-pro-ET-1 peptide to yield big ET-1, which is then converted to the mature 21-residue peptide ET-1 through the action of specific ET-converting enzymes (ECE), ECE-1 and ECE-2 (17, 74). The hypothalamus, including the preoptic area (POA), is endowed with all components necessary for a functionally active endothelin system. In fact, both mRNA expression and mature ET-1 and ET-3 proteins have been described in this area (24, 57); ETB (77) and, to a lesser extent, ETA receptors (75) are also present, and functional ECE activity has been reported (72). A high density of ETB receptors is also found in the subfornical organ and the organum vasculosum laminae terminalis (OVLT) (78), two circumventricular structures adjacent to the hypothalamus that lack a blood-brain barrier and are permeable to some circulating peptides (5, 64). Glial cells in both these structures also express ETs and functional ETA and ETB receptors (23). In addition, ET-1 and its precursor are detectable in the cerebrospinal fluid (CSF) under normal conditions (2) and appear to be substantially increased during brain diseases or injury (31, 51, 71).

Along with several proinflammatory cytokines, ET-1 is thought to be an endogenous pyrogen (19, 20). We have previously shown that intracerebroventricular ET-1 mimics intravenous LPS in causing fever (20), as well as in raising PG levels in the CSF of rats (21). Central injection of the selective ETB receptor antagonist BQ-788 counteracted fever induced either by ET-1 or LPS (20). The latter finding, along with the fact that ET-1 fever is observed after central (icv or intrahypothalamic) but not intravenous injection of the peptide (18), indicates that CNS-borne, but not peripheral, ET-1 is involved in febrile response.

Cyclooxygenase (COX)-2-derived PGs, produced in response to endogenous pyrogens, represent the final mediators of fever. PGs elicit its effects through the activation of PGE2 receptors expressed in neurons of the POA of anterior hypothalamus (41, 45). We have recently reported that ET-1-induced fever is accompanied by an increase in COX-2-derived PGs in the CSF of rats (21). The above features of ET-1 (i.e., induction of fever and increased PG production) are fairly similar to those previously reported with IL-1 (10, 32, 47, 69). It has also been reported that ETs stimulate proinflammatory cytokine production in different cell types, including macrophages, microglia, and astrocytes (14, 60). Moreover, we found that fever induced by intracerebroventricular ET-1 is completely blocked by intracerebroventricular IL-1 receptor antagonist (IL-1ra) (19). The above evidence suggests that IL-1 mediates certain central actions of ET-1. In the present study, we conducted in vivo and in vitro experiments in the rat to clarify 1) the hypothalamic region involved in ET-1-induced fever and PG biosynthesis and 2) whether hypothalamic IL-1 may play a role as a mediator of ET-1 activities, such as the induction of fever and the production of PGs.


Animals and surgery.

Experiments were conducted on male Wistar rats weighing 200–250 g, housed individually at 24 ± 1°C under a 12:12-h light-dark cycle (lights on at 0600) with free access to food and tap water. Approvals for experimental procedures and protocols were obtained from both the committee on ethical use of laboratory animals of the University of São Paulo, where the fever experiments were conducted, and the Italian Ministry of Health (licensed authorization to P. Navarra) for the in vitro studies. All protocols were performed in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals.

Under pentobarbital sodium (40 mg/kg ip) anesthesia, a permanent 22-gauge stainless steel guide cannula (0.7 mm outer diameter, 15 mm long) was stereotaxically implanted 2 mm above the POA at coordinates 0.6 mm lateral to the midline, 7.7 mm anterior to the interaural line, and 6.5 mm under the brain surface. In other rats, the cannula was placed 2 mm above the posterior hypothalamus (PHyp) at stereotaxic coordinates: 0.4 mm lateral to the midline, 4.6 mm anterior to the interaural line, and 6 mm under the brain surface. In both cases the incisor bar was lowered 3.0 mm below the horizontal zero (50). Cannulae were fixed to the skull with jeweller's screws embedded in dental acrylic cement. During the same surgical session, a battery-operated radiotelemetry transmitter (Data Science, St. Paul, MN) for telemetric monitoring of core body temperature (Tc) was aseptically implanted into the peritoneal cavity. Animals were then treated with oxytetracycline hydrochloride (400 mg/kg im) and allowed to recover in a room with temperature maintained at 25 ± 1°C for 1 wk before the experiments. After each experiment, the animals were anesthetized (as before), and the location of the cannula track was verified histologically. Animals showing cannula misplacement, blockage upon injection, or abnormal weight gain patterns during the postimplantation period were excluded from the study.

Core body temperature measurements.

Experimental measurements were conducted at the thermoneutral zone for rats in a temperature-controlled room (28 ± 1°C), following adaptation of the animals to this environment for at least 1 h. After this period, baseline temperature was determined before any injection; only animals displaying mean basal temperatures between 37.0 and 37.3°C were selected for the study. Tc was monitored continuously by implanted telemetry transmitters (Data Science). Data were acquired at 15-min intervals.

Intrahypothalamic injection.

In a first set of experiments, rats were microinjected into the POA or PHyp with 100 fmol ET-1 or saline. A dental injection needle (Mizzy, 200 mm OD) connected to a polyethylene (PE)-10 tube was used. The needle was protruded 2 mm beyond the cannula tip and a 200-nl volume was injected slowly (over 1 min) with a 5-μl Hamilton syringe coupled to a microinfusion pump (model 53100, Stoelting, Wood Dale, IL). After injection, the needle remained in place for one minute before it was withdrawn to prevent backflow of the injection fluid through the cannula.

In another series of experiments, rats were microinjected into the POA with 200 nl of a solution containing 100 fmol ET-1 or 600 pmol IL-1ra or both substances at the same concentrations as above. Control animals received sterile saline (200 nl). The dose of IL-1ra used in this study was based on the previous finding that a 1,000- to 10,000-fold excess of IL-1ra is required to block the IL-1-induced fever (49). Animals were always microinjected between 10:00 and 11:00 AM to minimize possible diurnal variability. Afterward, Tc was measured for 6 h.

Dissection of hypothalamic explants and incubation experiments.

Animals were killed by decapitation between 09:00 and 10:00 AM, and their brains rapidly removed. The whole hypothalamus was dissected from the brain with the following limits: the anterior border of the optic chiasm, the anterior border of the mammillary bodies, and the lateral hypothalamic sulci, with a depth of 2 mm. In certain experiments, the whole hypothalamus was divided into the POA (frontal cut 2 mm rostrally to the optic chiasm) and the remaining posterior part of hypothalamic explant [retrochiasmatic hypothalamus, (RCH)]. The latter was further divided longitudinally into two halves. The total dissection time was <2 min from decapitation (48).

The explants were incubated in a 24-well plate at 37°C in a humidified atmosphere consisting of 5% CO2-95% O2. The incubation volume was 350 μl/well for whole hypothalami or 250 μl/well for POA. The incubation medium was MEM with Earle's balanced salts, supplemented with 0.2% human serum albumin, 2 mmol glutamine/l, 60 μg/ml ascorbic acid and 100 IU/ml aprotinin, pH 7.4. Experiments were conducted on one explant per well, unless otherwise specified. After a 60-min preincubation period (during which the medium was changed every 20 min), the medium was aspirated and replaced with fresh medium alone (control), or medium containing the test substances. Hypothalamic tissues remain viable and functional during the timeframe of the experiments (1 h), as assessed by the lactate dehydrogenase assay for cell toxicity (53).

At the end of experiments, incubation media were collected and centrifuged, and the supernatant was stored at −35°C or −80°C until assay for PGE2 or IL-1β immunoreactivity, respectively. For RNA analysis, hypothalami were embedded in 2 ml of RNAlater solution (Ambion, Austin, TX) and kept at −20°C until RNA extraction.


PGE2 was measured by RIA, as described in detail elsewhere (69). The detection limit of the assay was 2 pg/tube. The intra- and interassay variability coefficients were 5 and 10%, respectively. The amounts of released PGE2 were expressed as picogram per milligram of wet tissue.

Supernatants of the incubation medium (50 μl) were assayed for the presence of mature IL-1β using a commercially available rat IL-1β ELISA kit (Pierce Biotechnology, Endogen, Rockford, IL), following the manufacturer's instructions. This ELISA kit utilizes an anti-rat IL-1β antibody that can recognize both recombinant and natural rat IL-1β. This antibody does not cross-react with rat interferon-γ (IFN-γ), IL-1α, IL-2, IL-4, IL-6, IL-10, monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein-1α (MIP-1α), or regulated on activation normal T-expressed and secreted (RANTES), or human, mouse or pig IL-1β. Detection limit to the assay was 12 pg/ml. Results were expressed as picograms IL-1β per milliliter.

RNA isolation and RNase protection assay.

Total RNA was extracted from homogenized tissue by the guanidine thiocyanate lysis method (7). The yield of RNA was ranging between 45 and 55 μg/ 100 mg of wet tissue.

To measure mRNA levels of a number of cytokine genes, the RiboQuant multiprobe template set rCK-1 (Pharmingen, San Diego, CA) containing cDNA templates for rat IL-1α, IL-1β, TNF-β, IL-3, IL-4, IL-5, IL-6, IL-10, TNF-α, IL-2, and IFN-γ, as well as the housekeeping genes L32 and glyceraldehyde-3-phosphate dehydrogenase, was used. Nucleotide antisense riboprobes for the above mentioned genes were synthesized using T7 RNA polymerase in the presence of [α-32P]UTP (800 Ci/mmol). [32P]RNA probes were extracted by phenol-chloroform and precipitated with ethanol.

RNase protection analyses were performed (52) by hybridizing 25 μg total RNA per sample in 24 μl deionized formamide plus 6 μl hybridization buffer containing 3 × 105 cpm of each riboprobe. A control sample of yeast tRNA was included in each assay as a negative control. After heating at 80°C, the samples were hybridized at 56°C for 15–18 h. Samples were then digested with RNase (40 μg/ml RNase A and 350 U/ml RNase T1) at room temperature for 60 min, followed by proteinase K treatment, phenol-chloroform extraction, and ethanol and ammonium acetate precipitation. Double-stranded RNase-protected fragments were resolved on a 5% polyacrilamide-8 M urea gel. After drying, gel was visualized by autoradiography. Quantitative analysis was performed by using the ImageMaster VDS and the Imagesystem software package (Amersham-Pharmacia Biotech, San Francisco, CA). The intensity of protected cytokine fragments was normalized to the intensity of the protected L32 fragment of the same sample, and results were reported as corrected arbitrary units.


ET-1 (stocked in 0.1% acetic acid sterile solution), LPS (Escherichia coli 026:B6; dissolved in sterile water), and pentobarbital sodium (solved in sterile saline) were obtained from Sigma (St. Louis, MO). Oxytetracycline hydrochloride (Terramicina, solved in sterile saline), was obtained from Pfizer Laboratories (São Paulo, Brazil). Human recombinant IL-1ra (stocked in saline for in vivo studies, or 0.05M PBS plus 0.1% human serum albumin for in vitro studies) was purchased from Bachem AG (Bubendorf, Switzerland). Interleukin-1 converting enzyme (ICE) inhibitor I, cell-permeable (stocked in DMSO) was from Calbiochem (San Diego, CA). All working solutions for in vitro experiments were prepared by appropriate dilutions in medium immediately before incubation.

Statistical analysis.

All variations in Tc were expressed as changes from the mean basal value (i.e., ΔT, in °C), and mean baseline temperatures were not statistically different among the groups in Fig. 1. Data of fever experiments are presented as means ± 1 SE of two different experiments and first analyzed by two-way ANOVA for repeated measures. When significant main effects or interactions were revealed, data for each time point were then analyzed by one-way ANOVA, and significance of differences between treatment groups was determined by the Newman-Keuls test.

Fig. 1.

ET-1 induces fever when microinjected into the preoptic area (POA) but not in the posterior hypothalamus; this effect is counteracted by IL-1 receptor antagonist. A: rats were microinjected in the POA with 200 nl of a solution containing 100 fmol ET-1 alone or 100 fmol ET-1 plus 600 pmol IL-1ra. Control animals received sterile saline (SAL) or 600 pmol IL-1ra alone. Basal temperatures (mean ± SE; °C) were the following: ○, 37.3 ± 0.17 (n = 4); •, 37.2 ± 0.03 (n = 5); □, 37.06 ± 0.051 (n = 10); ▪, 37.18 ± 0.08 (n = 8). By two-way ANOVA, significant differences were observed for main effects of treatment (P = 0.0316), time (P < 0.0001), and treatment-time interaction (P < 0.0001). *P < 0.05, #P < 0.01 compared with ET-1-treated group (ANOVA followed by Newman-Keuls test). B: ET-1 (100 fmol) or SAL (200 nl) were microinjected into the posterior hypothalamus (PHyp). Basal temperatures (means ± SE; °C) were the following: ▵, 37.2 ± 0.17 (n = 4); ▴, 37.13 ± 0.05 (n = 5). No significant differences were observed between the groups. Values represent the means ± SE of variation (ΔT) in body temperature. C: representative photomicrograph of coronal sections illustrating the site of ET-1 microinjection into the POA (left) and PHyp (right). Dorsal is at the top. 3V, third ventricle; LV, lateral ventricle; ac, anterior commissure; HIP, hippocampus; sox, supraoptic decussation.

Data from in vitro experiments were expressed as the means ± 1 SE of (n) replicates per group, and analyzed by one-way ANOVA followed by post hoc Newman-Keuls test for comparison between group means, or Student's t-test when appropriate.

All data were analyzed using a Prism computer software (GraphPad, San Diego, CA). Differences were considered significant when P < 0.05.


Fever induced by ET-1 microinjection in the POA is blunted by IL-1ra.

As previously reported by others (1, 8, 62), the injection of vehicle into the POA of rats evoked by itself a rise in Tc, which reached a maximum value of 0.7 ± 0.25°C after 2.5 h. Microinjection of ET-1 (100 fmol) into the POA caused a slowly developing and long-lasting increase in Tc, which was significantly higher compared with that induced by saline. ET-1-induced fever was significantly reduced when the rats were coinjected with IL-1ra (600 pmol). No statistical difference was observed between the group receiving ET-1 plus IL-1ra and the group receiving saline. IL-1ra given alone showed no differences with respect to normal saline (Fig. 1A). Injections of ET-1 or sterile saline into the PHyp did not modify Tc (Fig. 1B).

Light microscopic examination of serial coronal hypothalamic sections showed the site of ET-1 microinjection ventral to the anterior commissure and dorsal to the optic chiasm (Fig. 1C, left); this locus encompasses the classical thermosensitive and pyrogen-reactive regions of the diencephalon. Figure 1C (right), shows the site of ET-1 microinjection within the posterior hypothalamus, which was not associated to changes in Tc.

Endothelin-1 increases PGE2 from whole rat hypothalamus and POA.

Figure 2A shows that ET-1 increased PGE2 production and release from the whole hypothalamus in 1-h experiments. Such increase reached statistical significance at 100-nM concentration. The above stimulatory effect of 100-nM ET-1 was also observed when using POA explants, whereas no effect was seen on RCH explants (Fig. 2B).

Fig. 2.

Effect of ET-1 on PGE2 from the rat hypothalamus. A: ET-1 significantly increases PGE2 production and release from the whole hypothalamus. Results are from three independent experiments. B: ET-1 stimulates PGE2 production and release from POA but not from RCH explants. Results are from two independent experiments. Data are expressed as pg PGE2/mg of wet tissue, the means ± 1 SE of 7–9 replicates per group. *P < 0.05 and **P < 0.01 vs. control.

Endothelin-induced PGE2 release from the whole hypothalamus and POA is blocked by IL-1 receptor antagonist or ICE inhibitor.

To investigate whether ET-1 induces PG release via IL-1, 100 nM ET-1 was given alone or in the presence of human recombinant IL-1ra in the range of 0.06–6 nM. IL-1ra at the higher concentration completely prevented the increase in PGE2 release induced by ET-1 in the whole hypothalamus (Fig. 3A). Likewise, 6 nM IL-1ra fully antagonized the increase in PGE2 release caused by ET-1 in POA explants (Fig. 3B).

Fig. 3.

Interleukin-1 receptor antagonist or an interleukin-1 converting enzyme (ICE) inhibitor counteract ET-1-induced PGE2 release from rat hypothalamic explants. Interleukin-1 receptor antagonist (IL-1ra, 6 nM) significantly counteracts the increase in PGE2 release induced by 100 nM ET-1 from whole hypothalamus (A) and POA (B) explants. C: 10 nM cell-permeable ICE inhibitor displayed a similar inhibitory effect on ET-1-induced PGE2 release from the POA. Results are from a minimum of two independent experiments. Data are expressed as means ± SE in pg PGE2/mg of wet tissue of 5–9 replicates per group. *P < 0.05 vs. control; °, P < 0.05 and °°°P < 0.001 vs. ET-1 given alone.

A significant inhibition of ET-1 effects on PGE2 production was also achieved via the blockade of ICE, also referred to as caspase-1, by a specific cell-permeable peptide inhibitor (Fig. 3C). IL-1ra or ICE inhibitor given alone had no effect on PGE2 release.

Endothelin-1 or LPS stimulates the release of immunoreactive IL-1β from the rat hypothalamus.

Figure 4A shows that ET-1 caused an increase in immunoreactive IL-1β levels in the bath solutions of whole hypothalamic explants, with statistical significance at 100 nM, that is, the same concentration also stimulating PGE2 release in this paradigm. The effect of ET-1 was similar to that obtained by 10 μg/ml LPS (Fig. 4B) and could be attributed to an action at the level of POA, as the peptide was able to stimulate IL-1β release from isolated POAs but not from RCH explants (Fig. 4C).

Fig. 4.

ET-1 stimulates the release of immunoreactive IL-1β from rat hypothalamic explants. One hundred nanomoles ET-1 (A and B) and 10 μg/ml LPS (B) significantly increase IL-1β release from the whole hypothalamus. ET-1 (C) stimulates IL-1β release from the POA but not from retrochiasmatic hypothalamus (RCH) (2 explants × well). Results are from two independent experiments. Data are expressed as pg IL-1β/ml of incubation medium, the means ± 1 SE of 5–11 replicates per group. *P < 0.05 vs. control.

Lack of effect of ET-1 on cytokine gene expression in the POA or RCH.

Having shown that ET-1 increases the levels of mature IL-1β in the hypothalamus and that the blockade of either IL-1 receptors by IL-1ra or ICE by a caspase-1 antagonist blunt the effects of ET-1 on PG production, we next analyzed the gene expression of a number of cytokines, including IL-1α and IL-1β, to ascertain whether the effect of ET-1 on IL-1 production occurs at the mRNA production level. Cytokine mRNA levels were assessed in the POA or RCH explants incubated for 1 h in medium containing vehicle (control), 100 nM ET-1 or 10 μg/ml LPS as a positive control. Only bands for IL-1α, IL-1-β and IL-6 (data not shown) were detected in controls and treated explants. ET-1 did not affect the pattern of cytokine gene expression with respect to the control in both studied areas, whereas LPS induced increases in mRNA levels for IL-1α and IL-1β in the RCH (Fig. 5).

Fig. 5.

Absence of effect of ET-1 on IL-1α and IL-1β mRNA levels in the rat hypothalamus. POA or RCH explants were incubated for 1 h with medium + vehicle (control), or medium containing 100 nM ET-1 or 10 μg/ml LPS, and total RNA extracted. To quantify the effect of the stimuli, the intensity of the protected cytokines were normalized to the intensity of protected L32 housekeeping genes (see materials and methods). Data from three independent experiments are expressed as the means ± SE of 5–8 replicates per group. *P < 0.05 vs. control. A representative experiment is shown (top). Lanes 1 to 3 represent replicates of control treatment; Lanes 4 to 6 represent replicates of explants treated with ET-1; lanes 7 and 8 represent replicates of LPS-treated explants.


The present results from in vivo studies bring new insight to the notion of ET-1 as an endogenous pyrogen, pointing out two important aspects: 1) the hypothalamic site of ET-1 effect, and 2) the role of IL-1 as a mediator of ET-1 action in the hypothalamus. In vitro findings show that ET-1 acts mostly at the level of POA to increase the production and/or release of PGE2 and IL-1β, each of which is closely correlated to pyrogenesis in vivo. The effect of ET-1 on PGE2 production appears to be mediated by the upstream stimulation of IL-1 production. However, no induction of IL-1 gene expression both in the POA and RCH is observed after ET-1 treatment.

The preoptic anterior hypothalamic area as a putative pyrogenic zone for ET-1-induced fever.

We have previously shown that intracerebroventricular ET-1 acts on central ETB, but not ETA, receptors to raise the Tc of rats, thus mimicking the effects of intravenous LPS; the selective ETB receptor antagonist BQ-788, but not the selective ETA receptor antagonist BQ-123, markedly reduces LPS-induced fever (20). The microinjection of 100 fmol ET-1 into the POA elicits a sustained fever in the rat (18); this effect fully reproduces the febrile response that follows the intracerebroventricular administration of a 10-fold higher dose of the peptide (18, 19, 20, 21). Here, we have also compared the effects of ET-1 microinjections into the POA and into the PHyp, one of RCH nuclei, because neurons of both regions are sensitive to Tc and integrate central and peripheral thermal information (37, 44). ET-1 induced fever after injection into the POA, but not into the PHyp, showing a clear site-specific effect.

Here, we observed an increase in Tc after insertion of the injection needle into the POA, regardless of whether or not the vehicle was injected into the tissue (data not shown), and therefore, the increase in Tc reported in Fig. 1A cannot be attributed to the presence of pyrogens in the vehicle. Microinjection procedures into the POA have been shown to increase Tc in rats (1, 8, 62), cats (11), guinea pigs (56), and rabbits (9). Different authors have related such increases to increased local production of inflammatory mediators by damaged tissue. Moreover, we attempted to minimize the expected rise in Tc associated to the injection procedure by injecting a small volume (200 nl) over a relatively long period of time (1 min).

Hypothalamic IL-1 as a putative mediator of ET-1-induced fever.

Fever induced by ET-1 injections into the POA could be prevented by IL-1ra. It is known that endogenous IL-1ra plays a pivotal role in mitigating the magnitude of IL-1-mediated inflammatory response (29). Exogenous IL-1ra reduces febrile responses elicited in rats by different pyrogenic stimuli including muramyl dipeptide and IL-1β (27), adenoviral vectors (6), polyinosinic:polycytidylic acid (22), TNF-α (61) and leptin (39). Relevant to the present findings is the fact that intrahypothalamic injections of IL-1ra mimicked the blockade of ETB receptors in reducing LPS-induced fever (5, 20). Indeed, evidence from our previous study (19) indicates that other agents, namely CRF and the yet incompletely characterized preformed pyrogenic factor (PFPF) converge on both the ET-1 and IL-1 pathways to exert a role in the mechanism of fever: the blockade of ETB and/or IL-1 receptors inhibits fever induced not only by ET-1 but also CRF and PFPF.

Direct actions of IL-1 on neurons within the POA have been demonstrated in a number of electrophysiological studies showing that the application of IL-1 into the POA decreases the firing of warm-sensitive neurons, whereas increasing the activity of cold-sensitive neurons (27, 70, 73). The effects of IL-1 on POA neurons are also sensitive to blockade by IL-1ra (73). The above evidence is consistent with the view that IL-1 shifts the set point of Tc to a higher level through a direct action on sensitive neurons, thereby producing fever.

ET-1 vs. LPS-induced hypothalamic IL-1β formation.

Here, we found that ET-1 causes a significant increase in IL-1β immunoreactivity released into the bath solution of hypothalamic explants after 1 h of incubation. ET-1-induced changes in IL-1β release were observed in POA but not in RCH explants. However, during such time, ET-1 failed to modify the gene expression of IL-1β and other pyrogenic cytokines in the hypothalamus; LPS, which is known to increase IL-1β mRNA levels in hypothalamus after systemic injection (35) or to stimulate IL-1 protein release after incubation with hypothalamic explants (46), was used as a positive control. LPS significantly increased IL-1α and IL-1β mRNA levels in the explants of RCH, but not in the POA. This finding is apparently in conflict with other reports of increased IL-1 message in brain areas, including the OVLT, after LPS injection (55, 54). However, in these papers, the endotoxin was administered by the peripheral route, and therefore the two paradigms are not directly comparable. The sensitivity of OVLT to circulating LPS is not unexpected, because both CD14 and Toll-like receptor 4 (TLR4) (34) are constitutively expressed in this region, which can be reached by the bloodstream. However, in a study showing cytokine expression in the OVLT in response to peripheral subseptic dose of LPS (54), rats were killed 2 h after the injection of the endotoxin. This time is sufficient to the transcription of cytokines, including IL-1β itself. In fact, IL-1β can be detected in the bloodstream within 1 h after LPS injection (25). Because TLR4 seems to be the subtype essentially required to the most (if not all) of responses to LPS in vivo (63), at the conditions of the present study, that is, after an incubation period of 1 h, the effect of LPS can be attributable to a direct action on TLR4 receptors. On the other hand, 2 h after LPS injection, other mediators, such as the above cited peripherally formed IL-1β, could amplify its effect in increasing IL-1β mRNA expression inside the brain. The above considerations lead us to conclude that the absence of an effect of LPS in the POA explants in our model is not counter to the findings obtained using peripheral injections of endotoxin in vivo.

Lack of effect of ET-1 on IL-1 gene expression and putative posttranscriptional mechanisms.

The present data would suggest that the effect of ET-1 on the production and release of IL-1β occur at the posttranscriptional level. Putative mechanisms might involve an increase in IL-1 mRNA stability and/or the increased cleavage from a constitutive pool of IL-1 precursor protein. The latter event is usually associated with increased activity of ICE, also referred to as caspase-1, the enzyme responsible for the formation of mature IL-1 (65). IL-1 bioactivity and mRNA expression markedly increase in the rat hypothalamus within 1 h after the beginning of physico-emotional stress (42, 59). Although the increase in IL-1 gene expression occurring under these conditions was rapid, the increase in IL-1 bioactivity was even faster, being detectable as early as 5 min from the initiation of restraint stress. This finding led Shintani et al. (59) to postulate the existence of a preformed IL-1 pool coming into play during acute adaptive responses before newly synthesized IL-1 is available. Interestingly, the release of immunoreative IL-1β from rat hypothalamic explants is also stimulated by CRF (68), which induces a fever substantially dependent on ETs (19).

Alternative posttranscriptional mechanisms might involve the reported ability of ETs to increase NO production in the anterior hypothalamus via the neuronal nitric oxide synthase pathway (28). This phenomenon appears to be mediated by the activation of ETB receptors, and, in turn, produces downstream biochemical events such as increased cytosol cGMP levels, activation of GABA-A receptors, and decreased norepinephrine release in the anterior hypothalamus of rats (12, 28, 43). Interestingly enough, Di Nunzio et al. (13) also found diverging neuromodulatory effects of ET-1 when tested in different hypothalamic areas. While these ET-1 biochemical pathways have been well established, the relationships with the findings presented here remain to be clarified.

What is the cellular source of such constitutive IL-1 (and/or IL-1 precursor protein)? Although the large majority of available data indicates that IL-1 production is usually associated to cells of the immune-inflammatory lineage, some evidence concerning the hypothalamus also supports the hypothesis of a neuronal origin for IL-1. Breder et al. (3) first reported IL-1β-like immunoreactivity localized in neuronal cell bodies and fibers in different hypothalamic areas of human brains, including the POA. Similar findings were also obtained in rat brains (36). We have previously observed that immunoreactive IL-1β is released from rat hypothalamic explants following a typical neuronal pattern, that is, increased secretion by depolarizing stimuli, Ca2+- and Na+-dependent release and others (67). In this paradigm, stimulated IL-1β was also significantly reduced by the cell-permeable ICE inhibitor used in the present study (66). More recently, ICE has been localized in neurons of the rat brain; the presence of ICE in the nerve fibers of the POA and other hypothalamic areas suggests a functional role for ICE products in the control of fever and sleep-wake rhythm (38). In fact, the inhibition of ICE in vivo in the rat was found to protect against LPS-induced fever and inflammatory response (16).

Hypothalamic IL-1 as a putative mediator of ET-1-induced PGE2 production.

The in vitro paradigm was also very useful to unravel the complex interplay occurring among ET-1, IL-1, and PGs within the anterior hypothalamus. ET-1 induced a significant increase in PGE2 release from whole hypothalamus, and such an effect was observed at the same concentration that significantly stimulates IL-1β release. We also found that ET-1 induces the production of PGE2 from POA, but not from RCH explants (which include the PHyp region). This finding correlates well with our previous in vivo observation that ET-1-induced fever is accompanied by increased PG levels in the CSF of rats (21) and is in keeping with the notion that the POA is highly sensitive to PGE2 in terms of the febrile response (58) and presents a high density of PGE2 receptors (41). The increase in PG levels occurring within the CNS after stimulation by a number of different agents, including ET-1 and ET-3, is commonly associated with COX-2 isoform activity (4, 21, 33, 76). We have recently observed that low doses of celecoxib, a selective inhibitor of COX-2, did not significantly influence fever produced by ET-1, whereas it blunted the rise in CSF PG levels induced by the peptide (21). Higher doses of celecoxib (5 and 10 mg/kg) blocked both responses. Thus ET-1 appears to be able to induce fever even through PG- (and COX-2-) independent mechanism(s). One such putative mechanism might involve the above-mentioned ability of IL-1 to directly interact with thermosensitive neurons in the POA.

Here, we found that the increase in PGE2 production and release from hypothalamic explants induced by ET-1 is strictly dependent on IL-1, as it is abolished by IL-1ra. Indeed, PGs have been implicated as mediators and/or modulators in various biological activities of IL-1 (15). It has also been shown that the intravenous injection of IL-1β induces a significant increase in PGE2 levels in the POA, which is more rapid and larger than that observed in any other brain areas, including other hypothalamic nuclei (32). We also observed that the cell-permeable ICE inhibitor was able to block ET-1-induced PGE2 production in the POA. This finding, together with the increase in IL-1β production from the POA in response to ET-1, further reinforces the concept that ET-1 stimulates IL-1 production in this area via a posttranscriptional mechanism involving increased ICE activity.


This work was supported by grants from Fondi di Ateneo 2004–2005 (Italy) and by Fundação de Amparo a Pesquisa do Estado de São Paulo and the Conselho Nacional de Desenvolvimento Científico e Tecnológico (Brazil). ASC Fabricio was recipient of a fellowship from ‘Istituto G. Toniolo di Studi Superiori.'


The authors thank Rubens Fernado de Melo for technical assistance in histological processing.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract