| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Injury Research Group, University of Manchester (A.V.T., S.P., A.R.K., R.A.L.), Manchester M13 9PT, United Kingdom; and Hope Hospital (S.J.H.), Salford M6 8HD, United Kingdom
Address all correspondence and requests for reprints to: Dr. S. J. Hopkins, Injury Research Group, Clinical Sciences Building, Hope Hospital, Salford M6 8HD, United Kingdom. E-mail: shopkins{at}man.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
or IL-1ß, together with a prolonged (>18-h) activation of the HPAA. IL-6 plays a causal role in the TPS-induced elevation in HPAA activity, because the sustained (8–18 h) increases in 1) plasma corticosterone, 2) plasma ACTH, and 3) induction of c-Fos in the hypothalamic paraventricular nucleus are all markedly blunted in IL-6-deficient (IL-6-/-) mice. Peripheral administration of a neutralizing IL-6 antiserum inhibited the plasma corticosterone response of normal (C57BL/6) mice to hind limb inflammation to an extent similar to that seen in IL-6-/- mice, suggesting that the IL-6 responsible for the increased HPAA activity is produced, or acts, on the blood side of the blood-brain barrier. We also show that IL-6 in the circulation is induced almost exclusively at the local inflammatory site, where IL-1ß is produced. Induction of IL-6 and activation of the HPAA are dependent upon prior activation of an IL-1 type I receptor, as both are inhibited in type I IL-1 receptor-deficient mice. Furthermore, hind limb inflammation induced cyclooxygenase-2 protein expression around the cerebrovasculature of normal (IL-6+/+), but not IL-6-/-, mice. Based on these data, we propose that IL-6 is produced at the local inflammatory site under the control of IL-1ß and is the circulating afferent signal that is in part responsible for elevated HPAA activity, possibly acting via eicosanoid production within the cerebrovasculature. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Tissue trauma, inflammation, and infection represent particular types of stress, which, in contrast to purely psychogenic stress, impact on both peripheral tissues and the CNS. In these situations there is evidence of integration of neural and immune responses, which may occur at several levels, but appears principally to be integrated at the level of the hypothalamus (7). Mediation of signals associated with the immune system appears importantly to involve the production and action of several cytokines associated with the inflammatory response, particularly IL-1 and IL-6. These cytokines are induced during inflammatory responses and have been shown to have potent actions at several levels in the HPAA (8). Pharmacological effects of cytokines at the level of the hypothalamus, pituitary, and adrenal have been observed, but it is unclear which may be physiologically relevant and how they interface with pathways regulating the HPAA.
Because IL-6 is the only one of these cytokines consistently observed to be increased in plasma during inflammation, we sought to test the hypothesis that IL-6 is an important afferent signal in this situation. We investigated the role of IL-6 in a model of turpentine (TPS)-induced tissue inflammation, which is of particular value in corresponding more closely to tissue injury and inflammation, and avoids complex questions related to where lipopolysaccharide (LPS) may be acting. We found that IL-6, induced by IL-1 at the site of inflammation, is the key cytokine involved in systemic mediation of the signal from the tissues to the HPAA and is required for induction of cyclooxygenase-2 (COX-2) in the brain microvasculature.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Local inflammation and restraint stress
Sterile local inflammation was induced in mice by the im injection of 100 µl TPS in mice (10) and 600 µl TPS in rats (11). Control animals received a similar injection of 0.9% sterile saline.
Mice were subjected to restraint stress by placing them in a 50-ml conical tube with a small (5-mm) segment of the pointed end removed (to provide an air source). When placed inside the tube, mice have sufficient room to move, but not to turn. Control animals were left undisturbed in their home cage until sacrifice.
Plasma corticosterone and ACTH measurements
Plasma corticosterone measurements were made using a commercially available RIA (ICN Biomedicals, Inc., Oxon, UK). The limit of assay detection was 6.25 ng/ml. Plasma ACTH concentrations were measured using a commercial immunoradiometric assay (Allegro, Nichols Institute Diagnostics, San Juan Capistrano, CA) as described previously (5). Within- and between-assay coefficients of variation at low (50 pg/ml) and high (350 pg/ml) concentrations were less than 10%, and all plasma ACTH concentrations from a single experiment were measured in the same assay. The detection limit of this assay was 17 pg/ml. We found that plasma obtained from decapitated animals assayed for ACTH produced inconsistent and unreliable data, and plasma samples did not dilute parallel to the standard curve. Consequently, plasma ACTH concentrations were measured on plasma derived from a blood sample obtained by cardiac puncture of mice anesthetized with a high dose (100 µl of 30 mg/ml) of sodium pentobarbital per mouse. Assay of plasma samples so derived produced consistent ACTH values, and these samples diluted in parallel to the standard curve. The time taken from first handling the mouse to completion of sample collection varied between 1.5–3.5 min.
Plasma and tissue cytokine measurements
Cytokines in the muscle of injected limbs were measured after homogenization of the tissue. On thawing, 50 µl TPS were added to tissue from animals that had been injected with saline to approximately equalize the TPS concentration to that of animals injected with TPS and provide a similar assay matrix. All muscles were homogenized on ice for 90 sec in 1.2 ml PBS using an Ultra Turrax T25 homogenizer with an IKA 10-mm probe (Janke & Kunkel, Staufen, Germany). The homogenate was centrifuged at 8000 x g for 3 min, and the supernatant was frozen at -70 C, before assay. Rat IL-6 was measured, relative to the First International Standard for human IL-6 [89/548, National Institute for Biological Standards and Control (NIBSC), South Mimms, UK], using a bioassay as previously described (12). Mouse IL-6 was quantified using rat antimurine monoclonal antibodies to capture (BD PharMingen, San Diego, CA; clone MP5-20F3) and detect (BD PharMingen, clone MP5 32C11), followed by development with horseradish peroxidase-conjugated streptavidin (Vector Laboratories, Inc., Burlingame, CA) and orthophenylene diamine (Sigma-Aldrich Corp., Poole, UK). The standard was recombinant murine IL-6 (Immuno-Kontact, Frankfurt, Germany). Plasma samples were diluted between 1:4 and 1:10 in 100% fetal calf serum, and thawed muscle homogenate supernatants were diluted 1:10 in PBS containing 0.1% Tween and 0.25% BSA. Standards were diluted in the same matrix. Mouse IL-1 was similarly quantified, but using a hamster antimurine monoclonal antibody to capture (Biogenesis, Poole, UK; clone 2G7/43), a sheep antimurine IL-1 polyclonal serum to detect (gift from Dr. S. Poole, NIBSC), and development with horseradish peroxidase-conjugated antisheep IgG (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) and orthophenylene diamine. The standard was recMuIL-1 (93/672; NIBSC). Plasma samples, muscle homogenate supernatants and standards were diluted 1:2 to 1:10 in 100% fetal calf serum. TNF was quantified using mouse TNF DuoSet reagents (DY410, R\|[amp ]\|D Systems, Abingdon, UK). IL-1ß was quantified using a commercial kit (Quantikine M, MLB00, R\|[amp ]\|D Systems, Inc.). Dilutions were performed with the diluents provided. Muscle homogenate supernatants were assayed at dilutions of 1:2 or 1:10. The interassay coefficient of variation for the IL-6 assay was less than 12% within the assay range. Other cytokine comparisons were determined from measurements within a single assay, and all intraassay coefficients of variation were less than 10%. Within the range of values measured, recovery of exogenous recombinant cytokines was between 85% and 100%.
Assessment of limb edema
Limb edema was assessed by measuring the fluid content of the injected limb. After death the whole limb was severed at the hip joint and immediately weighed. The limbs were kept in an oven (100 C) for 2 d, by which time they had completely dried, as evidenced by the constancy of the weights of limbs. The dry weight of the limb was then measured, and the total loss in weight (fluid content) was calculated.
Immunohistochemistry
Animals were anesthetized with 300 µl 20% chloral hydrate (ip) and perfused transcardially with 0.9% saline (3 min, 20 sec, at 3 ml/min) and then 4% paraformaldehyde solution (6 min, 40 sec, at 3 ml/min). Brains were removed and postfixed overnight in a 10% sucrose/4% paraformaldehyde solution. Frozen brain sections (30 µm) were cut on a sliding microtome and stored at -20 C in 50% 0.1 M potassium PBS (KPBS), 30% ethylene glycol, and 20% glycerol as cryoprotectant. Before incubation with primary antibody, sections were washed several times in KPBS and incubated with 1% H2O2 to reduce background.
For c-Fos immunohistochemistry, sections were incubated for 24 h at 4 C with rabbit anti-c-Fos (Ab-5, Oncogene Research Products, Uniondale, PA), diluted 1:20,000 in KPBS with 2% normal goat serum. This was developed with a biotinylated goat antirabbit secondary antibody (1 µg/ml; Vector Laboratories, Inc., Peterborough, UK) and an avidin-biotin-peroxidase complex (Vector Laboratories, Inc.), in the presence of 3,3'-diaminobenzidine tetrachloride (Sigma-Aldrich Corp.), 0.003% H2O2, and 26.3 mg/ml nickel II sulfate. The brain slices were washed in acetate-imidazole and KPBS buffer, mounted onto poly-L-lysine-coated slides, vacuum-dried, and subsequently dehydrated, cleared, and coverslipped with DPX (Merck \|[amp ]\| Co., Lutterworth, UK). Sections were microscopically inspected for c-Fos immunoreactivity (black-stained nuclei). Preadsorbing the anti-c-Fos antiserum overnight at 4 C with 10 µg synthetic c-Fos peptide immunogen abolished specific staining. Anatomical localization was defined by comparison with an adjacent series of Nissl-stained sections. Positively stained cells in every second 30-µm section through the PVN were counted.
For COX-2 immunohistochemistry, sections were incubated for 24 h with a rabbit anti-COX-2 serum (catalog no. CAY-160106, Alexis Corp., Nottingham, UK) diluted 1:180 in KPBS with 2% normal goat serum. After several washes in KPBS, sections were developed and mounted as described for c-Fos staining. Single whole brain sections for each animal in the experimental group were microscopically inspected for COX-2 immunoreactivity (brown-stained cytoplasm) at the fourth section containing the PVN. Positive vessels were defined as those in which more than two cells were stained. Specific staining was abolished by preadsorbing the anti-COX-2 antibody overnight at 4 C with 50 µg synthetic peptide immunogen. Anatomical localization was determined by comparison with an adjacent series of Nissl-stained sections.
Occlusion of blood flow from the limb
Because of the relative technical difficulty involved in cannulating mice and limitations on blood volumes that may be taken, rats were used to evaluate the disappearance of IL-6 from plasma after occlusion of blood flow from an inflamed limb. Local inflammation was induced in the hind limb of male Sprague Dawley rats by im injection of 0.6 ml of the mineral oil TPS (Rowney, Berkshire, UK) into the left thigh muscle as described previously (13). Control animals were injected with a similar volume of sterile 0.9% saline. Approximately 5.75 h later rats were anesthetized with an iv dose of sagatal (60 mg/kg). The ventral tail artery was rapidly cannulated with PE5O tubing, such that the tip of the catheter was positioned at the junction of the tail artery and the aorta. A blood sample (0.2 ml) was obtained via the catheter, and an equal volume of sterile, heparinized saline was administered to replace the fluid lost. Tourniquets (rubber bands made taut by winding around a hollow aluminum tube) were applied to the inguinal region of either the inflamed limb (left) or the noninflamed limb (right) at 6 h after injection. Tourniquet application resulted in the limb turning blue within 30 sec, indicating that blood flow to (and therefore from) the inflammatory site had ceased. Blood samples were obtained 1, 3, 5 10, and 20 min after tourniquet application.
Immunoneutralization of IL-6
IL-6 was passively immunoneutralized by ip administration of 0.5 ml of a neutralizing antirat IL-6. Antirat IL-6 was raised in sheep using recombinant rat IL-6 as the immunogen and was provided by Dr. S. Poole (NIBSC). The cross-reaction of this antiserum was apparent with respect to its efficacy in immunoassays of both rat and mouse IL-6. Control animals received a similar injection of sheep preimmune serum, which was also provided by Dr. S. Poole. The timing of serum injections (10 h before sampling time) was determined from previous descriptions of the temporal profiles of antibody titers in plasma after ip administration of antisera to rats (14) and from successful immunoneutralization experiments in mice (10).
Data presentation and statistical analyses
The majority of data are presented as the mean ± SEM, and the numbers of subjects in each experimental group are indicated on either the figures themselves or in the figure legends. Statistical analyses of these data were performed using either paired or unpaired t test or a one-way ANOVA (followed by Tukey-Kramer multiple comparisons test) as appropriate. In all analyses a two-tailed probability of less than 5% (i.e. P < 0.05) was considered statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). The limb edema remained relatively well maintained at 36 h. The corticosterone response was also significant 1 h after the injection of TPS, peaking at 4 h and declining over the next 14 h (Fig. 1B). Increased plasma IL-6 was detected only after 4 h, peaking at 8–12 h and remaining raised until at least 18 h (Fig. 1C). Plasma IL-1ß was below the level of detection in these animals (<28 pg/ml), and although the assay for IL-1 detected a signal in control plasma, it did not increase significantly up to 36 h after injection of TPS (257 ± 71 pg/ml in controls at 0 h and a peak of 282 ± 117 at 8 h after TPS). The ACTH response was also rapid (Fig. 2A) and remained significantly elevated at 12 h (Fig. 2B). Activation at the level of the hypothalamus was similarly apparent at 1 and 12 h, as indicated by the increase in the number of c-Fos-positive cells in the PVN (Table 1). Interestingly, the number of c-Fos-positive cells was higher in the contralateral PVN, although this was only significant 12 h after injection of TPS.
|
|
|
). Similarly, there was no significant difference in corticosterone response 1 h after im injection of TPS (Fig. 3B; IL-6+/+, 797 ± 35 ng/ml; IL-6-/-, 719 ± 25 ng/ml; by one-way ANOVA followed by Tukey-Kramer multiple comparisons test). This demonstrates that IL-6-/- mice are capable of producing high plasma corticosterone concentrations (>700 ng/ml) and that the acute (neurally mediated) response to both noninflammatory stress and inflammation is independent of IL-6.
|
). Corticosterone responses of IL-6-deficient mice were inhibited by 36% at 8 h, 73% at 12 h, and 46% at 18 h. To examine whether a less severe inflammatory insult would be similarly IL-6 dependent, the 12 h point was repeated with a lower dose of TPS (5 µl, im). The plasma corticosterone response of IL-6-deficient mice was reduced to a similar extent (70%) as that observed with 100 µl TPS [256 ± 43 (n = 9) and 77 ± 15 (n = 8) ng/ml].
|
shows that mice without IL-6 did not exhibit a lesser degree of hind limb swelling compared with controls. The only difference seen was at 1 h in the IL-6-/- mice, where the deficient mice actually had a significantly greater response than IL-6+/+ mice.
|
), while again not affecting the inflammatory response (Table 3).
|
|
). In contrast, 12 h after induction of local inflammation in IL-6-/- mice, plasma ACTH concentrations (209 ± 23 pg/ml) did not differ significantly from saline-treated controls of either genotype, but were significantly (P < 0.05) lower than those in TPS-treated IL-6+/+ mice. Examination of c-Fos staining in the PVN revealed that the number of activated cells increased on both the contralateral and ipsilateral sides of both IL-6+/+ and IL-6-/- mice, although the staining was more marked on the contralateral side (Table 1
and Fig. 7B). However, in IL-6-/- mice the number of c-Fos-positive cells was reduced, although this was significant only on the ipsilateral side of the PVN (Fig. 8; P < 0.05).
|
|
|
or IL-1ß concentration in the plasma of mice after the injection of TPS. However, IL-1 is a potent inducer of IL-6, and it seemed likely that it would be important at the site of inflammation. Homogenization of muscle collected 4 h after injection of TPS into IL-6+/+ and IL-6-/- mice revealed that IL-1ß, IL-6, and TNF were increased, although there was no evidence of an increase in tissue IL-1 (Table 4). The increase in IL-1ß was actually significantly greater in IL-6-/- mice, although the TNF response was not significantly different. The importance of IL-1 in the overall response is shown by the fact that the corticosterone response of IL-1RI-/- mice is, like that of IL-6-/- mice, significantly blunted relative to the control value (Fig. 9), whereas local inflammation at 12 h is not affected (Table 5). Importantly, the injection of TPS into IL-1RI-/- mice produced no significant increase in plasma IL-6 at this time (saline, 67 ± 34 pg/ml; TPS, 143 ± 37 pg/ml) in contrast to the C57 x J129 controls [saline, 30 ± 0 pg/ml (30 pg/ml limit of detection); TPS, 1954 ± 150 pg/ml], supporting the concept that IL-6 is an important systemic intermediary between IL-1 induction and subsequent effects observed in the CNS and HPAA.
|
|
|
). In contrast, application of a tourniquet to the TPS-injected limb caused the plasma IL-6 concentration to fall rapidly, and by 10 min after tourniquet application plasma IL-6 fell by over 90% (half-life, 1.29 ± 0.13 min). This shows that the preponderance of IL-6 in the circulation after the injection of TPS is attributable to induction in the inflamed limb.
|
), whereas no staining was seen in controls (not shown). In contrast, few cells were stained in the brains of IL-6-/- mice injected with TPS (Fig. 11B). Although measurements were made in a single section of each brain, sections throughout the entire brain showed quantitatively the same result.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In contrast, models where inflammation is primarily restricted to peripheral tissues rely on local activation of neural afferent signals or local induction of mediators that must first gain access to the circulation. A feature of such models is that IL-1 and TNF are rarely reported in the circulation, whereas IL-6 is (11, 19), and this provides a rational explanation for why the response to TPS has shown a greater IL-6 dependency than that to LPS in the induction of acute phase proteins (20). Although we found IL-1ß and TNF to be induced locally after im injection of TPS, only IL-6 was detected in the circulation. Its relatively slow appearance, compared with the edema and HPAA responses, suggested that it would be likely only to affect later components of inflammation and HPAA. The relatively slow IL-6 response explains why corticosterone responses have previously been shown to be normal in IL-6-/- and IL-1-/- mice after the injection of TPS, as corticosterone responses were evaluated at 1.5–2 h thereafter (20, 21, 22) and is consistent with our finding that stress-induced corticosterone responses are similar in IL-6+/+ and IL-6-/- mice. However, there is at least one report demonstrating that anti-IL-6 antibody reduces ACTH production 2 and 4 h after the administration of LPS (23). Wang and Dunn (24) also showed that anti-IL-6 antibody reduced corticosterone and ACTH responses after the injection of LPS or IL-1ß. Possibly, the effect seen with LPS is dependent upon mouse strain or the amount of LPS used. For instance, Fattori et al. (20), who showed that LPS had little dependency on IL-6, used approximately 10-fold more LPS, of the same LPS serotype, than was used by Wang and Dunn (24).
The importance of IL-6 in the HPAA response has also been reported after infection with murine cytomegalovirus (MCMV) (25). As seen after systemic injection of LPS, Ruzek et al. (25) found that several cytokines were increased in the serum, including IL-6 and TNF. Serum IL-1 was also reported to be elevated, although the values appeared erratic and, as with our measurements, were unaccountably high in control animals (25). However, IL-6-/- mice showed significantly reduced HPAA responses 36 h after MCMV infection and 2 h after ip injection of the synthetic viral nucleic acid analog polyinosinic-polycytidylic acid as well as a significant, if modest, reduction 2 h after ip LPS (25). As with LPS, it is difficult to be certain where the MCMV or polyinosinic-polycytidylic acid may be acting. Indeed, in immunodeficient animals, MCMV causes destructive adrenalitis (26). However, the conclusion that the IL-6 action on the HPAA is specific to viral infection (25) is not consistent with the data presented here.
The reduced corticosterone responses in IL-6-/- mice from 8 h and beyond could not be attributed to a difference in the inflammatory response as determined by limb edema. Although modestly reduced extravasation 24 h after the injection of carageenan has been reported in IL-6-/- mice together with enhanced nociceptive responses (27), limb edema in the experiments here was similar in IL-6+/+ and IL-6-/- mice. Similarly, differences could not be attributed to intrinsic differences in HPA activity, because the acute corticosterone responses to both psychological stress and acute TPS-induced inflammation were similar in IL-6+/+ and IL-6-/- mice. This IL-6 independence has also been shown in mice subjected to prolonged (16-h) restraint stress (28), although Bethin et al. (29) did observe reduced corticosterone responses in IL-6-/- mice after 30-min restraint.
To test IL-6 dependency, we used IL-6-/- mice in our initial experiments, because anti-IL-6 sera have been shown in some circumstances to protect and carry, rather than neutralize, IL-6 (30), and at least one study has shown that anti-IL-6 sera did not affect the rise in plasma corticosterone at 12 h (31). However, our subsequent experiment showed that a cross-reactive antirat IL-6 serum produced an effect similar to that seen in IL-6-/- mice. The efficacy of IL-6 neutralization suggests that deficient corticosterone responsiveness is unlikely to be attributable to any developmental effect in IL-6-/- mice and that the IL-6 inducing the HPAA response is probably at least produced, even if it does not act, on the blood side of the blood-brain barrier. Our experiments incidentally show that IL-6 dependency is not sex dependent. As only female IL-1RI-/- mice were available for evaluation of the role of IL-1 we matched the use of these mice with an evaluation of the corticosterone response in female IL-6-/- mice, and the effect was similar to that in male mice.
It is important to note that we do not make the case that plasma IL-6 is the only or indeed the most important factor in HPAA activation. Early activation clearly cannot be induced in this way, because plasma IL-6 appeared after the corticosterone response was initiated. The early activation seems likely to be neurally mediated, and we have previously shown that early HPAA activation is reduced by deafferentation of the medial basal hypothalamus or neonatal capsaicin treatment (13). The sustained c-Fos expression on the contralateral side of the PVN supports the idea that neural afferent input is a significant factor at both early and later time points and explains why responses are not more completely inhibited in IL-6-/- mice, even at later time points. As pain causes stress, and we have demonstrated that stress-induced HPAA activation is IL-6 independent, it is not surprising that the pain and stress elements of inflammation should be IL-6 independent. However, pain and stress need not be proportional to tissue damage and inflammation. The ability of the HPAA to monitor tissue inflammation via a mediator such as IL-6 would therefore seem to be physiologically advantageous, given the potent antiinflammatory role of corticosteroids.
The present experiments do not exclude the possibility that IL-6 may have contributed to the pituitary or adrenal response by a direct action on these tissues, and such effects have certainly been shown (8). However, previous experiments have shown that the stimulatory effect of IL-6 on ACTH production is almost completely blocked by anti-CRF serum (32), and IL-6 clearly had some influence on the hypothalamus in the present experiments, as shown by the reduced c-Fos staining on the ipsilateral side of the PVN in IL-6-/- mice. The importance of CRF, after TPS-induced inflammation, has been demonstrated by reversal of the ACTH response with anti-CRF serum (5). In mice without CRF type 1 receptors, both corticosterone and ACTH were still increased in response to TPS (10). However, the IL-6 response in these animals was almost 10-fold greater than that in normal mice, and these high concentrations of plasma IL-6 may have had a direct effect on the pituitary and/or adrenal cortex of these animals. Similar observations have been made by Bethin et al. (29) and Venihaki et al. (33), and data from the latter study suggest that the marked absence of an early, stress-induced, corticosterone response in CRF-deficient animals results in a failure to down-regulate the subsequent IL-6 response, even though corticosterone concentrations increase later in the response. The high plasma concentration of IL-6 may then activate the pituitary and/or adrenal cortex directly in these animals, which were shown to produce more ACTH in response to directly injected IL-6 (29). Interestingly, the IL-6-/- mice in that study were also hyporesponsive to ACTH, which may explain why, unlike our study, the IL-6-/- mice produced significantly lower corticosterone responses than the IL-6+/+ controls in response to 30-min restraint stress. Although these studies of CRH-deficient mice by Bethin et al. and Venihaki et al. (29, 33) suggest the possibility of an action on the HPAA independent of CRH, they did not exclude the action of IL-6 via release of an alternate CNS-derived pituitary secretagogue, such as AVP.
IL-1 was the first cytokine clearly shown to have a direct action on the HPAA at the level of either the pituitary (34) or the hypothalamus (35, 36). However, such data could not indicate whether IL-1 had a physiological role. Further demonstrations that activation of the HPAA by agents such as LPS, Newcastle disease virus, or MCMV could be prevented by inhibition of IL-1 action also could not support the conclusion that IL-1 was acting via a direct action on the CNS (25, 37, 38). The importance of IL-1 in HPAA activation during inflammation is confirmed by our observation of the reduced corticosterone response by IL-1RI-/- mice and is consistent with the reduced glucocorticoid response to TPS in IL-1ß-/- mice (39). However, we found no evidence for an increase in the IL-1 concentration in the plasma of mice injected with TPS despite a significant local IL-1ß response in the tissue. The response was, in fact, significantly greater in IL-6-/- mice than in the controls, something we have also seen in the subcutaneous air pouches of IL-6-/- mice injected with LPS (unpublished observations). Increased IL-1 in IL-6-/- mice is consistent with data indicating that IL-6 plays an important negative regulatory role in IL-1ß induction (40, 41) and, together with the absence of an increase in circulating IL-6 in IL-1-/- mice, further indicates that IL-1 is extremely unlikely to be an important systemic mediator of HPAA activation. However, IL-1ß is important for the induction of IL-6 following the induction of tissue inflammation by TPS, although it is not critical after systemic activation by LPS (21). Unlike the increased systemic TNF responses to LPS reported in IL-6-/- mice (20), we did not find local TNF production to be significantly increased after im injection of TPS. We have shown previously that TNF is important in both the increase in plasma IL-6 and HPAA activation induced by TPS (11, 42). However, as the time point for tissue collection was selected to demonstrate IL-1 and IL-6 responses, it may be that TNF is produced earlier, which would be consistent with the relative timing of cytokine induction in tissues previously observed (19). Our data identify IL-1 as an important tissue cytokine, responsible for induction of the IL-6, which then acts as the systemic humoral mediator of HPAA activation.
As most tissues can produce IL-6, the question arises of exactly what is the contribution of inflamed tissues to the increase in systemic IL-6. This is particularly so because it has been shown that stress and peripheral catecholamines are able to induce peripheral IL-6 production (43, 44). However, the experiments in which tourniquets were applied to inflamed limbs showed that after the injection of TPS the vast majority of IL-6 originates at the inflammatory site. It also provides, for the first time, an indication of the half-life of endogenously produced (rather than injected) IL-6 (
1.3 min). These experiments additionally argue against the possibility that cytokines, such as IL-1 or TNF, circulate to other tissues in biologically effective concentrations. If they did so, it would be expected that they would have induced IL-6 production in those tissues, and the plasma IL-6 concentration would not have fallen so precipitously.
Although IL-6 has clearly been shown to increase fever and activate the HPAA, it is generally accepted to be less efficacious than IL-1 or TNF (45, 46). There may be several reasons for this. Firstly, it has a very short plasma half-life, as shown here. Secondly, IL-6 does not induce the production of additional circulating inflammatory mediators, whereas IL-1 induces sustained production of IL-6 (47). In this respect, constant infusion of IL-6 has been shown to induce c-Fos in the PVN, and this is COX-2 dependent (48). Wang and Dunn (49) also showed that, although less potent than IL-1ß, a relatively high dose of peripheral IL-6 produced a short-lasting induction of the HPAA, but it was not effective when given intracerebroventricularly. Thirdly, inflammation and IL-1 induce the production of soluble IL-6R (sIL-6R), which binds, but does not inactivate, IL-6. The IL-6:sIL-6R complex has an increased affinity for the signal transducing accessory receptor glycoprotein (gp130), and allows it to bind and activate cells, such as those of the endothelium, that have no IL-6R themselves (50). It has previously been shown that IL-6R augments the central effects of IL-6 on several CNS-regulated parameters (51). Interestingly, glucocorticoids themselves are able to increase the expression of IL-6R (52), which suggests that early glucocorticoid responses may facilitate later IL-6-mediated regulation of the HPAA.
The means by which inflammation and circulating signals such as IL-6 signal to the brain still requires satisfactory explanation (8). Some clues are provided by data showing PG synthesis to be important not only in the mediation of fever, but for other CNS responses as well, including HPAA activation (5). Experiments that show increased PGs, or COX-2 mRNA in the microvasculature of the brain are complicated by the issue of whether LPS may be acting on endothelial cells directly (53, 54, 55). However, further work has shown increased COX-2 expression in brain endothelium after the injection of TPS (56). However, in this latter study IL-6 did not induce COX-2 expression. One of the conundrums is that IL-6 is remarkably noninflammatory in tissues, and although there are some reports that IL-6 induces PGs in non-CNS cells, this seems to be uncommon and to occur at quite high concentrations (57) despite the presence of a nuclear factor IL-6 site upstream of the COX-2 gene (58). IL-6 is apparently unable to induce PGs in peripheral mononuclear cells, although IL-6-induced fever is blocked by COX inhibitors, and IL-6 induces PGs in cerebrospinal fluid (59). Both IL-1 and IL-6 induce PGs in hypothalamic explants (60), and arachidonate products are involved in IL-6-induction of CRF release (61). However, some experiments have shown that IL-6 alone is unable to induce PGs in brain microvessels (62). It therefore seems likely that, as with CNS responses (51), the presence of sIL-6R and the short half-life of IL-6 may be critical to whether IL-6 modulates COX-2, as our data clearly show that IL-6 deficiency prevents the expression of COX-2 after the injection of TPS.
The COX-2 expression observed by us and others may seem too generalized to activate specific areas of the brain. However, PG receptor subtypes appear to be differentially expressed in different regions of the brain, and this, rather than the site of PG expression, may be a critical factor (63). Whether PGs are directed basolaterally from the endothelium or whether intravascular PGs impact upon particular brain regions before they leave the brain circulation is open to speculation. As peripheral inflammation regulates several CNS functions that are associated with IL-6 induction and PG dependency, it is conceivable that IL-6 may operate via a mechanism not requiring specific PG induction at selected endothelial sites. Given the short plasma half-life of PGs, attributable to its extremely efficient removal in the pulmonary circulation, generalized induction of PGs within the CNS vasculature might be an effective mechanism of delivering these potent mediators to the CNS.
The importance of understanding the regulatory events within the HPAA network is exemplified by the role that glucocorticoids play in regulating immune activation and inflammatory disease (64, 65, 66). Several steps remain unclear with respect to how cytokines influence the HPAA in physiological situations, particularly with respect to cellular and molecular events occurring beyond the blood-brain barrier. However, the present experiments verify that IL-6 is a physiologically important, systemic mediator in this pathway and add to previous data supporting a role that is summarized in Fig. 12.
|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: AVP, Arginine vasopressin; CNS, central nervous system; COX-2, cyclooxygenase-2; CRF, corticotropin-releasing factor; HPAA, hypothalamo-pituitary-adrenal axis; IL-1R-/-, IL-1 receptor deficient; KPBS, potassium PBS; LPS, lipopolysaccharide; MCMV, murine cytomegalovirus; NIBSC, National Institute for Biological Standards and Control; PG, prostaglandin; PVN, paraventricular nucleus; sIL-6R, soluble IL-6 receptor; TP, turpentine.
Received September 12, 2002.
Accepted for publication January 28, 2003.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and fever after peripheral inflammation in the rat. Am J Physiol 36:R1431–R1436
, IL-1ß, IL-1/ß, and IL-l receptor antagonist shows that IL-1ß is crucial in turpentine-induced fever development and glucocorticoid secretion. J Exp Med 187:1463–1475 action within the CNS markedly reduces the plasma adrenocorticotropin response to peripheral local inflammation in rats. J Neurosci 17:3262–3273 and IL-6: a dose response study. J Endocrinol Invest 19:175–182[Medline]
B and nuclear factor-interleukin-6 in the tumor necrosis factor -dependent induction of cyclooxygenase-2 in MC3T3-E1 cells. J Biol Chem 270:31315–31320This article has been cited by other articles:
 |
|
 |
|
  C. Nilsberth, L. Elander, N. Hamzic, M. Norell, J. Lonn, L. Engstrom, and A. Blomqvist The Role of Interleukin-6 in Lipopolysaccharide-Induced Fever by Mechanisms Independent of Prostaglandin E2 Endocrinology, April 1, 2009; 150(4): 1850 - 1860. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Aguilar-Valles, S. Poole, Y. Mistry, S. Williams, and G. N. Luheshi Attenuated fever in rats during late pregnancy is linked to suppressed interleukin-6 production after localized inflammation with turpentine J. Physiol., August 15, 2007; 583(1): 391 - 403. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Jankord, J. R. Turk, J. C. Schadt, J. Casati, V. K. Ganjam, E. M. Price, D. H. Keisler, and M. H. Laughlin Sex Difference in Link between Interleukin-6 and Stress Endocrinology, August 1, 2007; 148(8): 3758 - 3764. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Boelen, J Kwakkel, W M Wiersinga, and E Fliers Chronic local inflammation in mice results in decreased TRH and type 3 deiodinase mRNA expression in the hypothalamic paraventricular nucleus independently of diminished food intake J. Endocrinol., December 1, 2006; 191(3): 707 - 714. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Rummel, C. Sachot, S. Poole, and G. N. Luheshi Circulating interleukin-6 induces fever through a STAT3-linked activation of COX-2 in the brain Am J Physiol Regulatory Integrative Comp Physiol, November 1, 2006; 291(5): R1316 - R1326. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Wernstedt, A. Edgley, A. Berndtsson, J. Faldt, G. Bergstrom, V. Wallenius, and J.-O. Jansson Reduced stress- and cold-induced increase in energy expenditure in interleukin-6-deficient mice Am J Physiol Regulatory Integrative Comp Physiol, September 1, 2006; 291(3): R551 - R557. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F.-y. Gong, Y.-f. Shi, and J.-y. Deng The regulatory mechanism by which interleukin-6 stimulates GH-gene expression in rat GH3 cells. J. Endocrinol., August 1, 2006; 190(2): 397 - 406. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Chida, T. Imaki, T. Suda, and Y. Iwakura Involvement of Corticotropin-Releasing Hormone- and Interleukin (IL)-6-Dependent Proopiomelanocortin Induction in the Anterior Pituitary during Hypothalamic-Pituitary-Adrenal Axis Activation by IL-1{alpha} Endocrinology, December 1, 2005; 146(12): 5496 - 5502. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Rummel, S. W. Barth, T. Voss, S. Korte, R. Gerstberger, T. Hubschle, and J. Roth Localized vs. systemic inflammation in guinea pigs: a role for prostaglandins at distinct points of the fever induction pathways? Am J Physiol Regulatory Integrative Comp Physiol, August 1, 2005; 289(2): R340 - R347. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Callahan, X. Ming, S. K. Lu, L. J. Brunner, and M. A. Croyle Considerations for Use of Recombinant Adenoviral Vectors: Dose Effect on Hepatic Cytochromes P450 J. Pharmacol. Exp. Ther., February 1, 2005; 312(2): 492 - 501. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Rummel, T. Hubschle, R. Gerstberger, and J. Roth Nuclear translocation of the transcription factor STAT3 in the guinea pig brain during systemic or localized inflammation J. Physiol., June 1, 2004; 557(2): 671 - 687. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) or saline (
) before collection of trunk blood and removal of the injected hind limb at the times indicated. The water content of the limb (A), plasma concentrations of corticosterone (B), and plasma concentrations of IL-6 (C) were determined.
), before collection of blood by cardiac puncture at the times indicated. The individual concentrations of plasma ACTH at 1 h (A) and 12 h (B) are indicated. The significance of differences between saline controls and TPS-treated animals is shown (by unpaired t test).
) or 0.5 ml anti-IL-6 serum (
