Role of endogenous interleukin-1 receptor antagonist in regulating fever induced by localised inflammation in the rat

Abstract

1. Interleukin (IL)-1 is a mediator of host defence responses to inflammation and injury, including fever, but its sites of synthesis and action have not been fully elucidated. The actions of IL-1 are antagonised by IL-1 receptor antagonist (IL-1ra). The present study tested the hypothesis that IL-1 and IL-1ra are produced locally at sites of peripheral inflammation in rats, and that endogenous IL-1ra acts to limit the fever resulting from the inflammation.

2. Injection of lipopolysaccharide (LPS; 100 μg kg⁻¹) into a subcutaneous air pouch (i.po.) of rats induced a significant increase in body temperature. Virtually all (∼85 %) of the injected LPS was recovered from the pouch between 1 and 8 h (when the experiment was terminated) after injection of LPS, but LPS was undetectable (< 50 pg ml⁻¹) in plasma at any time. Concentrations of immunoreactive IL-1α and IL-1β were increased significantly in the pouch at 1, 2, 3, 5 and 8 h after injection of LPS, corresponding with the rise in body temperature and the fever peak. The appearance of IL-1ra was delayed until 2 h. Thereafter, the concentrations of IL-1β and IL-1ra increased in parallel with the development of fever, while the concentrations of IL-1α remained constant. IL-1ra, but not IL-1α or IL-1β, was detected in significant quantities in the plasma of LPS-injected animals.

3. Treatment of rats with an anti-IL-1ra serum (2 ml, i.po.) at the time of injection of LPS (10 or 100 μg kg⁻¹, i.po.) abolished the appearance of IL-1ra in the circulation. Although neutralisation of endogenous IL-1ra did not affect the maximum body temperature reached after injection of submaximum (10 μg kg⁻¹, i.po.) or maximum (100 μg kg⁻¹, i.po.) doses of LPS, the duration of the fever was significantly prolonged, and was associated with a 3- to 4-fold increase in immunoreactive IL-1β concentrations in the pouch fluid, but not in the plasma, at the 8 h time point.

4. These data show that effects of local (i.po.) injection of LPS are not due to its action in the circulation or at distant sites (such as at the blood-brain barrier). These data also show that locally produced IL-1ra, in response to injection (i.po.) of LPS, inhibits the production and/or action of locally produced IL-1β. The ability of IL-1ra to limit the duration, rather than the magnitude of the fever, is consistent with its delayed production, relative to IL-1. IL-1ra, therefore, appears to play a key role in the resolution of fever induced by localised inflammatory responses.

The pro-inflammatory cytokine interleukin-1 (IL-1) is a pivotal mediator of local and systemic responses to infection and inflammation, of which fever is the most widely studied, experimentally and clinically (see Kluger, 1991; Dinarello, 1996, for reviews). The IL-1 family comprises two agonists, IL-1α and IL-1β, and a highly selective, endogenous IL-1 receptor antagonist (IL-1ra) (reviewed by Dinarello, 1996). Administration of recombinant IL-1α or IL-1β, systemically or directly into the brains of experimental animals, causes fever (Anforth et al. 1998) which is prevented by IL-1ra (Opp & Kreuger, 1991). Inhibition of the actions of IL-1, peripherally or in the brains of rodents, by administration of neutralising anti-IL-1 sera or IL-1ra, markedly attenuates fever induced by systemic injection of the (exogenous) pyrogen bacterial endotoxin (lipopolysaccharide, LPS) (Long et al. 1990; Smith & Kluger, 1992; Klir et al. 1994; Luheshi et al. 1996; Cartmell et al. 1999). Despite its pyrogenic action in the periphery, little or no IL-1 is detected in the circulation of febrile animals or humans during infection or injury (Damas et al. 1992; Engel et al. 1994; Luheshi et al. 1997; Miller et al. 1997b). Our previous studies demonstrated that the main source of IL-1, in response to a localised inflammatory response, is within infected or inflamed tissues, and it is here, rather than in the circulation, that the biological activity of IL-1 is largely manifest (Miller et al. 1997a,b).

The overall bioactivity of IL-1 appears to be determined by the relative concentrations of IL-1 and IL-1ra, rather than by the absolute concentration of IL-1 alone (see Dinarello, 1996, for review; Hirsch et al. 1996; Gabay et al. 1997). IL-1ra, like IL-1, is induced by inflammatory stimuli, and prevents the actions of IL-1 (Dinarello & Thompson, 1991; reviewed by Arend, 1993; Dinarello, 1996), albeit at molar ratios of 500:1 or greater. The concentration of circulating IL-1ra in disease states is much higher than that of IL-1, and the IL-1ra production appears to be delayed and prolonged relative to that of IL-1 (Fischer et al. 1992b; Arend, 1993; Matsukawa et al. 1993; see Dinarello, 1996). Apart from fever, IL-1ra also inhibits other aspects of host defence responses in rodents. For example, large doses of recombinant IL-1ra improve survival rates during endotoxic shock (Ohlsson et al. 1990; Alexander et al. 1991; Wakabayashi et al. 1991; Fischer et al. 1992a; Fisher et al. 1994), attenuate the manifestations of experimental colitis (Cominelli et al. 1990), decrease IL-1-induced lethality in adrenalectomized mice (Besedovsky et al. 1986; Mengozzi et al. 1991), and reduce inflammation in experimental arthritis (Henderson et al. 1991; Wooley et al. 1993; Makarov et al. 1996). These studies, together with studies involving administration of neutralising anti-IL-1ra sera (Dinarello & Thompson, 1991; Chensue et al. 1993; Ferretti et al. 1994; Hirsch et al. 1996), illustrate an important role for IL-1ra in responses to infection and inflammation. Interestingly, IL-1ra also has been shown to be critical in a normal developmental process (linear growth) in the absence of a specific pathogenic stimulus (Hirsch et al. 1996; Horai et al. 1998).

In spite of these observations, the sites of production of IL-1ra during localised infection or inflammation and its functional role in fever are unknown. The objective of this study, therefore, was to investigate the peripheral sites of endogenously produced IL-1 and IL-1ra during fever induced by a localised injection of LPS in the rat air pouch model of inflammation, and to test the hypothesis that locally released endogenous IL-1ra acts to limit this fever.

METHODS

Male Sprague-Dawley rats (Charles River, Margate, Kent, UK) (250-300 g) were used in all experiments. Animals were housed in a controlled environment at an ambient temperature of 21 ± 2°C on a 12 h:12 h light:dark cycle (light on from 08.00 to 20.00 h). Food (pelleted rat chow, Beekay International, UK) and water were provided ad libitum. All procedures were performed under the UK Animals (Scientific Procedures) Act, 1986.

Measurement of body temperature and air pouch formation

Core body temperature of rats was measured (to an accuracy of 0.1°C), by remote biotelemetry (Data Quest IV system, Data Sciences, St Paul, MN, USA), using pre-calibrated radiotransmitters (TA10TA-F40, Data Sciences) implanted intraperitoneally whilst animals were under halothane (Fluorothane, Zeneca, Cheshire, UK) anaesthesia (3 % in oxygen). Animals were housed individually for 24 h before the experiments. Transmitter output frequency (Hz) was monitored, at 10 min intervals, by an antenna mounted in a receiver board, situated beneath the cage of each animal, and the data were logged into a peripheral processor (BCM 100, Data Sciences) connected to a personal computer.

A subcutaneous air pouch was formed immediately after implantation of the radiotransmitter (day 1), while animals were under halothane anaesthesia, as described previously (Edwards et al. 1981). Briefly, 20 ml of sterile air (0.2 μm Acrodisc, Gelman Sciences, USA) was injected into the subcutaneous tissue of the dorsal mid-line, caudal to the scapulae. Three days after the initial pouch formation, animals were again briefly anaesthetised (3 % halothane in oxygen) and the air pouches were reinflated with a further 10 ml of sterile air, to maintain open cavities. On day 6, treatments or vehicle were injected directly into the air pouches of lightly restrained (hand held), conscious animals.

Experimental protocol

Experiment 1

To investigate the distribution and time course of appearance of LPS and locally produced and circulating cytokines, animals were injected i.po. with LPS (100 μg ml⁻¹ kg⁻¹) (Escherichia coli serotype 0128:B12, Sigma, UK) or vehicle (pyrogen-free saline, 1 ml kg⁻¹). Pouch fluid and plasma samples were collected for LPS and cytokine analyses (IL-1α, IL-1β and IL-1ra), from groups of animals (n= 5 per treatment per time point) under terminal anaesthesia (with halothane), before (0 h) and at 1, 2, 3, 5 or 8 h after injection. Blood was collected by cardiac puncture, into sterile tubes containing heparin (10 u ml⁻¹) and centrifuged (5300 g, 4°C, 10 min). Plasma was stored at -70°C until assayed. Animals were killed by cervical dislocation and pouch fluid extracted. Sampling of the inflammatory exudate within the pouch was achieved by lavaging the pouch with 1 ml of sterile, pyrogen-free saline. The lavage fluid was quickly aspirated and centrifuged (3000 g, 4°C, 10 min), and the resultant supernatant stored at -70°C until assayed.

Experiment 2

The role of locally produced endogenous IL-1ra was investigated by inhibiting its action, at the site of inflammation, with a sheep anti-rat IL-1ra serum (raised against recDNA rat IL-1ra; National Institute for Biological Standards and Control (NIBSC), Potters Bar, Herts, UK). Control animals were injected with an equivalent volume of normal (pre-immune) sheep serum (NSS; NIBSC). The anti-IL-1ra serum recognises both recombinant (E. coli-derived) and natural rat IL-1ra, but does not cross-react with rat recombinant (rr)IL-1α or IL-1β (NIBSC). Animals were co-injected (i.po.) with NSS (n= 9) or sheep anti-rat IL-1ra serum (2 ml, n= 10) and either saline (1 ml kg⁻¹, n= 4) or a dose of LPS which induces either a submaximum fever (10 μg ml⁻¹ kg⁻¹, n= 5) or a maximum fever (100 μg ml⁻¹ kg⁻¹, n= 5) (data not shown). Body temperature was monitored continuously throughout the experiment. Plasma and pouch fluid samples (for assay of cytokines, see above) were collected 8 h after injection, from animals under terminal anaesthesia (with halothane). Animals were killed by cervical dislocation.

Endotoxin assay

LPS concentrations in rat pouch fluid and plasma were measured using a Limulus amoebocyte lysate (LAL) test (European Pharmacopoeia, 1999). The First International Standard for endotoxin (LPS, 84/650, WHO) was used as the standard, and endotoxin concentrations were expressed in nanograms per millilitre (1 ng of the LPS injected = 6 international units (i.u.)). Before testing, samples were diluted 1:16 to 1:10 000 with endotoxin-free water to reduce interference of plasma components in the LAL test. Spiked plasma controls were included in the assay to ensure that the plasma itself did not interfere with the test. The sensitivity of the LAL assay was 5 pg ml⁻¹.

Measurement of cytokine concentrations

Rat-specific enzyme-linked immunoabsorbent assay (ELISA)

Concentrations of IL-1α, IL-1β and IL-1ra in the various biological fluids (see above) were quantified using rat-specific sandwich ELISAs (NIBSC), as described previously for IL-1β and IL-1ra (Safieh-Garabedian et al. 1995; Cunha et al. 2000). The ELISA for rat IL-1β detects both precursor and mature IL-1β but not rat IL-1α or rat IL-1ra (NIBSC). The ELISA for rat IL-1ra does not cross-react with rat IL-1α or rat IL-1β (NIBSC). The protocol for the ELISA for rat IL-1α was similar to the protocols for the ELISAs for rat IL-1β and rat IL-1ra. Briefly, sheep polyclonal anti-rat IL-1α IgG, raised against rat recombinant IL-1α (2 μg ml⁻¹) in 100 μl PBS buffer, was used to coat microtitre plates (Nunc Maxisorb). After incubation (4°C overnight) and washing the plates in assay buffer (0.01 m phosphate, 0.05 m NaCl, 0.1 % Tween 20, pH 7.2), 100 μl of standard (rat recombinant IL-1α), or sample, was added to each well and incubated overnight at 4°C. After washing the plates, 100 μl of biotinylated, sheep polyclonal anti-rat IL-1α IgG (diluted 1:500 with assay buffer + 1 % normal sheep serum) was added to the plates and incubated for 1 h at room temperature. Subsequent steps were as for the ELISA of rat IL-1β (Safieh-Garabedian et al. 1995). The ELISA for rat IL-1α does not cross-react with rat IL-1β or rat IL-1ra (NIBSC). Quality controls for all ELISAs consisted of aliquots of pooled rat plasma samples that had been spiked with recombinant cytokine specific for that ELISA (100 and 1000 pg ml⁻¹) and frozen at -70°C. The sensitivities of the IL-1β and IL-1ra ELISAs were similar at ∼5 pg ml⁻¹, and that of IL-1α was ∼10 pg ml⁻¹. Given the high concentrations of cytokines at the site of inflammation compared to those in the circulation, some samples required large dilutions. The detection limit, which allows for the sample dilution factor, therefore differs markedly between different biological fluids and treatment groups, and is reported in Results.

Bioassay

IL-1 bioactivity in plasma and pouch fluid samples was measured using the D10(N4)M cell line as described previously (Hopkins & Humphreys, 1989). This bioassay does not differentiate between the α- and β-subtypes, but allows measurement of mature, biologically active IL-1 (Hopkins & Humphreys, 1989). The assay is based on the survival of the D10(N4)M subline (Hopkins & Humphreys, 1989) of the D10.G4.1 murine T cell clone in response to IL-1 (Kaye et al. 1983). Cell responses were determined by their ability to metabolise 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (Sigma) at the end of the assay periods. Plasma samples were pretreated with polyethylene glycol (PEG) before IL-1 bioassay to reduce non-specific interference from uncharacterised factors present in physiological fluids (Hopkins & Humphreys, 1990), which may obscure low concentrations of bioactive IL-1 in the plasma. Bioassay values are expressed in terms of the First International Standard for human IL-1β (86/680, 1 μg = 100 000 i.u., NIBSC). Results were calculated by use of the ELIPS program (Bowman & Co. Ltd (Software), Boreham Wood, UK). Quality controls for assays of IL-1 plasma consisted of aliquots of pooled rat plasma samples that had been spiked with IL-1 (1 and 10 i.u. ml⁻¹) and frozen at -70°C. The assay detection limits, after allowing for dilution of samples into the assays, in the various biological fluids, are reported in the Results section.

interleukin-1 in vitro cleavage assay

Western blot analyses of IL-1 were performed on pouch fluid exudates. Since the IL-1β ELISA recognises both precursor (31 kDa) and mature (17.5 kDa) IL-1β, an in vitro cleavage assay (as described in Culhane et al. 1998) was performed on samples to distinguish whether the larger IL-1β band was indeed proIL-1β and could, therefore, be cleaved by ICE (caspase-1), to mature IL-1β. Briefly, 10 μl of pouch fluid, with or without the caspase inhibitor zVAD-DCB (100 μm) was pre-incubated at 37°C for 30 min in 25 μl caspase-1 reaction buffer. Five hundred units of recombinant human caspase-1 and 10 mm DTT, required for enzyme activation, were then added, the reaction volume was made up to 75 μl with caspase-1 reaction buffer and the cleavage assay incubated at 25°C for 30 min. The products were immunologically detected with both rabbit and sheep anti-rat IL-1β antibodies (NIBSC).

Statistical analysis

All results are expressed as means ±s.e.m. for the number of animals given. Temperature responses were plotted as abdominal temperature-time curves and data analysed using either analysis of variance (ANOVA), followed by a Tukey-Kramer multiple comparisons post hoc test for differences between more than two groups, or Student’s t test for differences in maximum temperatures, at the same time point, between two groups. Comparisons between cytokine concentrations, in pouch fluid and plasma, at the various time points were made by comparing treatment with control values at the same time point, using Student’s t test. Comparisons across groups and/or time points were not analysed. Where cytokine concentrations were undetectable, samples were assigned a value equivalent to the detection limit of the assay. A two-tailed probability P < 0.05 was considered statistically significant.

RESULTS

experiment 1: relationship between il-1α, il-1β and il-1ra at the site of inflammation and in plasma after localised (intrapouch) injection of lps or saline

Injection (i.po.) of LPS (100 μg ml⁻¹ kg⁻¹, 198 000 ng ml⁻¹ kg⁻¹) caused a rise in core temperature that was apparent 1.5 h after injection and peaked (at 38.8 ± 0.1°C) 3 h after injection (Fig. 1A). The temperature of vehicle-injected controls did not change during the course of the experiment (Fig. 1A). The pouch concentrations of LPS and immunoreactive IL-1α, IL-1β and IL-1ra were below the detection limit of each respective assay before the injection of LPS (Figs 1B and 2A and B) and in vehicle-injected (control) animals (data not shown). LPS was detected in the pouch at all the time intervals tested between 1 and 8 h after LPS injection (25 920 ± 7325 ng ml⁻¹ at 1 h, declining to 12 150 ± 1350 ng ml⁻¹ at 8 h, Fig. 1B). In contrast, plasma concentrations of LPS were below the detection limit of the LAL test (50 pg ml⁻¹) at all time points (data not shown). Fever evoked by LPS (100 μg kg⁻¹, i.po.) was associated with significant increases in concentrations of immunoreactive IL-1β and IL-1ra in the pouch (Fig. 2A). Immunoreactive IL-1β (2423 ± 738 pg ml⁻¹ at 1 h) and IL-1α (1339 ± 270 pg ml⁻¹ at 1 h) appeared before immunoreactive IL-1ra (< 40 pg ml⁻¹, at 1 h; 1387 ± 598 pg ml⁻¹, at 2 h). The increase in IL-1α and IL-1β in the pouch (1 h), preceded the rise in body temperature (Figs 1A and 2A). In response to LPS, pouch concentrations of immunoreactive IL-1α remained relatively constant during the 8 h following LPS injection (Fig. 2A). Immunoreactive IL-1β concentrations in the pouch increased gradually and were maximum 5 h after injection of LPS (5 h, 53 800 ± 5358 pg ml⁻¹vs. vehicle (controls), < 40 pg ml⁻¹, n= 5), and then decreased again by the 8 h time point (15 800 ± 2996 pg ml⁻¹, n= 4). Pouch IL-1ra concentrations increased proportionately more than IL-1β between 2 and 3 h, but then remained relatively stable until the last measurement (8 h after injection of LPS: IL-1ra, 6259 ± 1124 pg ml⁻¹, n= 4 at 8 h, compared with 1387 ± 598 pg ml⁻¹, n= 5 at 2 h and control values of < 40 pg ml⁻¹, n= 4). The concentrations of bioactive IL-1 (D10(N4)M) in the pouch fluid mirrored the changes in immunoreactive IL-1β (ELISA) concentrations. Bioactive IL-1 concentrations were elevated significantly (144 ± 43 i.u. ml⁻¹) 1 h after injection of LPS (data not represented in a figure), and were maximum 5 h after injection of LPS (i.u. ml⁻¹: 2 h, 326 ± 97; 3 h, 388 ± 56; 5 h, 493 ± 115; 8 h, 347 ± 71), compared to vehicle-injected (control) animals (5 h, 3 ± 2 i.u. ml⁻¹). Bioactive IL-1 concentrations in response to i.po. LPS challenge were, however, lower than immunoreactive IL-1β concentrations detected throughout the time course investigated (8 h). Pouch fluid exudates were examined for the presence of the pro- and mature forms of IL-1β by Western blot analysis. Pouch fluid from animals injected with LPS showed the presence of two IL-1β bands of immunoreactivity. The lower band in the supernatants migrated at the same molecular size as recombinant mature (active) 17 kDa IL-1β and the upper band at ∼33 kDa corresponding to the size of proIL-1β. After incubation with caspase-1, the 33 kDa band was no longer present, and all IL-1β migrated at 17 kDa (data not shown). These data indicate that both pro- and mature IL-1β were induced by LPS. Plasma immunoreactive IL-1α and IL-1β (Fig. 2B) and bioactive IL-1 concentrations (data not represented in a figure) were below the assay detection limits (IL-1α, < 100 pg ml⁻¹; IL-1β, < 38 pg ml⁻¹; bioactive IL-1, < 3 ± 1 i.u. ml⁻¹) in all samples tested. IL-1ra first appeared in the circulation 3 h after (i.po.) LPS (IL-1ra, 1685 ± 74 pg ml⁻¹ compared with controls < 38 pg ml⁻¹) and was still present at 8 h (IL-1ra, 223 ± 111 pg ml⁻¹).

Figure 1. Fever and local and plasma concentrations of lipopolysaccharide (LPS) after injection (intrapouch, i.po.) of LPS.

A, temperature responses (means ±s.e.m.) of rats injected with vehicle (saline, ○, 1 ml kg⁻¹, i.po.) or with LPS (•, 100 μg kg⁻¹, i.po.) at time 0 h (arrow). B, concentrations of LPS at the site of inflammation (air pouch lavage fluid) at specific time points after injection (i.po.) of vehicle (saline, □) or LPS (▪). Data are expressed as means ±s.e.m.*** The earliest significant differences compared with vehicle-injected (saline control) animals, at the same time point, P < 0.001 (Student’s t test).

Figure 2. Time course of appearance of cytokines during fever.

Time course of the appearance of immunoreactive IL-1α (□), IL-1β (▪) and IL-1ra () at the site of inflammation (pouch, A), and in plasma (B) at various time points after injection of LPS (100 μg kg⁻¹, i.po.) (n= 5 per time point). Data are expressed as means ±s.e.m. IL-1α, IL-1β and IL-1ra were undetectable in pouch fluid and plasma of vehicle-injected (saline control) animals (data not shown). *** The earliest significant differences compared with vehicle-injected (saline control) animals, at the same time point, P < 0.001 (Student’s t test). Horizontal dotted line indicates assay detection limit.

Experiment 2: effect of neutralising endogenous IL-1ra within the pouch on responses to submaximum and maximum doses of LPS

In a separate series of experiments naive rats were co-injected (i.po.) with LPS, together with anti-IL-1ra serum (2 ml) to observe any effects locally released IL-1ra may have had on fever. A low (10 μg kg⁻¹) and high (100 μg kg⁻¹) dose of LPS were used to allow for the fact that the antiserum might be insufficient to block the large amounts of IL-1ra induced by the highest dose of LPS. Injection (i.po.) of antiserum with vehicle had no effect on basal body temperature for 24 h after injection (data not shown). The body temperature of rats injected with NSS and LPS (4 h, LPS, 10 μg kg⁻¹, 38.5 ± 0.2°C, n= 5; 4 h, LPS, 100 μg kg⁻¹, 38.7 ± 0.2°C, n= 4) was significantly elevated (P < 0.001, ANOVA) compared with rats injected with saline and NSS (4 h, 37.0 ± 0.1°C, n= 4). Anti-IL-1ra serum had no effect on the LPS-induced peak fever (4 h, LPS, 10 μg kg⁻¹, 38.7 ± 0.3°C, n= 5; LPS, 100 μg kg⁻¹, 38.9 ± 0.3°C, n= 4, Fig. 3A and B), but prolonged the duration of the fever (Fig. 3A and B). Pouch fluid immunoreactive IL-1β concentrations were significantly elevated in animals injected with NSS and LPS (LPS, 10 μg kg⁻¹, IL-1β: 11 516 ± 271 pg ml⁻¹, n= 5; LPS, 100 μg kg⁻¹, IL-1β: 13 028 ± 348 pg ml⁻¹, n= 4) compared with controls (NSS and saline, IL-1β: 190 ± 55 pg ml⁻¹, n= 4), but were markedly lower than in animals injected with anti-IL-1ra serum and LPS (LPS, 10 μg kg⁻¹, IL-1β: 46 636 ± 1134 pg ml⁻¹, n= 5, P < 0.001; LPS, 100 μg kg⁻¹, IL-1β: 40 060 ± 1677 pg ml⁻¹, n= 4, P < 0.001; Fig. 4A). IL-1β was undetectable in the plasma of animals injected with NSS and LPS (< 38 pg ml⁻¹), or NSS and saline (< 38 pg ml⁻¹), or anti-IL-1ra serum and LPS (< 38 pg ml⁻¹). Treatment of animals with anti-IL-1ra serum had no effect on plasma concentrations of IL-1β resulting from injection of LPS (Fig. 4B).

Figure 3. Effect of anti-IL-1ra serum, at the site of inflammation, on LPS-induced fever.

Temperature responses (mean ±s.e.m.) of rats injected (i.po.) with anti-IL-1ra serum (2 ml, filled symbols) or normal sheep (pre-immune) serum (NSS, 2 ml, open symbols) together with either LPS (10 μg kg⁻¹, n= 4 or 5 per group, circles) or saline (1 ml kg⁻¹, n= 4, triangles) (A) or LPS (100 μg kg⁻¹, n= 4 or 5 per group, squares) or saline (1 ml kg⁻¹, n= 4, triangles) (B) at time 0 h (arrow).

Figure 4. Effect of anti-IL-1ra serum, at the site of inflammation, on LPS-induced concentrations of cytokines.

A, concentrations of immunoreactive IL-1β at the site of inflammation (pouch) 8 h after co-injection (i.po.) of NSS and saline (n= 4, □), NSS and LPS (10 μg kg⁻¹, n= 4, ▪), anti-IL-1ra and LPS (10 μg kg⁻¹, n= 5,), NSS and LPS (100 μg kg⁻¹, n= 4,) and anti-IL-1ra and LPS (100 μg kg⁻¹, n= 5,). B, plasma concentrations of immunoreactive IL-1β and IL-1ra in the same animals (see A for details). ***P < 0.001 (Student’s t test, compared with NSS + treatment, at the same time point). Horizontal dotted line indicates assay detection limit.

DISCUSSION

The objective of the present study was to investigate, in rats, the relationship between local and circulating concentrations of IL-1 and IL-1ra during localised LPS-induced inflammation (which causes fever), and to test the hypothesis that endogenous IL-1ra limits the fever. Our results agree with earlier suggestions (Miller et al. 1997a,b) that the actions of IL-1 are largely manifest locally, in infected or inflamed tissues, where it acts to induce fever. Further, we provide direct evidence that endogenous IL-1ra acts locally (at the site of inflammation) to reduce local concentrations of IL-1β, and ultimately to limit fever.

Most previous studies investigating the relationship between endogenous IL-1 and IL-1ra have examined the presence of these two cytokines in the circulation, in response to intraperitoneal or intravenous injection of LPS. These protocols are intended to mimic severe conditions, such as endotoxaemia or sepsis. However, in such conditions, endotoxin would have access to numerous potential sites of action, including the blood-brain barrier and thus, potentially, the brain itself (see Dinarello et al. 1999). In the present study, in which fever was induced by LPS injection into a subcutaneous air pouch, LPS was undetectable in the circulation at any time (up to 8 h). This cannot be ascribed to insensitivity of the LAL assay, since the plasma concentrations of LPS of a control group of animals injected intraperitoneally with LPS (100 μg ml⁻¹ kg⁻¹) was 205 ng ml⁻¹ at 2 h (data not shown), and inclusion of spiked plasma controls in the assay confirmed that plasma did not interfere with the assay at the dilutions tested. The slow disappearance of LPS from the pouch and failure to appear in the circulation may be ascribed to degradation of LPS in pouch macrophages, or neutralisation by endogenous anti-endotoxin antibodies (Barclay, 1990) since the LAL assay detects active endotoxin only. Moreover, induction of inflammation in a 6-day-old air pouch provokes migration of polymorphonuclear leukocytes into the pouch (Konno & Tsurufuji, 1983; Martin et al. 1994; Ahluwalia & Perretti, 1996) and degradation of LPS has been reported in neutrophils (Hall & Munford, 1983) and macrophages (Munford & Hall, 1985), the latter being 10-40 % in the lining of a 6-day-old air pouch (Edwards et al. 1981). Since neutrophil and macrophage infiltration were not investigated in the present study, our hypothesis as to the disappearance of LPS from the pouch is speculative and requires further investigation. Importantly, these data indicate that injection of LPS into a subcutaneous air pouch is a useful protocol for studying the effects of localised inflammation and suggest that cytokine production and actions in response to intrapouch injection of LPS are not due to the release of LPS in the circulation or LPS acting at distant sites (such as at the blood-brain barrier).

Injection of LPS into the air pouch produced a dose-related fever, which was preceded by rapid production of immunoreactive IL-1α and IL-1β, and accompanied by production of IL-1ra at the site of inflammation. However, IL-1ra only, but not IL-1α or IL-1β, was detected in the circulation. These data do not support the possibility that IL-1α or IL-1β is a major circulating cytokine in this model. The time course of events in the present study shows that local IL-1α and IL-1β first appeared about 1 h before IL-1ra. This is in agreement with previous findings in vivo (Fischer et al. 1992b; Matsukawa et al. 1993) and in vitro (DeRochemonteix et al. 1993), and is consistent with reports that IL-1 induces IL-1ra (Bargetzi et al. 1993; Ilyin & Plata-Salamán, 1996). Mice lacking the IL-1β gene show reduced expression of IL-1ra in the brain (< 1/10 of the wild-type level), after injection of turpentine (Horai et al. 1998), though IL-1β does not induce substantial IL-1ra release from peripheral blood mononuclear cells in vitro (Poutsiaka et al. 1991).

IL-1α concentrations increased at 1 h and remained constant thereafter (8 h). Measurements of IL-1α are relatively rare because this cytokine remains primarily cytosolic or membrane associated and is released only under unusual in vitro or in vivo conditions, possibly as a result of cell death (see Dinarello, 1996, for review). The significance of increased IL-1α at the site of inflammation is intriguing, therefore, given that endogenous IL-β, but not IL-1α, is crucial for mediating fever in response to LPS (Long et al. 1989, 1990) or localised inflammation (Zheng et al. 1995; Horai et al. 1998) in rats and mice.

In the present study, pouch IL-1β immunoreactivity was significantly greater than IL-1ra immunoreactivity in response to LPS at all time points investigated. Immunoreactive IL-1ra, but not immunoreactive IL-1β, was detected in plasma after LPS was injected into the pouch. Previous studies of endotoxaemia or sepsis in man (Granowitz et al. 1991; see Dinarello, 1996) and experimental animals (Fischer et al. 1992b; Matsukawa et al. 1993) reported plasma concentrations of IL-1ra significantly greater than those of IL-1. In the present study, Western blot analysis of pouch fluid revealed that both the precursor and mature forms of IL-1β were present within the pouch. Incubating pouch fluid samples with caspase-1 resulted in the appearance of mature IL-1β alone (data not shown). Thus, LPS induces both pro- and mature (active) IL-1β, which could account for the higher local concentrations of IL-1β in relation to IL-1ra. Locally released mature (i.e. bioactive) IL-1 is likely to have been sufficient to mediate the fever. We have shown recently that injection of IL-1β (0.3 μg kg⁻¹) directly into the pouch of rats induces a robust fever (1.5°C change in body temperature) (Cartmell et al. 2000), and inhibiting IL-1 action, either at the site of inflammation, or in the brain, attenuates LPS-induced fever (Luheshi et al. 1996; Miller et al. 1997a; Cartmell et al. 1999).

Despite the increased local production of IL-1α, IL-1β and IL-1ra in response to LPS administration, only IL-1ra was detected in the circulation. The absence of significant increases in circulating immunoreactive IL-1α and IL-1β and bioactive IL-1 is not attributable to a failure of the assays to detect IL-1 in plasma: plasma controls spiked with small amounts of IL-1 yielded positive results. Also, it is unlikely that the IL-1ra induced in the pouch, in response to LPS, prevented the release of IL-1 from the pouch into the circulation, since inhibiting endogenous IL-1ra action with anti-IL-1ra serum at the site of inflammation (such that local concentrations of IL-1β were increased 4-fold), did not result in detectable concentrations of IL-1 in plasma (present study). In the light of the present data it is unclear why IL-1 is selectively retained at the site of inflammation, while IL-1ra enters the circulation. It is interesting that TNF-α similarly is detectable in the pouch only and not the plasma, after injection of LPS (Miller et al. 1997b). A possible explanation is that both IL-1 and TNF bind to soluble forms of the receptors, which are induced by LPS. These ligand-receptor complexes, unlike those of IL-6 and its receptor, are biologically inactive. Their size might inhibit them from entering the circulation readily and they may be immunoreactive. Clearly, IL-1ra also binds to IL-1 receptors but soluble IL-1RII, which is in any case a non-signalling, decoy receptor, has a particularly low affinity for IL-1ra (Symons et al. 1995). Therefore, IL-1, like LPS, remains localised in the pouch during inflammation and its presence in the circulation is not required for the induction of fever. We have shown previously that IL-1 can act locally, at the site of inflammation, to induce the synthesis and release of IL-6 into the circulation (Luheshi et al. 1997; Miller et al. 1997a). This endogenous circulating IL-6 appears to be an essential mediator of the febrile response to local LPS-induced inflammation in rats (Cartmell et al. 2000). Experimental evidence has been presented recently supporting the possibility of the transport of peripheral immune signals to the brain via vegetative and peripheral afferent nerves (see Zeisberger, 1999, for review) contributing to the pyrogenic signalling to activate host defence responses such as fever.

Our earlier studies have shown that injection of exogenous IL-1ra into the subcutaneous air pouch, but not intraperitoneally or into the brain, decreases pouch IL-1 bioactivity and reduces LPS-induced fever (Miller et al. 1997a). IL-1ra has been described as an acute phase protein (Gabay et al. 1997), and may suppress the inflammatory consequences of early IL-1 release during the acute phase response. The observed increase in local and circulating concentrations of endogenous IL-1ra in response to peripheral inflammation supports the hypothesis that IL-1ra acts in vivo to limit inflammation by inhibiting the actions of IL-1. IL-1ra, therefore, like IL-1, may act locally and inhibit the fever that follows injection (i.po.) of LPS. Interestingly, data from the present study reveal that animals treated with anti-IL-1ra serum have significantly elevated concentrations of plasma immunoreactive IL-6 (data not shown), and we have shown previously that injection (i.po.) of IL-1ra inhibits the rise in plasma bioactive IL-6 concentrations in response to localised injection (i.po.) of LPS (Miller et al. 1997a). It is plausible therefore that IL-1ra may act as a regulator of IL-6 at the site of inflammation, via its regulation of IL-1 (which stimulates IL-6 production).

In the present study, the role of endogenous IL-1ra in inflammation was investigated by neutralising rat IL-1ra by injection of specific antiserum (anti-rat IL-1ra serum), into the site of inflammation. Although inhibition of endogenous IL-1ra augmented local concentrations of IL-1 (at 5 h (data not shown) and at 8 h), it did not exacerbate the peak fever response to injection into the pouch of LPS (at doses which induced either a submaximum or a maximum fever), but did significantly prolong (for at least 8 h after injection of LPS compared with 5 h in controls) the duration of the LPS-induced fever. This is consistent with the delayed appearance of IL-1ra, and suggests that endogenous IL-1ra may act to limit the duration of localised LPS-induced fever. IL-1ra was not detected until 1 h after IL-1α and IL-1β appeared in the pouch (and even then at lower concentrations), and it was about 1-2 h after the commencement of LPS-evoked fever (3-4 h after injection) that temperature responses in animals given LPS and anti-rat IL-1ra serum began to deviate from those given LPS without anti-IL-1ra. That treatment with anti-IL-1ra did not influence the maximum magnitude of the fever cannot be explained in the light of the present data and cytokine levels in the pouch and plasma were not measured during the peak increases in body temperature (3-4 h). LPS also induced elevated circulating concentrations of IL-1ra in preimmune serum-treated animals yet, not surprisingly, in animals injected with anti-IL-1ra serum, IL-1ra was not detectable in plasma. Consequently, it is possible that the anti-IL-1ra serum used entered the circulation, or plasma IL-1ra derives mainly from the pouch, or IL-1ra is induced in the circulation, possibly via another mediator. Although our data suggest that IL-1ra appearing in the circulation originates from the pouch this evidence is largely circumstantial and further studies are needed to clarify this. It is possible that the IL-1ra detected in the circulation 3 h after LPS injection is in fact induced in the plasma by another circulating mediator, such as IL-6. Our recent observations have demonstrated that fever evoked in response to a localised injection of LPS into the air pouch, is mediated by circulating IL-6 (Cartmell et al. 2000), and that this IL-6 increases in the circulation over the time course of the febrile response (Miller et al. 1997b; Cartmell et al. 2000). Other investigators have shown that IL-6 is a potent inducer of IL-1ra in vivo (Tilg et al. 1994). The source from which plasma IL-1ra derives requires further investigation.

In conclusion, the data presented here confirm that the action of IL-1 is principally manifested in the tissues where it is produced, during infection or inflammation, where it may act to induce fever. Our data also suggest that locally released endogenous IL-1ra, like IL-1, may act locally to induce its effects and may play a key role in the resolution of the fever induced by a localised injection of LPS in the rat air pouch model of inflammation.