| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print September 22, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Research Unit, Laboratory for Clinical and Experimental Pharmacology, University Hospital of Valme, Seville, Spain;
Department of Pharmacology, Paediatrics and Radiology, Faculty of Medicine, University of Seville, Seville, Spain; and
Faculty of Life Sciences, University of Manchester, Manchester, UK
1 Correspondence: Research Unit, Laboratory for Clinical and Experimental Pharmacology, University Hospital of Valme, Avda Bellavista s/n, Seville 41014, Spain. E-mail: jminano{at}us.es
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: cyclophoshamide • ibuprofen • lypopolysaccharide • subfornical organ
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bacterial endotoxins cause much of the pathophysiology associated with gram-negative sepsis [3 ], including hemodynamic disorders and changes in leukocyte counts and fever [4 ]. Much of the pathophysiology is mediated by proinflammatory cytokines synthesized mainly by leukocytes, perivascular microglia, and meningeal macrophage cells. Thus, the febrile response to LPS administered systemically is mediated by endogenous pyrogenic cytokines originating from polymorphonuclear leukocytes (PMN) [5 , 6 ]. Following administration of LPS, the number of circulating granulocytes is decreased in a manner temporally related to the fever response in both normal and granulocytopenic animals [7 ]. Observations such as these have led to the suggestion that endotoxin fever is a central response to endogenous pyrogen(s) released from activated PMN. This hypothesis is challenged, however, by facts such as granulocytopenic patients developing persistent fever and animals made granulocytopenic by immunosuppressant drugs responding with enhanced febrile responses to LPS [9 10 11 ].
A large body of evidence shows PGE2 is a principal downstream mediator of fevers produced by LPS and most pyrogenic cytokines, including IL-1β and IL-6. In the biphasic febrile response to LPS, the first rise in temperature is thought to be due to pyrogenic cytokines, mainly IL-1β, acting via an extravascular component of the circumventricular organs accessible from only the blood side (e.g., the organum vasculosum laminae terminalis, OVLT) or to release arachidonate metabolites, especially PGE2 [12 ]. The second phase is due to the cytokines acting via the blood-brain interface, accessible from both the blood and brain, to release metabolites other than PGE2 [12 ]. However, our understanding of how peripheral inflammatory messengers cross the blood-brain barrier (BBB) to produce fever has changed. The view that the OVLT is the major port of entry for blood-borne cytokines into the brain has lost its once dominant position [13 ]. Instead, it has been proposed, for example, that the vagus nerve transmits pyrogenic cytokine information from the periphery to the brain [14 ]. But this theory has also been superseded. In the endothelial and perivascular microglia of the BBB, upstream signaling molecules, such as proinflammatory cytokines, are switched to the downstream mediator, PGE2 [15 ]. Hence, fevers induced by most pyrogenic cytokines require cyclooxygenase (COX), the key enzyme synthesizing PG from arachidonic acid [16 ]. Consequently, production of PGE2 and activation of hypothalamic PGE2 receptors provides a unifying mechanism for fevers induced by both exogenous and most endogenous pyrogenic cytokines [17 ]. Although PGE2 produced centrally is considered most important in the development of fever [18 , 19 ], circulating peripheral PGE2 [12 , 13 , 20 ], as well as alternative mechanisms of the PGE2-synthesizing cascade (e.g., by transcriptional upregulation of secretory phospholipase A2 or induction of microsomal PGE2 synthase-1) may also contribute [21 ].
In immunosuppressed patients with persistent fever, PMN depletion is thought to be the major, if not the sole mechanism underlying the increased susceptibility to bacterial infection. The absolute neutrophil count in people receiving cytotoxic chemotherapy can drop below 500 cells/mm3, while the incidence of severe infection, the number of days spent on antibiotics, and the number of days of fever increase [22 ]. Such patients have increased serum levels of pyrogenic cytokines, including IL-1β preceding febrile episodes [23 , 24 ]. Although PGE2 was not measured in these patients, the COX inhibitor, indomethacin has been reported to accelerate hematopoietic recovery and have a protective effect during bacterial infection in granulocytopenic mice [25 ]. As indicated earlier, fever persists in immunosuppressed patients and is resistant to COX inhibitors [2 ]. Fever in leukopenic subjects could be mediated by cellular and humoral pathways other than PG-dependent ones, but information is lacking on this topic for hosts receiving cytotoxic chemotherapy.
In light of these observations, we hypothesized that LPS-induced fever in leukopenic subjects requires neither significant numbers of circulating leukocytes nor increased PGE2 concentrations in the blood. The present study investigates the role of circulating leukocytes and PGE2 in the pathogenesis of fever using LPS-induced fever in rats made granulocytopenic with cyclophosphamide (CP). This is an established animal model of the hematopoietic depletion that occurs in patients receiving CP, an alkylating agent with cytotoxic and immunosuppressive activities useful in the treatment of malignancies and transplant rejection [26 ]. We conducted time course studies of fever, plasma levels of PGE2, and IL-1β, as well as expression of COX-2 following intraperitoneal (i.p.) administration of LPS in leukopenic and normal rats. We use an i.p. dose of 50 µg LPS that may be pathophysiologically relevant because it mimics most of the clinical features of endotoxemia in normal and immunosuppressed rats [10 , 11 ]. The link between LPS and PGE2, in the signal transduction pathway leading to fever, was investigated using the COX inhibitor, ibuprofen (IBU).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Animals
Male Wistar rats weighing 200 to 250 g were obtained from Charles River Laboratories. Rats were housed individually in a temperature-controlled room (26±1°C) on a 12 h light/12 h dark diurnal cycle (light on 0700 h) with unrestricted access to food and water. Studies were performed on conscious, unrestrained rats acclimatized to experimental procedures. Each rat was used for only a single experiment. All treatments were carried out at the same time of day during the light period so that the circadian rhythms of the rats were identical across studies. All animal experimental procedures were reviewed and approved by the Institutional Animal Review Committee and were conducted according to the European Communities Council Directive (86/609/EEC).
Measurement of body temperature
Battery-operated biotelemetry devices (model PDT-4000; Mini-Mitter Co., Inc., Sunriver, OR) were inserted into the peritoneal cavities of rats, under general anesthesia induced by ketamine (100 mg/kg) and xylazine (4 mg/kg), as described elsewhere [11
]. After implantation, rats were allowed to recover for at least five days before experimental use. Deep body temperature (Tb) and locomotor activity (data not shown) of animals were recorded at 5-min intervals beginning at least 24 h before injection of LPS and continued for at least 3 days after injections. For analysis and clarity of graphical presentation, temperature data of 15 min time intervals have been used.
Production of granulocytopenia
The rat model of leukopenia used in this study has been described in detail elsewhere [10
, 11
]. Following recovery from surgical implantation of the biometric monitor, animals were injected i.p. with CP at doses that produced maximum granulocytopenia four days after the initial injection of 150 mg/kg. This was followed 72 h later by a second injection of CP 50 mg/kg. No animals died following the second CP administration. The ability of granulocytopenic rats to develop fever in response to LPS 50 µg/kg i.p. was measured 24 h after the second CP injection when leukopenia was maximal (Fig. 1
). To avoid development of tolerance to bacterial endotoxin [27
], each animal was injected with LPS only once. Separate groups of normal age-matched, nongranulocytopenic rats with similar weights, and treated i.p. with the same dose of LPS, were used for comparison in each experiment. Control groups of normal and leukopenic rats received an equivalent volume of vehicle (PFS, 1 mL/kg i.p.).
|
In another set of experiments, the effects of IBU, administered 3 h after i.p. injection of LPS, were determined. Further new groups of normal and leukopenic rats received similar treatment with LPS (n=8 per treatment group), followed 3 h later by IBU (at 1100 h) before the maximum peak of LPS fever. This interval between LPS and IBU administrations allowed sufficient time for any induction of COX-2 transcription, development of the highest fever peak, and the maximal rise in PGE2 after LPS injection [33 ].
Blood and CSF sampling
Blood and CSF samples were collected as terminal procedures under general anesthesia induced by a mixture of ketamine and xylazine (same dose as above) at times indicated in Results. Peripheral blood (5–8 mL) was aspirated by heart puncture from the right ventricle with a needle and a syringe [11
]. Whole blood for hematological parameters was collected into sterile tubes containing potassium EDTA and counted by an automatic hematological analyzer equipped with veterinary software (Cell Dyne 3500 Abbot, Allentown, PA). Counts of total peripheral leukocytes, lymphocytes, monocytes, and neutrophils were determined for blood collected before and at 1, 2, 4, 8, 12, 24, and 48 h after LPS, in both normal and leukopenic rats (n=10 rats/group/time point).
The effects of LPS, as well as IBU on plasma IL-1β and PGE2, as well as CSF PGE2 levels were determined in further groups of normal and leukopenic rats (n=8 per treatment group) (see Results). On the day of the experiment, each rat received an i.p. injection of either IBU (10 mg/kg), LPS (50 µg/kg i.p.) or drug vehicle. For plasma IL-1β and PGE2 determinations, blood was placed in polypropylene test tubes treated with 50 µl of solution containing 9 mg EDTA and 0.057 mg sodium carbonate, and stored on ice for not more than 5 min before centrifugation (1,600 g at 8°C for 12 min). Plasma samples from each rat were then divided in aliquots of 500 µl and stored at –20°C until determination of PGE2 and IL-1β concentration immediately before and 3, 6, 12 (before the dark period), and 24 h (before the light period) after LPS injection (n=12 rats/group/time point). Plasma IL-1β and PGE2 levels were assayed with commercially available sandwich ELISA kits (R&D). Repeated freeze-thaw cycles for reagents and samples were avoided. This assay recognizes both recombinant and natural rat IL-1β. This immunoassay was calibrated against a highly purified E. coli-expressed, recombinant rat IL-1β produced at R&D Systems. This recombinant rat IL-1β contains 153 amino acid residues and has a predicted molecular mass of 17 kDa. On the basis of total amino acid analysis, the absorbance of a 1 mg/mL solution of E. coli-expressed recombinant rat IL-1β at 280 nm was determined to be 0.64 AU. No significant cross-reactivity or interference was observed for IL-1β. The lower limits of detection were typically <5 pg/mL for IL-1β and <8.0 pg/mL for PGE2 (see below).
CSF was collected from normal (n=6 rats/group/time point) and leukopenic (n=12 rats/group/time point) rats anesthetized as described above. Each rat was placed in a stereotaxic apparatus, and the neck flexed so that a 29-gauge needle could be lowered between the base of the skull and the first cervical vertebra, into the cisterna magna. Needle placement was verified by drawing a small amount of CSF into the needle and observing the clear CSF in the Tygon tubing connecting the needle to a 50-µl syringe (Hamilton). CSF (40-60 µl) was collected in polypropylene test tubes that were put immediately on dry ice and then stored at –80°C until determination of PGE2 concentration. After dilution (1:10) of the CSF samples, PGE2 was measured with an overnight high-sensitivity immunoassay (PGE2, High-Sensitivity Immunoassay Kit, R&D Systems) following the procedures detailed in the manufacturer’s instructions. The sensitivity of this ELISA kit was <8.0 pg/mL. The intra-assay and interassay coefficients of variation were 2.6 and 2.2%, respectively. No interference was observed. Cross-reactivities of the PGE2 samples were <18%, <12%, and <8% for PGE3, PGE1, and PGF series, respectively. Cross-reactivity for PGA2, PGB1, arachidonic acid, and thromboxane B2 was <0.1%.
COX-2 immunohistochemistry
Under general anesthesia induced by ketamine and xylazine, rats were rapidly perfused transcardially with 0.9% saline, followed by 4% paraformaldehyde in 0.1 M PBS (pH 7.4) at 3, 6, and 24 h after LPS injection. Control animals were injected i.p. with PFS and killed 3 h later. Brains were removed, immersed in the same fixative for 4 h at 4°C, then cryoprotected by immersion in 0.1 M PBS containing 30% sucrose at 4°C for 48 h. Coronal sections (20 µm) of brain regions containing the SFO implicated in fever [34
] were cut.
Immunocytochemistry was performed as described previously for COX-2 in the rat brain using an immunoperoxidase detection system (Vectastain ABC Kit, Vector Laboratories), in free-floating sections at room temperature [33 , 35 ]. Briefly, sections were preincubated in 3% normal rabbit serum (Vector) with 0.3% Triton X-100 and subsequently incubated with the primary antibody, goat anti-murine COX-2 (dilution 1:5000; Santa Cruz Biotechnology) overnight. Endogenous peroxidase was blocked using an avidin-biotin blocking kit according to the manufacturer’s instructions. After several washes with PBS, sections were incubated with biotinylated rabbit anti-goat IgG (dilution 1:200) in secondary antibody dilution buffer (0.01M PBS with 0. 3% Triton X-100, pH 7.4) for 50 min. After a rinse, sections were incubated with peroxidase-linked ABC (Vector) for 50 min. COX-2-ir was visualized with diaminobenzidine (DAB; DAKO, Carpinteria, CA). The reaction was terminated after 5 min with two successive rinses in PBS. Tissue sections were mounted on gelatin-coated slides, air dried, counterstained with cresyl violet, dehydrated in alcohol, cleared in xylene, and coverslipped with Permount. The COX-2 primary antibody had been previously characterized by preadsorption and Western blotting experiments in rat brain [36 ]. Control experiments included coincubation of the COX-2 primary antibody with the COX-2 antigen peptide (1µg/mL); negative controls omitted the primary antibody. In both cases, no positive immunostaining was detected. Brain sections of normal and leukopenic rats were postfixed on glass slides under the same conditions, in order to avoid the possibility of varied COX-2 staining. All antibodies were diluted with 3% normal rabbit serum in 0.1 M PBS. Photomicrographs were produced using a digital camera (Olympus C3030 Zoom) mounted on the microscope (Olympus BX41). Adobe Photoshop was used to combine photomicrographs into plates. Only the sharpness, contrast, and brightness have been adjusted.
Statistical analysis
All data are reported as means ± SEM. In graphs of thermal responses to drugs, the mean changes in absolute values of Tb (°C) for each rat are plotted against time (h). ANOVA followed by Student-Newman-Keuls multiple comparisons test was used to analyze statistical differences. Animals with consecutive missing temperature values recordings, due to failure of the telemetry system, were excluded from statistical analysis. Concentrations of IL-1β and PGE2 were compared by ANOVA. Calculations and graph plots were carried out on a personal computer using SigmaPlot and SigmaStat software (SPSS Science, Chicago, IL). Two tailed P values <0.05 were considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
60 cells/mL, P<0.05), but lower numbers of lymphocytes (8±1 cells/mL) and monocytes (10±1 cells/mL) also contributed to the decreased total cell count (P<0.05 CP-treated vs. normal counts). Some toxic effects of immunosuppressive chemotherapy were noted; leukopenic rats appeared less active and lost a little body weight (15±3 g), but they were not significantly different from normal rats at the time of LPS treatment. The effects of LPS (50 µg/kg, i.p.) on the total and differential white blood cell counts for normal and CP-treated rats are shown in Fig. 1 . All normal rats displayed a transient leukopenia 1–4 h after receiving LPS (Fig. 1A , P<0.001), due mainly to a fall in the number of lymphocytes (Fig. 1B , P<0.001). This leukopenia was followed by leukocytosis attributable to a marked recovery in neutrophil counts to levels exceeding normal from 8 to 24 h (P<0.001, Fig. 1C ). At 48 h, total peripheral leukocyte, lymphocyte, monocyte, and neutrophil counts were similar to pre-LPS values (time 0 h, P>0.05, Fig. 1 ). In contrast to normal rats, the same dose of LPS had no effect (P>0.05) on leukocyte populations in immunocompromised rats with cell counts below the normal range throughout the experimental period (Fig. 1) .
Effect of LPS on body temperature in normal and leukopenic rats
Core body temperature (Tb) responses of normal and immunosuppressed rats to LPS (50 µg/kg, i.p.) are shown in Fig. 2
. Before LPS or control vehicle (PFS) injections, all groups had virtually identical baseline Tb (37.2±0.1°C; P>0.05). Injection of PFS in normal and immunocompromised rats resulted in small, nonsignificant changes from preinjection Tb values throughout the whole experimental period. Under the experimental conditions used in this study, both normal and leukopenic rats receiving PFS showed a normal circadian rhythm in Tb with low daytime and high nighttime values with no significant differences between groups (Fig. 2)
.
|
150 and 400 min after injection (Fig. 2)
. LPS induced elevations in Tb that began 15 min after injection in both normal and leukopenic rats. During the daytime period following injection of LPS (0800–1900), both groups of animals exhibited a biphasic rise in Tb, with the first peak occurring within 3 h. This peak was followed by a second hyperthermic phase, which peaked 6–8 h after injection of LPS, at 1500. During both phases, increases in Tb in LPS-treated rats were significantly greater than those in time-matched controls injected with PFS (P<0.05; Fig. 2
). In normal rats, fever decreased during the nighttime period. In marked contrast, Tb remained dramatically elevated in leukopenic rats during this period (P<0.001 LPS vs. PFS, Fig. 2 ). The enhanced fever evoked by LPS in leukopenic rats was significantly greater than the rise in Tb observed in normal animals (P<0.001, normal vs. leukopenic rats, Fig. 2 ) throughout the experimental period. By 96 h, Tb in LPS-treated leukopenic rats returned to, or near to preinjection values (37.6±0.2°C) (data not shown).
Effects of LPS on plasma IL-1β and PGE2 in normal and leukopenic rats
Plasma concentrations of IL-1β and PGE2 were assayed in both normal and leukopenic rats during the febrile response to LPS (Fig. 3
). There were no significant differences between the groups of PFS-treated rats used at the different sampling times (3, 6, 12, and 24 h); basal plasma IL-1β levels in normal rats were 19.9 ± 8.4 pg/mL and 7.7 ± 3.8 pg/mL in immunosuppressed animals (Fig. 3)
. Both normal and leukopenic rats showed peak elevations in plasma IL-1β 3–6 h after injection of LPS (50 µg/kg, i.p.) when compared with respective PFS controls (P<0.001). It is important to note that 6 h after LPS injection, when elevations in Tb and IL-1β were highest in normal, i.e., nongranulocytopenic rats, the severe depletion of leukocytes in CP-treated rats (Fig. 1)
was associated with a marked reduction in plasma IL-1β levels (normal rats: 491.2 ± 76.8 pg/mL; leukopenic rats: 152.9 ± 12.5 pg/mL; P < 0.001, Fig. 3A
). LPS-induced increases in plasma IL-1β decreased to control levels by 24 h after injection in both normal and leukopenic rats (Fig. 3A)
.
|
Effect of IBU on LPS-induced fever in normal and leukopenic rats
IBU (10 mg/kg, i.p.) was administered either 1 h before (Fig. 4C
4D
) or 3 h after LPS 50 µg/kg (Fig. 4E
4F)
in different groups of normal and leukopenic rats. No significant changes in Tb were observed after IBU or its vehicle (PFS, equivalent volume 1 mL/kg) in either normal or leukopenic rats not receiving LPS (Fig. 4A
4B)
.
|
The same dose of IBU (10 mg/kg i.p.) given 3 h after LPS, that is immediately before the second phase of LPS-induced fever, abolished the febrile response to LPS in normal rats (P<0.001, Fig. 4E ). In leukopenic rats, however, IBU treatment after LPS caused a more pronounced rise in Tb during the second phase of fever, with Tb elevated throughout the night-time phase (P<0.001, Fig. 4F ).
Effects of IBU on plasma IL-1β and PGE2 during LPS fever in normal and leukopenic rats
Pretreatment with IBU (10 mg/kg, i.p.) 1 h before LPS or vehicle (PFS) in normal and leukopenic rats did not affect basal plasma concentrations of either IL-1β or PGE2 at any sample time (P>0.05, Fig. 5
). Pretreatment with IBU enhanced LPS-evoked increases in IL-1β concentrations at 3 h in both normal (3-fold) and leukopenic rats (1.8-fold) (P<0.001 vs. PFS+LPS, Fig. 5A
). Plasma levels of IL-1β in LPS-injected normal and leukopenic rats remained higher than their respective controls (IBU+PFS) at 6 h (P<0.001, Fig. 5
). By 24 h, plasma IL-1β levels in IBU+LPS treated rats in both normal and leukopenic groups did not differ from their respective controls (P>0.05 vs. IBU+PFS, Fig. 5
).
|
85% (Fig. 5B
; P<0.001 IBU+LPS vs. PFS+LPS). PGE2 levels returned to pretreatment values 6 h after LPS injection (Fig. 5B)
. In leukopenic rats, IBU 1 h before LPS had no significant effect on plasma PGE2 concentrations (Fig. 5B)
.
Effects of LPS and IBU on CSF PGE2 in normal and leukopenic rats
Basal PGE2 levels in CSF from normal and leukopenic rats were 243 ± 86 and 324 ± 102 pg/mL, respectively (Fig. 5C)
. The CSF PGE2 level in normal rats treated with IBU (10 mg/kg, i.p.) was similar to control. Intraperitoneal injection of LPS (50 µg/kg) resulted in a significant (P<0.05) and steady increase in CSF PGE2 levels for 0–6 h after LPS administration in both normal and leukopenic rats. LPS administration resulted in a twofold increase in the concentration of PGE2 at 6 h after the injections (Fig. 5C
, LPS/vehicle vs. PFS/vehicle). The CSF PGE2 levels of normal rats treated with LPS and IBU were similar to control. However, IBU administration in LPS-treated rats did not block the LPS-induced increase in CSF PGE2 level in leukopenic rats (Fig. 5C
, LPS/IBU vs. LPS/vehicle).
LPS-induced fever and brain COX-2 expression
It has been shown previously that systemic LPS induces COX-2-ir in rat brain barrier cells [37
, 38
]. To examine whether LPS-induction of COX-2 is altered during granulocytopenia, immunohistochemical staining of this isoform was applied in this study. Intraperitoneal injection of LPS, but not PFS, induced expression of COX-2 in brain blood vessels (Fig. 6
). COX-2 induction occurred in blood vessels throughout the brain but was more prominent in circumventricular structures such as the subfornical organ (SFO). COX-2-positive structures were oval or round, and located in the luminal side of vessels, corresponding to perivascular microglia. Figure 6
shows LPS-induced COX-2-positive cells in normal (Fig. 6C
6E
6G
6I)
and leukopenic rats (Fig. 6D
6F
6H
6J)
in the SFO at 3, 6, and 24 h following LPS injection. Normal rats showed a number of intensely stained COX-2-positive structures along the blood vessel wall, whereas COX-2 staining in granulocytopenic animals was generally strong in terms of both the number of COX-2-positive cells and the intensity of staining.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
PMN are an important early source of endogenous pyrogenic and antipyretic factors [14 , 39 ], but what role PMN have in the pathogenesis of fever is unclear. Herion et al. [7 , 8 ] used nitrogen mustard to induce neutropenia and were the first to report that leukocytic pyrogen is a necessary intermediate in LPS fever; a normal febrile response to LPS in leukopenic animals was related to the persistence of granulocytes [8 ]. CP is often prescribed in the chemotherapy of cancer and induces a profound leukopenia with neutrophils being the most susceptible granulocyte to suppression [40 , 41 ]. When patients treated with CP suffer bacterial infections, they develop high fevers (febrile neutropenia). We have reported previously that rats develop a profound leukopenia of <100 cells /mL lasting 6–7 days in response to CP but still develop high fevers in response to LPS [10 , 11 ]. In this rat model of granulocytopenia, there is no significant change in Tb before LPS treatment, but a biphasic febrile response to LPS occurs, which is significantly greater in magnitude and duration than the LPS fever in normal animals. Observations presented here support and extend our previous studies and confirm that agranulocytic subjects can develop fever that is both greater than normal and resistant to antipyretic therapy.
It is now widely accepted that LPS fevers are generally polyphasic. Depending on several methodological factors (including ambient temperature, route, and dose of LPS), the febrile response evoked by LPS can be monophasic, biphasic, or polyphasic [14 , 15 , 19 , 42 , 43 ]. As expected [10 , 11 ], and possibly because of methodological factors, the first phase of LPS fever is missing in rats used in this study. Immunocompetent and leukopenic rats responded to LPS 50 µg/kg i.p. with a biphasic fever, corresponding to the second and third febrile phases of the typical polyphasic febrile response to LPS administered i.v. [15 , 43 ].
It has been shown that all phases of the typical polyphasic febrile response to i.v. LPS are dependent on the Toll-like receptor 4 (TLR-4), at least in rats and mice [44 ]. The first phase is triggered via the TLR4 on hematopoietic cells. The second and third phases involve TLR4 signaling in both hematopoietic and nonhematopoietic cells. The peripheral production of pyrogenic cytokines at the times corresponding to the second and third febrile phases (2–6 h after LPS) has been attributed mostly to hematopoietic cells (e.g., macrophages), whereas both nonhematopoietic cells (e.g., endotheliocytes, microglia, astroglia) and hematopoietic cells (e.g., perivascular and meningeal macrophages) were found to be responsible for cytokine production within the brain [45 ]. In the same time period, PGE2 synthesis is activated both in the periphery and the brain [46 ]. Thus the present data indicate that nonhematopoietic cells other than those depleted by CP in the blood play an indispensable role in the febrile response induced by LPS fever in leukopenic rats. Leukocytes, however, do appear to be involved in the limitation and reduction of fever, since in their absence the febrile response to peripheral LPS is greater. These data also suggest the possibility that there are several mediators or multiprocesses underlying the pathogenesis of fever.
Tissue resident macrophages are required for host defense against acute bacterial infection, and they may have a role in the enhanced fever seen in immunocompromised hosts. For example, production of pyrogenic cytokines by Kupffer cells, rather than by circulating cells, could explain the febrile neutropenia syndrome. Hence, guinea pigs made leukopenic with vinblastine, and exhibiting hyperpyrexia, have Kupffer cells exhibiting increased uptake of LPS [9 ]. However, vinblastine differs from other cytotoxic drugs, including CP, by inducing nonspecific leukopenia without affecting Kupffer cells [41 , 47 ]. Overall, indications are that a variety of cells types, including fibroblasts, endothelial cells, and glial cells, are involved in the acute inflammatory response, and that one or more of these could be implicated in the enhanced LPS fever in CP-treated rats. CP not only induces a selective depletion of lymphoid tissue but also abolishes the production of antipyretic mediators such as vasopressin [48 ] and the antiinflammatory cytokine IL-10 [11 ] required for the defervescence of LPS fever. As indicated earlier, PMN appear to be implicated in the limitation and reduction of fever in vivo, since in their absence, in CP-treated rats, the febrile response to a given dose of LPS is enhanced.
Enhanced fever in response to LPS in leukopenic animals has been observed previously in vinblastine [9 ] and CP-treated [10 , 11 ] animals. Proinflammatory and anti-inflammatory cytokines were present in the serum of granulocytopenic hosts after endotoxin injection, indicating tissues other than leukocytes are implicated in the febrile response to systemic LPS. How leukopenia causes an intensified rather than a reduced fever with decreased production of circulating inflammatory mediators is undetermined. Granulocytes incubated in vitro inactivate endotoxin [49 ]. If granulocytes are important in the in vivo inactivation of LPS, their absence may enhance the biological effects of bacterial endotoxins in leukopenic hosts.
Present findings show COX activation is involved in the production of fever in normal (i.e., non-granulocytopenic) rats after i.p. LPS administration. Inhibition of COX by IBU 10 mg/kg i.p. almost completely prevented the febrile response (Fig. 4C 4E) . IBU is a reversible, short-acting inhibitor of COX, which is not selective for the known COX isoforms. Peripheral administration of anti-inflammatory, antipyretic doses of IBU is used widely in most infectious diseases causing inflammation and/or fever. A single dose of IBU decreases PGE2 levels in blood, as demonstrated in this study, and inhibits PGE2 synthesis in tissues, most relevantly here, the brain, including the hypothalamus [50 , 51 ]. In this study, IBU also inhibited LPS-induced expression of COX-2 in brain microglial and endothelial cells, which is thought to play a critical role in the pathogenesis of fever [52 ]. In marked contrast with observations in normal rats, IBU pretreatment had only a small antipyretic effect on the early febrile response to LPS fever in leukopenic rats (Fig. 4D) , while IBU given after LPS resulted in further enhancement of the later second phase of LPS fever (Fig. 4F) . Also, in contrast with normal animals, plasma levels of PGE2 in granulocytopenic rats, which were not significantly raised during LPS fever, were unaffected by pretreatment with IBU. The same LPS challenge (50 µg/kg i.p.) was used in both normal and leukopenic rats, and yet there was this pronounced difference in circulating PGE2 between the two groups. Furthermore, although plasma PGE2 levels were unchanged after LPS injection in leukopenic rats, CSF PGE2 levels began to rise by 3 h and were increased twofold compared with controls by 6 h (see Fig. 3C ). Clearly, there is an association between the increase in CSF PGE2 levels and the enhanced febrile response but no association between plasma PGE2 and fever in CP-treated rats. The increase in CSF PGE2 probably reflects the induction of COX-2 in the brain capillary endothelial cells, where LPS or cytokines enhance COX-2 mRNA and protein levels, which is compatible with the development of LPS-induced fever [37 ].
The increase in CSF PGE2 induced by LPS in leukopenic rats was not reduced by a dose of IBU (10 mg/kg i.p.), which did lower both CSF PGE2 and fever in normal rats. This apparent lack of effect by IBU on COX in CP-treated animals is indicated by the assay of PGE2 in CSF sampled from the cisterna magna. It is possible that 1) cisternal CSF PGE2 levels may not reflect PGE2 concentrations at the neuronal site of fever action and/or 2) PGE2 does not accurately reflect COX activity in CP-treated rats. Although PGE2 is a major COX product and mediator of fever in normal immunocompetent hosts, other arachidonic acid metabolites synthesized by COX-2 (not measured in this study) could contribute to the febrile response to LPS and be of increased importance in immunocompromised hosts. After systemic LPS administration, plasma 6-keto-PGF1 and thromboxane B2 have been shown to be increased after selective inhibition of COX-2 [53 ]. Substances other than arachidonic acid metabolites, for instance cytokines, such as IL-6 [54 ], not measured in this study and linked to downstream mediators other than arachidonic acid derivatives, may also be involved. Finally, IBU 10 mg/kg is a near threshold dose for an inhibition of LPS fever in normal rats. The apparent resistance of leukopenic rats to the central COX inhibition and antipyresis normally produced by this dose of IBU may represent a decrease in drug sensitivity, which may be addressed by an increase in dose.
Blood-borne PGE2 has been proposed as an endogenous mediator of fever in response to peripheral LPS [12 , 15 , 20 ]. The absence of PGE2 in the circulation of CP-treated rats could conceivably explain the failure of antipyretic drugs acting as peripheral COX inhibitors to abrogate fever, but if peripheral PGE2 is an important endogenous pyrogen then, its absence would be associated, we propose, with reduced fever rather than an enhanced fever. These data support the conclusion that LPS fever in CP-treated rats is not mediated by blood-borne PGE2 acting as an endogenous pyrogen. This is consistent with the observation that peripheral PGE2 infused into the internal carotid arterial blood supply to the brain does not increase Tb in conscious rabbits [55 ].
Results presented here indicate that increased expression of COX-2 in the brain may be more important than peripheral in the febrile response to LPS in CP-treated rats. If so, the lack of antipyretic effect by IBU in leukopenic animals could be explained theoretically by 1) IBU not crossing the blood-brain barrier and/or 2) PGE2 not being produced in the brain of CP-treated rats in sufficient quantity to affect body temperature. However, we consider both these hypotheses unlikely. Previous immunohistochemical studies have demonstrated constitutive expression of COX-2 protein in neurons in circumventricular thermoregulatory brain areas involved in fever [33 34 35 ]. Hence, it is possible that some components of the CNS response to LPS may be mediated by constitutive neuronal COX-2, rather than by newly synthesized enzyme in perivascular cells. COX-2-ir was detected in perivascular cells of the SFO at all sampling times following LPS administration. IBU prevented LPS induction of COX-2 expression in both normal and granulocytopenic animals but did not prevent LPS fever in granulocytopenic rats. Although the amount of immunoreactive protein was not determined in this study, observations presented indicate that the relationship between central COX-2 expression and raised Tb following i.p. injection of LPS is not simple.
Other hypotheses that might explain the lack of antipyretic effect of IBU in leukopenic rats include the existence of COX-independent pathways [31 , 56 57 58 ]. Whatever the basis for the reduced antipyretic effect by IBU in CP-treated rats, it remains necessary to ascertain the role of PGE2 in the pathogenesis of fever [59 ]. Our observations in rats made leukopenic by CP suggest that the mechanisms underlying the manifestation of fever depend not only on the qualitative and quantitive nature of the inflammatory/pyrogenic stimulus but also on the host’s immune status.
Fever involves not only the increased synthesis and release of PGE2 in response to endogenous pyrogens e.g., IL-1β, but also that of endogenous antipyretic substances, which limit fever [5 ]. PGs are inhibitory regulators of IL-1β production [17 ]; consequently, IBU-like antipyretic drugs, presumably acting as COX inhibitors, can increase cytokine production. Increased plasma IL-1β was evident in both normal and leukopenic rats in this study 3 h after LPS following pretreatment with IBU (Fig. 5A) . During hemopoietic recovery after intensive chemotherapy, this effect would paradoxically oppose the desired antipyretic action of COX inhibitors.
IL-1β was assayed in this study as a cytokine implicated in many fevers as an endogenous pyrogenic mediator. Plasma IL-1β levels in leukopenic rats were low compared with those in normal rats and did not correlate with the late (>3 h) enhanced fever induced by LPS in CP-treated animals. This suggests that a pyrogenic cytokine other than IL-1β may mediate the potentiated late fever phase. Regardless of the identity of the actual cytokine(s)-mediating fever, a role for COX products e.g., PGE2 in the down-regulation of immune function is indicated by this study. In immunocompromised rats, with the normal balance between the many cells and factors orchestrating the immune response modified by CP, the enhanced fever resistant to IBU may be attributable to disinhibition of cytokine production from PGE2 down-regulation.
Although assay of PGE2 did not show statistically significant changes in plasma from IBU-treated leukopenic rats, this does not mean that important pathophysiological changes in COX and PGE2 activities did not occur at the cell or tissue level. Production of cytokines at the tissue level, rather than in the circulation, has been suggested as an alternative pathway for induction of fever [60 ]. Tissue-resident macrophages are required, in addition to peripheral blood leukocytes, for host defense against acute bacterial infection, and they also produce cytokines and PGE2. In addition, the immunomodulatory effect of PGE2 is complex, being dependent, for example, on both the composition of the cell population and the concentration of PGE2 [61 ]. Caution is needed, therefore, when interpreting data for PGE2 and IL-1β in plasma.
PGE2 appears to have a role in regulating the production of proinflammatory cytokines and antagonists during hemopoietic recovery after intensive chemotherapy [62 , 63 ]. The results of the present study are consistent with PGs down-regulating IL-1β levels in vivo at least during the first phase of fever in response to LPS. At the time of LPS injection, lower levels of IL-1β were found in the circulation of leukopenic rats than in normal animals. Low plasma levels of this early proinflammatory cytokine are consistent with the decreased number of circulating PMN in CP-treated rats. LPS increased plasma IL-1β in both normal and leukopenic rats, and these increases were in turn augmented (at 3 h, Fig. 5A ) by pretreatment with IBU. This apparent up-regulation of IL-1β in the presence of the COX inhibitor IBU is consistent with a COX product, possibly local PGE2, having an inhibitory action on cytokine expression in vivo, in addition to PGE2 acting as an endogenous pyrogen. Overall, the present results are consistent with the hypothesis that the endothelial and perivascular cells of the brain vasculature act as an interface between blood and brain by producing secondary mediators, such as PGs, which, in turn, act on neurons involved in fever.
In summary, this study addresses an important clinical problem, namely antipyretic resistant fever in immunocompromised patients, by using a rat model of the pathophysiological events that occur in humans with bacterial infection after immunosuppressive chemotherapy. Results indicate that IBU, a common antipyretic drug acting by inhibition of COX production of PGE2, not only failed to reduce the enhanced fever in granulocytopenic rats, but increased fever further when administered therapeutically, i.e., after the onset of the febrile response to LPS. This drug enhancement of fever was associated with an apparent up-regulation of cytokine production, evident in this in vivo study during the early, but not the late phase, of fever as increased plasma IL-1β. These observations are consistent with cytokine production ordinarily being subject to inhibitory control by a product of COX activity. Induction of COX-2 in the brain appears to contribute to the pathogenesis of fever in granulocytopenic rats. These observations could help explain the failure of antipyretic drugs to abrogate the persistent fever in immunosuppressed patients with infections. Thus, although eicosanoid inhibitors are widely used in clinical medicine because of their effective antifebrile and anti-inflammatory activities, further studies are necessary to determine whether COX-2 inhibition can be beneficial in attenuating the fever associated with chemotherapy in humans.
|
ACKNOWLEDGEMENTS |
|---|
Received January 27, 2006; revised July 17, 2006; accepted August 2, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
