Circulating Inflammatory Mediators during Start of Fever in Differential Diagnosis of Gram-Negative and Gram-Positive Infections in Leukopenic Rats

Gram-negative and gram-positive infections have been consideredthe most important causes of morbidity and mortality in patientswith leukopenia following chemotherapy. However, discriminationbetween bacterial infections and harmless fever episodes isdifficult. Because classical inflammatory signs of infectionare often absent and fever is frequently the only sign of infection,the aim of this study was to assess the significance of seruminterleukin-6 (IL-6), IL-10, macrophage inflammatory protein-2(MIP-2), procalcitonin (PCT), and C-reactive protein (CRP) patternsin identifying bacterial infections during start of fever innormal and cyclophosphamide-treated (leukopenic) rats followingan injection of lipopolysaccharide (LPS) or muramyl dipeptide(MDP) as a model for gram-negative and gram-positive bacterialinfections. We found that, compared to normal rats, immunosuppressedanimals exhibited significantly higher fevers and lesser productionof all mediators, except IL-6, after toxin challenge. Moreover,compared to rats that received MDP, both groups of animals thatreceived an equivalent dose of LPS showed significantly higherfevers and greater increase in serum cytokine levels. Furthermore,in contrast to those in immunocompetent rats, serum levels ofIL-6 and MIP-2 were not significantly changed in leukopenicanimals after MDP injection. Other serum markers such as PCTand CRP failed to discriminate between bacterial stimuli inboth groups of animals. These results suggest that the use ofthe analyzed serum markers at an early stage of fever couldgive useful information for the clinician for excluding gram-negativefrom gram-positive infections.

INTRODUCTION

Top

Introduction

Bloodstream infection is an important cause of morbidity andmortality, especially in febrile patients undergoing intensivechemotherapy for the treatment of cancer and other illnesses(10). This susceptibility appears to be directly related tothe severity and length of leukopenia (28). Most infectionsin immunosuppressed patients are due to gram-negative and gram-positivebacteria (21), and fever is frequently the only sign of infection(7, 34). Thus, because in the majority of these patients withfever episodes the causative infectious agents cannot be identifieddue to the nonspecific nature of signs and symptoms (16), thesepatients are treated empirically with broad-spectrum antibioticsas soon as fever develops (6, 28). Therefore, because thereare increasing numbers of patients who are immunodeficient followingchemotherapy and because microbes are becoming antibiotic resistant,the use of inflammatory markers for early detection of bacterialinfections could be essential to reflect the nature of the infectionand distinguish the fever episodes in immunosuppressed patientsin order to guide the choice of specific antibiotic therapies,even before culture results are available (27).

The systemic host responses to gram-negative and gram-positivebacterial infections involve similar clinical symptoms, includingleukocytosis and fever (19, 44). Leukocytes, mainly neutrophils,are essential for resistance to bacterial infections, as theyare the main source of proinflammatory mediators (25, 32). Thesecells express specific receptors for immunogenic bacterial cellwall constituents released during infection (1, 37).

Lipopolysaccharide (LPS) is believed to be the most importantmarker for gram-negative bacteria, whereas peptidoglycan by-productssuch as muramyl dipeptides (MDP) serve as salient stimuli fromgram-positive bacteria (9, 19, 20, 44). Although gram-negativebacterial cell walls also contain peptidoglycan, its concentrationis far greater in the walls of gram-positive bacteria (20, 44).Muramyl peptides are key elements of the immune response togram-positive bacteria, which lack LPS. Presently, much workis focused on LPS effects, but many immunocompromised patientsdie of gram-positive infections (10). Immune cells activatedby binding of these substances release many inflammatory mediators,including eicosanoids, cytokines, and chemokines, that activateother immune cells and can also serve to signal the brain.

In humans and experimental animals, MDP and LPS cause an inductionof the acute-phase response, which is characterized by feverand the production of cytokines in vitro and in vivo (9, 20,30), with cells and animals being more sensitive to LPS thanto MDP (42). These studies indicate that the patterns of productionof endogenous mediators involved in the primary nonspecificacute host response may be dependent upon the specificity ofthe microbial pathogens and the host recognition pathways invoked.Like for LPS, it is assumed that MDP produces fever via cytokinesand prostaglandin E₂ induction (20, 44). The mechanisms underlyingthe febrile response to LPS and MDP in leukopenic animals werenot elucidated in this study. This point remains to be resolved,although it is presently under investigation in our laboratory.

Although there are discrepancies regarding the involvement ofspecific receptors in recognition of gram-negative and gram-positivecell wall components (1, 38, 39, 43), LPS and MDP elicit similarfevers and apparently similar sickness behaviors. Much evidencenow indicates that LPS and MDP activate different cell surfacereceptors and that the stimulation of these different receptorscan lead to different pathways and profiles of proinflammatorymediators (13, 17, 18, 29). More recently, evidence has accumulatedto indicate that, in contrast to the case for LPS, the responseto MDP is CD14 independent and is recognized by nucleotide-bindingoligomerization domain 2 but not by Toll-like receptor 2 (TLR2)or TLR4 MD-2 complex (14, 37).

Circulating proinflammatory mediators such as interleukin-6(IL-6), IL-8, macrophage inflammatory protein-2 (MIP-2), C-reactiveprotein (CRP), and, more recently, procalcitonin (PCT) havebeen suggested to be predictive for a systemic microbial bloodstreaminfection in neutropenic hosts (4, 8, 11, 12, 24, 35, 36). However,the early cellular response to gram-negative and gram-positivebacteria is still unclear, and little is known about how theprofile of these mediators during severe leukopenia might beaffected by different microbial stimuli.

The present study was established to compare the induction offever and proinflammatory mediators (IL-6, IL-10, MIP-2, CRP,and PCT) in normal and leukopenic rats following an injectionof LPS or MDP as a model to discriminate between gram-negativeand gram-positive infections. For this purpose, we used an experimentalmodel that resembles the pathophysiologic events that may occurin patients receiving cyclophosphamide, an alkylating agentwith cytotoxic and immunosuppressive activities used in thetreatment of malignancies and as a combatant of transplant rejection.

MATERIALS AND METHODS

Top

Materials and Methods

Animals. Specific-pathogen-free male Wistar rats (Charles River BreedingLabs, Barcelona, Spain), weighing between 200 and 250 g, wereused for the experiments. Rats were kept in a specific-pathogen-freefacility and housed in individual plastic cages. All rats weremaintained during all experiments in a temperature-, humidity-,and light-controlled chamber set at 26 ± 1°C witha 12-h, 12-h light-dark cycle, with light on at 0700 h. Rodentlaboratory chow and drinking water were provided ad libitum.One rat was used for one experiment only and was not used repeatedly.Half of the rats were rendered leukopenic by intraperitoneal(i.p.) injections of cyclophosphamide before bacterial challenge,whereas the nonleukopenic group (normal, immunocompetent rats)did not receive any cyclophosphamide. All rats were acclimatedfor 1 week before the study. Experimental interventions werecarried out with the rats remaining in their home cages. Careand use of the animals were in accordance with protocols approvedby the Animal Ethics Committee of the University of Sevilleand were conducted according to the European Communities CouncilDirective (86/609/EEC).

Drug solutions. Cyclophosphamide, purified lipopolysaccharide from Escherichiacoli (serotype 0111:B4), and MDP (N-acetyl-muramyl-L-alanyl-D-isoglutamine)were obtained from Sigma Chemical Co. (St. Louis, MO). The vehiclefor all solutions was sterile, pyrogen-free 0.9% saline (PFS)(Fresenius Kabi, Barcelona, Spain). The same lot and stock solutionswere used for all experiments. All drug solutions were preparedin pyrogen-free glassware, passed through 0.22-µm-pore-sizeMillipore bacterial filters, warmed to 37°C, and brieflysonicated before i.p. administration. None of the MDP and cyclophosphamidesolutions used in this study induced gelation in the Limulusamebocyte lysate assay, so any contamination with endotoxinwas below the level of 25 pg/ml.

Body temperature measurement. Changes in core body temperature were measured every 5 min usingremote radiobiotelemetry (Vitalview Series 4000 system; MiniMitter, Sunriver, OR) with an accuracy of ±0.1°C.Briefly, anesthetized animals (100 mg/kg ketamine and 5 mg/kgxylazine, both administered by i.p. injection), were implantedi.p. with precalibrated temperature-sensitive radio transmitters(model PDT-4000 E-Mitters; Mini Mitter). Each animal was housedin a separate cage and allowed to recover for 1 week. The regularday-night body temperature rhythm was monitored for anotherweek before the experiment. Animals remained in their home cagesduring the experiment, while the frequencies emitted by thethermistors were monitored remotely. These frequencies werethen converted to temperature (°C) values based on eachtransmitter's calibration data. For the analysis and graphicaldocumentation, temperature data from adequate time intervalsof 15 min were used.

Induction of leukopenia. Chemotherapy-induced profound leukopenia in rats was achievedas described previously (24, 26). In brief, rats were renderedleukopenic by i.p. administration of cyclophosphamide at 150and 50 mg/kg, 3 days and 1 day, respectively, before injectionor any further treatment. This protocol resulted in leukocytecounts of <0.1 x 10⁹/liter on the day of drug injection (day0), mainly due to a very low number of neutrophils. Also, thenumbers of lymphocytes and monocytes are strongly reduced andcontribute to the profound leukopenia observed in this experimentalmodel (see Table 1). Animals were monitored for evidence ofovert illness, activity, and body temperature after the treatments.No animals died following cyclophosphamide administration.

View this table:
[in this window]
[in a new window]
TABLE 1. Effect of cyclophosphamide on total leukocytes and differential cell counts in rats^a

Experimental procedures. All experiments were conducted between 0700 and 1900 h. so thatthe circadian rhythms of the animals were identical across studies.Before the experiment, rats were weighed and arbitrarily assignedto body weight-matched experimental groups. Only animals whosebody temperatures were stable and in the range of 36.9 to 37.3°C,which is within the range usually observed for rats under similarexperimental conditions (23, 24), and which were acclimatedto experimental procedures were used to determine the effectof drug application. All treatments were carried out at thesame time of day, with each rat being used only once for theexperiment. On the day of the experiment, the animals were separatedinto three treatment groups, namely, LPS, MDP, and PFS; eachrat was handled gently so that the body temperature stabilizedfor a period of at least 90 min before any injection. For bodytemperature and hematological studies as well as the quantitativedetermination of inflammatory mediators, separate groups ofnormal and leukopenic rats were used in a time course experiment(n = 10 to 12 rats/group/time point).

Fever induction. Fever was induced by i.p. administration of a single dose ofLPS (100 µg/kg) or MDP (100 µg/kg), as previouslydescribed (23, 30, 33, 41). Control animals received an equivalentvolume of PFS as vehicle (1 ml/kg). The order of PFS and pyrogeninjections was randomized. All LPS and MDP treatments were carriedout 24 h after the second injection of cyclophosphamide, whenleukocyte counts are very low (see Table 1). To avoid the developmentof toxin tolerance, each animal was injected with LPS or MDPonly once. All injections were done at the same time of day(0700 h), so that the circadian rhythms of the animals wereidentical across studies. The changes in core body temperatureswere monitored for 12 h posttreatment. For hematological studiesand the quantitative determination of inflammatory mediators,additional groups of normal and leukopenic rats were used. Bloodsamples were collected from rats anesthetized as described abovevia cardiac puncture (under terminal anesthesia) at selectedtimes after drug injection for later measurement of cell countsand serum mediators.

Hematological parameters. For hematological studies, whole blood for cell counts was collectedinto sterile tubes containing K₃EDTA, and cells were countedby an automatic hematological analyzer equipped with veterinarysoftware (Cell Dyne 3500; Abbot, Allentown, PA), which usesa laser beam and measures the light scattered by the cells togive the total and differential leukocyte counts. Total peripheralleukocytes, lymphocytes, monocytes, and neutrophils counts weredetermined at 1, 2, 4, 8, and 12 h after LPS, MDP, or PFS administrationfor both normal and cyclophosphamide-treated rats.

Inflammatory mediators. For measurement of IL-6, IL-10, MIP-2, CRP, and PCT productionin serum, blood was collected into sterile tubes containingtrisodium citrate and a proteinase inhibitor mixture (125 mmol/mlcitrate, 5000 mmol/ml EDTA, 6000 mmol/ml N-ethylmaleimide, and500 kIU/ml aprotinin). The blood samples were immediately centrifugedat 2,000 x g for 10 min at 4°C and stored at -80°C untilthe day of analysis (within 1 month). Quantitative determinationof rat cytokines in serum was done by enzyme-linked immunosorbentassays, using specific kits for IL-6 (R&D Systems Europe,Abingdon, United Kingdom), IL-10 (IBL, Hamburg, Germany), andMIP-2 (BioSource International, Camarillo, CA). The lower detectionlimits were 10 pg/ml for IL-6, 5 pg/ml for IL-10, and 1 pg/mlfor MIP-2. The serum PCT concentrations were measured usinga chemiluminescent immunoassay kit (LUMItest PCT; B.R.A.H.M.S.Diagnostica GMBH, Berlin, Germany) with a lower detection limitof 0.1 ng/ml. CRP in serum was measured by a particle-enhancedimmunoturbidometric method, using a Cobas Integra 700 (RocheDiagnostic System, Basel, Switzerland). All samples were analyzedin duplicate and were assayed at optimal concentrations, accordingto the manufacturer's instructions.

Data analysis. Data are presented as means ± standard errors of themeans (SEM) and were analyzed using SigmaStat software (SPSSInc., Chicago, IL). In graphs of the thermal and cytokine responses,the mean changes were plotted over time. Integrated areas underthe curve (AUC), from 0 to 12 h posttreatment, were calculatedby trapezoid analysis and were used to compare the differencesin the responses of each rat to pyrogen and PFS injections.Data were analyzed by one-way analysis of variance (ANOVA).In cases where the comparisons using ANOVA were shown to besignificant, a further test using Student-Newman-Keuls posthoc analyses was performed to compare the effect of each treatment.A P value of <0.05 was considered significant.

RESULTS

Top

Results

Induction of peripheral blood leukopenia by cyclophosphamide. As described before, i.p. administration of cyclophosphamideinduced a severe leukopenia in rats. As shown in Table 1, theabsolute leukocyte count (immunocompetent rats, 5,890 ±350 cells/ml) was significantly reduced by cyclophosphamide(to 100 ± 20 cells/ml) during all experimental periods(P < 0.05), indicating a successful leukopenic model. Thiseffect remains for 5 to 7 days (28, 32). Although the reducednumber of blood neutrophils characterizes the present leukopenia(absolute neutrophil count, $≤$ 60 cells/ml; P < 0.05), reducednumbers of lymphocytes (8 ± 1 cells/ml) and monocytes(10 ± 1 cells/ml) also contributed to the decreased totalcell count (P < 0.05). Some characteristic toxic effectsof immunosuppressive chemotherapy were noted; e.g., the leukopenicrats appeared more inactive and may have lost some weight. Figure1 shows the effects of i.p. administration of 100 µg/kgLPS or 100 µg/kg MDP on the total count of white bloodcells for immunocompetent and leukopenic rats. All immunocompetentrats had a transient leukopenia at 1 h after MDP or LPS administration(P < 0.05). This leukopenia was followed by leukocytosisdue to a marked recovery in neutrophils to a level that exceededthe normal level at 4 or 8 h after MDP or LPS administration,respectively (data not shown). At 12 h, total peripheral leukocytecounts were similar to preinjection values (time zero) (P >0.05). However, in contrast to the case for immunocompetentrats, the same doses of either LPS or MDP did not modify theleukocyte populations deteriorated from a normal range due tocyclophosphamide during all experimental periods (P > 0.05).

View larger version (28K):
[in this window]
[in a new window]
FIG. 1. Leukocyte (WBC) counts in peripheral bloodstreams of immunocompetent and leukopenic rats injected i.p. with 1 ml/kg PFS, 100 µg/kg LPS, or 100 µg/kg MDP. Leukopenia was induced by administration of cyclophosphamide 3 days (150 mg/kg) and 24 h (50 mg/kg) before PFS or pyrogen injection. The values presented are the means ± SEM for 10 to 12 rats per group. *, P < 0.05 (significantly different from the control values by one-way analysis of variance).

To determine the influence of peripheral leukocytes on bodytemperature and serum IL-6, IL-10, MIP-2, CRP, and PCT levelsfor distinguishing gram-negative from gram-positive bacterialinfections, we serially measured the body temperature and theserum levels of these mediators in this model of leukopenicrats following i.p. administration of the gram-negative pyrogenLPS or the gram-positive pyrogen MDP.

Effects of LPS and MDP on body temperature. To test for the development of a pyrogenic response to LPS orMDP, normal rats and rats rendered leukopenic with cyclophosphamidewere injected i.p. with 100 µg of either LPS or MDP perkg (Fig. 2). Importantly, the mean body temperature did notdiffer among the experimental groups, and all animals exhibiteda similar basal body temperature that ranged between 36.9 and37.3°C at the time of injection. The body temperature inthe PFS-treated control groups remained stable throughout thestudy and was not different among groups (P > 0.05). Intraperitonealinjection of either LPS or MDP caused fever in both normal andleukopenic rats (Fig. 2A and B). Fifteen minutes after challengewith 100 µg LPS or 100 µg MDP, body temperaturesin immunocompetent rats started to rise, and they reached thepeak level at 6 to 8 h in normal rats (Fig. 2A). The returnto the preinjection level was gradual but was essentially completedby 12 h after challenge (Fig. 2A). In contrast, the fevers inducedby LPS or MDP in leukopenic rats remained elevated up to 12h following the injection and were significantly higher thanthose of the immunocompetent animals throughout (P < 0.05)(Fig. 2B). Thus, although the onset latencies of both groupsof rats were not significantly different, the postfebrile recoveryafter LPS and MDP administration in leukopenic animals was notcompleted after 12 h after pyrogenic stimuli, and there wasa significant enhancement in the fever (Fig. 2C). By 96 h postinjection,body temperature declined to or near to preinjection values(data not shown). As shown in Fig. 2C, injection of LPS or MDPin leukopenic rats resulted in a significant increase in corebody temperature, which was significantly higher than that ofthe immunocompetent group (Fig. 2C). Furthermore, the febrileresponse evoked by LPS was significantly higher than that evokedby an equivalent dose of MDP in both groups of rats (Fig. 2C)(P < 0.05).

View larger version (27K):
[in this window]
[in a new window]
FIG. 2. Effect of LPS and MDP on body temperature in immunocompetent and leukopenic rats. Mean changes of core body temperature ( ${Delta}$ Tc) were determined in separate groups of immunocompetent (A) and leukopenic (B) rats during 12 h after i.p. injection of either PFS ( ${blacktriangledown}$ ), 100 µg/kg LPS (•), or 100 µg/kg MDP ( ${circ}$ ). The mean baselines of absolute body temperatures at time zero for the groups were not significantly different (36.9 to 37.3°C). Integrated thermal responses (C) (AUC) of immunocompetent and leukopenic rats after administration of LPS (black bars), MDP (hatched bars), or PFS (white bars) were calculated for the whole 12-h recording period from the data shown in panels A and B (10 to 12 rats under each condition). Data are expressed as means ± SEM. $§$ and *, statistically significant differences (P < 0.05) between LPS- and MDP-treated rats and normal rats, respectively. ${ddagger}$ , significantly different (P < 0.05) from corresponding values between immunocompetent and leukopenic rats (ANOVA, Student-Newman-Keuls post hoc test).

Effects of LPS and MDP on serum inflammatory mediators. In addition to measurement of core body temperatures, the kineticsof IL-6, IL-10, MIP-2, CRP, and PCT in response to LPS or MDPchallenge were also studied. Significant differences were foundbetween the kinetics of production of these mediators in normaland leukopenic rats, as well as in LPS- and MDP-treated rats.High secretion of IL-6, IL-10, MIP-2, PCT, and CRP was notedin sera of both immunocompetent and leukopenic rats after injectionof a pyrogenic dose of either LPS or MDP (Fig. 3 and 4). Wefound for normal (nonleukopenic) rats that, in parallel withthe body temperature change, the concentrations of IL-6 (Fig.3A), IL-10 (Fig. 3C), MIP-2 (Fig. 3E), and CRP (Fig. 4C) inserum also started to rise at 1 h, reached the peak level at2 to 4 h after LPS or MDP injection, and returned to the preinjectionvalues at 12 h after challenge. However, unlike the case forthese mediators, PCT started to rise at 2 h and reached itspeak at 8 h after LPS or MDP administration (Fig. 4A), comparedto the respective PFS controls. Furthermore, as shown in Fig.5, LPS evoked a significantly higher production of serum mediatorsin both immunocompetent and leukopenic rats compared with theresponses evoked by an equivalent dose of MDP (P < 0.05).

View larger version (30K):
[in this window]
[in a new window]
FIG. 3. Effects of LPS and MDP on serum IL-6, IL-10, and MIP-2 levels in immunocompetent and leukopenic rats. Serum levels of IL-6 (A and B), IL-10 (C and D), and MIP-2 (E and F) in separate groups of immunocompetent and leukopenic rats were determined at 1, 2, 4, 8, and 12 h after i.p. injection of either PFS ( ${blacktriangledown}$ ), 100 µg/kg LPS (•), or 100 µg/kg MDP ( ${circ}$ ). Each experiment involved 10 to 12 rats per time point. At all time points tested after injection of PFS (control), immunocompetent and leukopenic rats showed no significant elevation in serum levels of these cytokines (P > 0.05). Symbols represent means; bars indicate SEM. *, P < 0.05 between LPS- and MDP-injected rats (ANOVA, Student-Newman-Keuls post hoc test) for the same time point.

View larger version (34K):
[in this window]
[in a new window]
FIG. 4. Effects of LPS and MDP on serum CRP and PCT levels in immunocompetent and leukopenic rats. Serum levels of CRP (A and B) and PCT (C and D) in separate groups of immunocompetent and leukopenic rats were determined at 1, 2, 4, 8, and 12 h after i.p. injection of either PFS ( ${blacktriangledown}$ ), 100 µg/kg LPS (•), or 100 µg/kg MDP ( ${circ}$ ). Each experiment involved 10 to 12 rats per time point. At all time points tested after injection of PFS, both immunocompetent and leukopenic rats showed no significant elevation in serum levels of these mediators (P > 0.05). Symbols represent means; bars indicate SEM. *, P < 0.05 between LPS- and MDP-injected rats (ANOVA, Student-Newman-Keuls post hoc test) for the same time point.

View larger version (28K):
[in this window]
[in a new window]
FIG. 5. Mean integrated IL-6 (A), IL-10 (B), MIP-2 (C), CRP (D), and PCT (E) responses of immunocompetent and leukopenic rats after PFS (white bars), LPS (black bars), or MDP (hatched bars) challenge. Statistical analysis was performed on AUC data from 0 to 12 h after the time of injection and was calculated from the data shown in Fig. 3 and 4. Bars represent means ± SEM for each group of rats (nd, not detected). $§$ and *, statistically significant differences (P < 0.05) between LPS-and MDP-treated rats and their respective PFS-treated groups, respectively. ${ddagger}$ , significantly different (P < 0.05) from corresponding values between immunocompetent and leukopenic rats (ANOVA, Student-Newman-Keuls post hoc test).

In leukopenic rats, the concentrations of IL-6 (Fig. 3B), IL-10(Fig. 3D), and MIP-2 (Fig. 3F) in serum also started to riseat 1 h after LPS challenge. However, in contrast to those inimmunocompetent rats, serum IL-6 and MIP-2 levels in rats treatedwith MDP did not differ from those in leukopenic rats treatedwith their respective PFS controls at any time point. Comparedto leukopenic animals, the peak levels of all mediators, exceptfor IL-6, were significantly higher in immunocompetent rats(Fig. 5) (P < 0.05). Similar to the case for immunocompetentrats, PCT and CRP failed to discriminate between bacterial stimuliin leukopenic rats (Fig. 4B and D).

The comparative study using analysis of integrated areas underthe curve during the early phase of LPS- or MDP-induced fevershowed significant differences between immunocompetent and leukopenicrats (Fig. 5). Baseline levels of IL-6, CRP, and PCT in serumwere significantly higher in leukopenic rats. In contrast, baselinelevels of MIP-2 were significantly higher in immunocompetentanimals, and those of IL-10 were not detectable in either groupof rats. Injection of LPS or MDP evoked a significant decreasein IL-10, MIP-2, and PCT levels in leukopenic rats comparedto immunocompetent rats (P < 0.05) (Fig. 5). By contrast,IL-6 production was significantly higher in leukopenic ratsthan in immunocompetent rats after LPS administration (P <0.05) but not after MDP administration (P > 0.05). Both groupsof rats that received MDP showed lesser increases in serum IL-6(Fig. 5A), IL-10 (Fig. 5B), and MIP-2 (Fig. 5C) production thanwere observed in rats that received a similar dose of LPS. Finally,PCT and CRP failed to discriminate between bacterial stimuliin both groups of animals (Fig. 5D and E).

DISCUSSION

Top

Discussion

The present findings indicate that cytokines such as IL-6, IL-10,and MIP-2 are rapid and sensitive markers to differentiate gram-negativefrom gram-positive stimuli in immunocompetent and, especially,in leukopenic rats when measured at an early stage of feverinduced by specific pyrogens, such as LPS and MDP. Our resultsfurther suggest that CRP and PCT are not specific markers todiscriminate between these bacterial infections. We have chosenthese inflammatory markers because of interrelationships inthe primary nonspecific host response to microbial infection(11, 12, 17, 18). The levels of other cytokines, such as IL-1or tumor necrosis factor, were not determined in the presentstudy because of poorer predictive value for the presence ofbloodstream infection in patients with fever (15).

Gram-negative and gram-positive organisms have been consideredthe most important causes of infection in leukopenic patientswho undergo chemotherapy. It is therefore not surprising thatbacterial constituents that are present on the outside of thecell and therefore readily accessible to detection have beenselected by the immune system as indicators of bacterial presenceand potent inducers of a host response. During a bacterial infectionin neutropenic patients, it is of vital importance to the infectedhost that the presence of pathogenic bacteria is detected assoon as possible, since these patients present a high morbidityand this may lead to mortality if the infection is treated inappropriatelyor too late. The present knowledge regarding endogenous inflammatorymediators in leukopenic patients is limited but indicates thatthe host response to microbial infection is mediated by therelease of some pyrogenic mediators into the bloodstream (seethe introduction).

Studies with immunocompromised patients have shown that thesepatients are capable of producing high serum concentrationsof several mediators during severe systemic bacterial infection(4, 11, 21). In febrile neutropenic patients, elevated circulatingIL-6 and IL-8 also predicted microbial infection (8, 12). Insome reports, significant differences in the serum levels ofdifferent inflammatory mediators between gram-negative and gram-positiveinfections were described. Several investigators found differencesbetween neutropenic patients with and without bacteremia and/ordifferences between gram-negative and gram-positive bacteremia.The latter observations, however, are difficult to interpret.Many patients with other causes of fever are omitted from suchan analysis, but they need to be considered in order to calculatetrue sensitivities and specificities.

A key event in the initiation of fever in response to acutebacterial infection is the induction and release of endogenouspyrogens (i.e., IL-1, IL-6, IL-8, RANTES, and MIPs) by polymorphonuclearand mononuclear leukocytes into the bloodstream (19, 23, 32,40, 41, 44). Normal leukocytes contain no detectable pyrogenand do not produce any during in vitro incubation. After stimulation,however, pyrogen production begins within a few hours and continuesat a steady rate for 12 h or more (5). The key event is a rapidactivation of humoral cascade systems and of leukocytes thatsupport the inflammatory reaction. These cells are thought tobe an essential target for gram-positive and gram-negative bacteria,as they are the main source of inflammatory mediators involvedin the host acute response to these organisms (32). We foundthat leukopenic rats exhibit higher fevers after systemic injectionof LPS and MDP than immunocompetent rats, despite lower increasein circulating mediators. Moreover, we found that, independentlyof the immune status, rats that received similar doses of LPSshowed higher fevers and greater increases in serum cytokinelevels than those observed after MDP challenge. Interestingly,febrile responses actually subsided before the disappearanceof elevated levels of serum cytokines after injection of LPSand MDP in leukopenic rats. These results are consistent withthe idea that even when cytokines are detected in the blood,their timing does not always correlate with acute-phase activationafter administration of immune stimuli (19).

Concerning the role of administration of cyclophosphamide, ithas previously been shown that this immunosuppressive drug inducesa nonspecific leukopenia (fewer than 100 cells/ml for 6 to 7days) (24, 26) and a selective depletion of the lymphoid tissue(2, 42). When such treated patients become afflicted with abacterial infection, they develop high fevers. These resultssuggest that a lesser contribution of leukocytes may be importantto the exaggerated fever that occurs in leukopenic patients.However, in addition to peripheral blood leukocytes, the tissueresident macrophages are required for host defense against acutebacterial infection. In this context, production of cytokinesat the tissue level, rather than in circulation, could be analternative pathway for the induction of the signal leadingto fever in leukopenic animals.

Because the liver is the principal organ for LPS clearance andits resident macrophages, the Kupffer cells, are a major sourceof pro- and anti-inflammatory cytokines, it has been suggestedthat these cells could be central to production of fever byLPS (44). However, the activity of Kupffer cells is also depressedby cyclophosphamide in mice and rats (2, 22), and temporaryblockade of these cells with gadolinium chloride, a substanceknown to eliminate Kupffer cells (31), did not prevent the febrileresponse to i.p. LPS in leukopenic rats (E. Tavares, unpublisheddata). These results indicated that other cells types (fibroblasts,endothelial cells, glial cells, etc) involved in the acute inflammatoryresponse are major sites of the action of bacterial toxins inimmunosuppressed rats.

The present results indicate that cytokines such as IL-6, IL-10,and MIP-2 are regulated or released independently of the leukocytecount. How the absence of leukocytes leads to an intensifiedrather than a reduced febrile response to LPS and MDP and anassociated decreased production of circulating inflammatorymediators is still unknowable. That is, the reduction of circulatingleukocytes by cyclophosphamide caused significantly enhancedfevers (in both magnitude and duration), and this effect wasassociated with a significant reduction of circulating inflammatorycytokines compared to those in immunocompetent animals. Thepresent findings confirm previous experimental and clinicalstudies that show that the production of fever was not dependenton the presence of leukocytes in the peripheral circulation(3, 24, 28). Furthermore, they suggest that other, yet-to-beidentified mediators as possible mechanisms of immune signalingare involved in the development of the acute-phase responseto gram-negative and gram-positive infections. Together, thesefindings suggest that in immunosuppressed animals leukocytesmay play a "negative" role in the development of a fever inducedby exogenous pyrogens. However, further studies on the pathophysiologyof fever and the contribution of early markers regulated orreleased independently of the leukocyte count are necessaryto verify this hypothesis.

Our findings suggesting that the patterns of early productionof proinflammatory mediators may be dependent upon the specificityof the microbial pathogens are in agreement with recent studies(13). The differences in the cytokine pattern and the febrileresponses to LPS and MDP may correspond to the different cellsurface receptors stimulated by these bacterial pyrogens, withleukopenic animals being more sensitive to LPS than to MDP,which is in agreement with previous in vitro and in vivo studies(43). Functional characterization of TLRs has established thatinnate immunity is a skillful system that detects invasion ofmicrobial pathogens. Recognition of microbial components byTLRs initiates signal transduction pathways, which triggersexpression of genes. These gene products control innate immuneresponses and further instruct development of antigen-specificacquired immunity (37). In agreement with these findings, thepresent results indicate that the patterns of production ofendogenous mediators involved in the primary nonspecific acutehost response may be dependent upon the specificity of the microbialpathogens and the host recognition pathways invoked.

The present findings also illustrate the potential differencesin the patterns of whole-blood leukocytes in response to twodisparate, but common, bacterial stimulants. Taking all thesedata together, it is improbable that the pyrogenic effects ofLPS and MDP seen clinically in leukopenic patients are directlyand exclusively due to stimulation of pyrogen production byleukocytes. Overall, serum IL-6, IL-10, and MIP-2 levels areable to discriminate between gram-negative and gram-positiveinfection at start of fever in both immunocompetent and leukopenicrats with acceptable specificity. Furthermore, they suggestthat the exaggerated fever response occurring after LPS injectionin leukopenic animals is partially mediated via IL-6. On theother hand, CRP and PCT are not sensitive markers in distinguishingthese types of infections. However, although we used an experimentalrat model that resembles the pathophysiologic events that mayoccur in leukopenic patients, the findings reported here cannotnecessarily be extrapolated to the response to different gram-negativeor gram-positive pathogens or their constituents in immunosuppressedpatients. The mechanisms underlying the febrile response toLPS and MDP in leukopenic animals were not elucidated in thisstudy. This point remains to be resolved, although it is presentlyunder investigation in our laboratory.

The data presented here extend the argument that LPS and MDPinduce differential fever responses and cytokine profiles innormal and immunodeficient animals. Finally, these results indicatethat the determination of the serum pattern of specific inflammatorymarkers at the onset of fever may be used as an early diagnostictool to differentiate gram-negative from gram-positive infectionsin both immunocompetent and leukopenic hosts. A rapid bedsidetest may thus help guide the choice of specific antibiotic therapies,especially for neutropenic patients, even before culture resultsare available.

ACKNOWLEDGMENTS

We thank the Division of Hematology/Oncology at the Valme Hospitalfor helping with the study of hematological parameters.

This study was supported by Fundación Valme, Junta deAndalucía, and by a grant from the Spanish Fondo de InvestigacionesSanitarias (FIS 00-1066).

FOOTNOTES

* Corresponding author. Mailing address: Departamento de Farmacología, Pediatría y Radiología, Facultad de Medicina, Universidad de Sevilla, 41009 Seville, Spain. Phone: 34-955-015877. Fax: 34-954-905970. E-mail: jminano{at}us.es.

REFERENCES

Top