Background. Listeria monocytogenes tends to survive in phagocytes. Granulocyte macrophage colony-stimulating factor (GM-CSF) protects mice against L. monocytogenes infection, and mice knocked out for the GM-CSF gene are more susceptible to these infections.
Methods. THP-1 cells were used to characterize the GM-CSF receptor (binding isotherms; STAT5 phosphorylation), measure the intracellular growth of L. monocytogenes (5 h after phagocytosis), examine the influence of a 24-h incubation with GM-CSF before infection, measure the production of tumor necrosis factor (TNF)-α and the expression of nitric oxide synthase (iNOS), and evaluate the influence of anti-CGM-CSF receptor (GMCSFRα) and anti—TNF-α antibodies and the addition of Nω-nitro-l-arginine methyl ester (l-NAME) and catalase.
Results. THP-1 cells display functional GM-CSFRα. GM-CSF impairs the intracellular growth of L. monocytogenes to ∼65% of its value in unstimulated cells. This effect is abolished by anti–CGM-CSFRα, anti–CTNF-α antibodies, and catalase (and, to a lesser extent, by l-NAME). GM-CSF stimulates the release of TNF-α and the expression of iNOS. TNF-α added to unstimulated cells (even in large amounts) does not fully reproduce the impairment in the growth of L. monocytogenes caused by GM-CSF.
Conclusions. GM-CSF impairs the intracellular growth of L. monocytogenes by a synergistic action of the GMC-SF-triggered release of autocrine TNF-α and hydrogen peroxide and the production of NO (associated with the stimulation of the expression of iNOS).
Listeria monocytogenes is a facultative intracellular food-borne pathogen that causes life-threatening infections that mostly affect immunocompromised patients [1, 2]. These infections are often associated with a high rate of recurrence and dissemination, which has been related to the lack of eradication of intracellular forms, particularly those hosted in phagocytes . The results of previous studies have shown that interferon (IFN)-γ promotes the clearance of L. monocytogenes in vivo , and the results of in vitro studies have suggested that this effect could be mediated by a containment of L. monocytogenes in phagosomes  and the release of nitric oxide (NO) and hydrogen peroxide . Granulocyte macrophage colony-stimulating factor (GMC-SF) is another cytokine that protects mice against L. monocytogenes infection . It has also been shown that mice knocked out for the GM-CSF gene are more susceptible to L. monocytogenes infection . GM-CSF stimulates the production of tumor necrosis factor (TNF)-α  and enhances the oxidative burst that occurs after phagocytosis . The aim of the present study was to establish whether GM-CSF has an inhibitory effect on the intracellular growth of L. monocytogenes in macrophages and to try to unravel the mechanism of this effect in relation to the release of TNF-α and the expression of inducible NO synthase (iNOS). For this purpose, we used a previously established model of human macrophages (THP-1 cells). These cells are highly permissive for intracellular infection when they are unstimulated [11–15], but they can impair the intracellular growth of L. monocytogenes when they are activated by IFN-γ . Our results show that GM-CSF also impairs the intracellular multiplication of L. monocytogenes. This effect is related to the binding of GM-CSF to specific surface receptors and is mediated by a synergistic action of GM-CSF-triggered autocrine secretion of TNF-α and the GM-CSF-stimulated expression of iNOS.
Materials and Methods
Cytokines, antibodies, and main reagents. Recombinant human GM-CSF, recombinant human TNF-α, and monoclonal anti-human TNF-α antibodies were purchased from R&D Systems, and recombinant human IFN-γ was purchased from Roche Diagnostics. Monoclonal anti-GM-CSF receptor-α (GMCSFRa) antibodies were purchased from Maine Biotechnology Services. All cytokines were aliquoted in sterile PBS supplemented with 0.1% bovine serum albumin (BSA), stored at −20°C, and thawed immediately before use. Nq-nitro-l-arginine methyl ester (l-NAME) and catalase were purchased from Sigma-Aldrich. Gentamicin was kindly donated by Shering- Plough Belgium in the form of Geomycin (the brand name of the product registered for clinical use in Belgium). Cell-culture media and serum were purchased from Gibco Biocult (Life Science Technologies). Unless stated otherwise, all other reagents were purchased from E. Merck.
Cell cultures. All experiments were performed with THP- 1 cells (ATCC TIB-202), a human myelomonocytic cell line that displays macrophage-like activity . These cells were maintained as a loose suspension in RPMI 1640 medium supplemented with 10% decomplemented fetal calf serum (FCS) and 2 mmol l-glutamine/L in an atmosphere of 95% air/5% CO2. Cells were subcultured every third day by gentle shaking, followed by pelleting and reseeding at a density of ∼3×105 cells/mL. The cultures were used after 2 days, when they had reached a density of ∼0.5–1×106 cells/mL.
Characterization of GM-CSF receptors. Binding assays were performed by exposing 107 cells/mL at 4°C for 2 h to increasing concentrations of [125I]-GM-CSF (specific radioactivity, 47.8 TBq/mmol; NEN Perkin Elmer Life Science) in RPMI 1640 medium supplemented with 2% FCS. After incubation, cells were washed 4 times with ice-cold PBS supplemented with 2% FCS, and cell-associated radioactivity was thereafter determined by γ-scintillation counting. Nonspecific binding was determined by repeating these experiments in the presence of a 100-fold excess of unlabeled GM-CSF. Specific binding was determined by subtraction of the nonspecific binding from the total binding. Binding parameters were determined by nonlinear regression (1 site-binding hyperbola). Phosphorylation of the transcription activator of GM-CSF (STAT5) [17, 18] was assessed by Western Blot analysis, by use of specific antiphosphorylated STAT5 antibodies (see details below).
Evaluation of the influence of GM-CSF on the intracellular growth of L. monocytogenes. Unless stated otherwise, the following protocol was adopted. Cells were exposed to GMCSF for 24 h at 37°C, washed in prewarmed sterile PBS, and then challenged with L. monocytogenes infection (5 h of intracellular growth) in the absence of GM-CSF, according to a method described in detail elsewhere . In brief, 106 cells were exposed to a fresh inoculum of 5×106 cfu of L. monocytogenes (hemolysin-producing strain EGD, serotype 1/2a) in a volume of 2 mL, for 1 h at 37°C. Nonphagocytosed and non-firmly adherent bacteria were removed by 4 successive centrifugations and washing in prewarmed sterile PBS. Cells were thereafter placed in fresh medium and collected immediately (time 0) or cultured for 5 h. As has been shown elsewhere [14, 19], no extracellular growth of L. monocytogenes occurs during this time frame. Cells were harvested by centrifugation, washed with ice-cold sterile PBS, and lysed in distilled water. Samples were then plated on tryptic soy agar after appropriate dilutions, for determination of the number of viable bacteria by counting colony-forming units after 24 h of incubation at 37°C, using a Gel Doc 2000 apparatus (Bio-Rad Laboratories) operated with Quantity One software (Bio-Rad Laboratories). The actual number of colonies counted was typically between 200 and 800 and was never <80. The number of colony-forming units in each sample was related to its cell protein content, and results were expressed as growth index by calculating the ratio of the number of colony-forming units per milligram of protein observed in samples collected 5 h after phagocytosis to the number observed after phagocytosis. The absence of extracellular bacteria was ascertained by plating aliquots of the culture medium collected at the end of the incubation. No bacterial growth was detected after 48 h at 37°C.
Measurement of TNF-α production. This was performed by use of an ELISA with OptEIA Set human TNF-α (Pharmingen), by use of a standard protocol. In brief, supernatants collected by centrifugation of THP-1 cells for 24 h (with or without GM-CSF) were added to 96-wells plates coated with mouse anti-human TNF-α antibody (dilution, 1:500). Results were revealed by using biotinylated mouse anti-human TNF-α antibody (dilution, 1:500) and avidin-horseradish peroxidase (HRP; dilution, 1:250), followed by 400 mg/L of o-phenylenediamine and 10% urea peroxide in citrate buffer (0.1 mol/ L, pH 5). The reaction was stopped after 30 min in the dark with 2 N H2SO4. Optical densities were read with a Novapath plate reader (Bio-Rad Laboratories), by use of a 490- and 650- nm filter for background.
Western Blot analyses (phospho-STAT5, iNOS, and actin). Cells were collected by centrifugation, resuspended, and lysed in O'Farrel 1× sample buffer (Laemmli sample buffer; Bio-Rad Laboratories, β-mercaptoethanol and water in a ratio of 95:5:100) at a concentration of 106 cells/mL. Lysates were then subjected to SDS-PAGE and immunoblotted with appropriate antibodies: rabbit monoclonal anti-human phosphoSTAT5 (Cell Signaling Technology; diluted 1:500 in 0.8% BSA), iNOS, rabbit polyclonal antihuman NOS 2 (Santa Cruz Biotechnology; diluted 1:500 in 0.8% BSA), actin, and mouse monoclonal anti-human β-actin (Sigma-Aldrich, diluted 1:1000 in 0.8% BSA). The membranes were applied to Clear Blue X-ray films (CL-Xposure film; Pierce Biotechnology), and the bands were revealed with goat antirabbit (phospho-STAT5 and iNOS) or goat anti-mouse (actin) polyclonal antibodies (diluted 1:500 in 0.8% BSA) coupled to HRP (1:20,000; Transduction Laboratories), by use of Super Signal West Pico Chemiluminescent Substrate (Pierce Biotechnology). The films were then developed and scanned using an Image Station 440 CF (Kodak Digital Sciences) and DS-1-D software (Eastman Kodak). The apparent molecular weight of the proteins of interest was estimated using standards of known molecular weight ran in parallel with the samples (BenchMark Pre-Stained Protein Ladder; Invitrogen).
Biochemical assays. The protein content of the samples was assayed by use of the Folin-Ciocalteu/biuret method , and NO release was measured by use of the nitrite/nitrate colorimetric method and a commercial kit (Roche Diagnostics).
Statistical analysis. Curve-fitting analyses were performed with Prism software (version 2.01; GraphPad Software). The statistical significance of differences between means was determined using Student's t test.
Characterization of GM-CSF receptors. In a first series of experiments, we tested for the presence and functionality of high-affinity GM-CSF receptors on THP-1 cells. As shown in figure 1, the binding of [125I]-GM-CSF was saturable, with an apparent KD13.87 pmol/L (0.2 ng/mL) and an estimated D maximum binding capacity of ∼100 molecules/cell (under the assumption of monovalent receptor-ligand interaction). The phosphorylation of STAT5 (the transcription activator of GMCSF) [17, 18] was then examined by Western Blot analysis of lysates of cells incubated with GM-CSF (5–40 ng/mL for 24 h at 37°C), and results were revealed by use of an antibody directed against the phosphorylated STAT5. The upper inset of figure 1 shows a definite reaction for a protein with an estimated molecular weight of 90 kDa, which is typical of STAT5.
Influence of GM-CSF on the intracellular growth of L. monocytogenes. In our model, L. monocytogenes achieved an ∼10-fold intracellular growth within 5 h after phagocytosis, compared with the initial inoculum . Figure 2A shows that this growth was impaired to ∼65% of the value of controls when cells had been preincubated for 24 h with 20 ng/mL GMCSF. In a next series of experiments, we examined the concentration dependency of the impairment of the growth of L. monocytogenes caused by GM-CSF. Figure 2B shows that a marked decrease in growth was observed once the concentration of GM-CSF was >5 ng/mL and reached its maximal effect at ∼20 ng/mL.
To ascertain that GM-CSF acted on the host cell and not directly on bacteria, we first tested that this cytokine had no direct effect on the growth of L. monocytogenes in broth. No influence was seen on growth curves, for concentrations up to 200 ng/mL. We then measured the capacity of THP-1 cells to phagocytose L. monocytogenes and found no significant differences in initial inoculum between control and GM-CSF-exposed cells. To further discount an effect on extracellular bacteria, the experiments described in figure 2A were repeated using gentamicin (18 mg/L; 20× the MIC, added during the 5 h after phagocytosis incubation). As shown in table 1, gentamicin caused an apparent reduction in growth of ∼15%, compared with controls (this effect has been observed before and was ascribed to the killing activity of gentamicin toward adherent, but not phagocytosed, bacteria and bacteria that re mained in phagosomes ; this decrease was indeed observed during the first 1 h of incubation, after which bacterial growth proceeded at the same rate as in cells not exposed to gentamicin). Under these conditions, pretreatment of cells with GM-CSF reduced the growth of L. monocytogenes to ∼57% of the value observed without pretreatment, versus 66% for cells not exposed to gentamicin. Although this difference was not statistically significant, it illustrates that the effect of GM-CSF is probably directed against intracellular bacteria.
To establish whether the impairment of the growth of L. monocytogenes caused by GM-CSF was related to its binding to its receptor, these experiments were repeated in the presence of monoclonal blocking antibodies raised against the α-subunit of GM-CSFRα (2.5 mg/mL; this concentration was ∼2.5 times the concentration needed to neutralize all GM-CSF receptors present in the system, on the basis of the titration of the batch of antibody used in these studies). Antibodies were added to the culture medium before the addition of GM-CSF and were maintained until GM-CSF was washed off and cells were challenged by infection with L. monocytogenes. Table 1 shows that the impairment of intracellular bacterial growth caused by GMCSF was completely abolished under these conditions. We tested whether the addition of gentamicin (18 mg/L) modified the neutralizing effect of the anti-CGM-CSFRα antibodies; to ensure that they did not act in an unspecific fashion on cell defense mechanisms or by a mere toxic effect, we measured the effect of IFN-γ on the intracellular growth of L. monocytogenes in the presence and absence of these antibodies. In both cases, the growth of L. monocytogenes was inhibited to ∼20% of the value observed in unstimulated cells, which is similar to what has been observed elsewhere in the same cells .
Role of TNF-α. GM-CSF is known to induce the production of TNF-α in macrophages . We therefore measured its release from THP-1 cells on exposure to GM-CSF, at increasing concentrations, for 24 h. As shown in figure 3A, a detectable level of release was noted at a concentration of 5 ng/mL, but it increased markedly thereafter, with a tendency to reach a plateau at ⩾2.5 ng/mL for cells incubated with ⩾20 ng/mL GM-CSF. The impairment of the growth of L. monocytogenes was correlated with the level of the production of TNF-α (figure 3A, inset). We then examined the influence of neutralizing monoclonal antibodies raised against human TNF-α. These antibodies, added in an amount sufficient to neutralize ∼10 ng/mL TNF-α, were maintained during the entire exposure of the cell to GM-CSF. As shown in figure 3B, although GM-CSF caused a significant decrease in bacterial growth (P > .05), the addition of anti–TNF-α antibodies completely abolished this effect. These antibodies had no effect on the growth of L. monocytogenes in cells that were not exposed to GM-CSF. To further rule out any unspecific action of these anti–TNF-α antibodies, we examined their influence on the GM-CSF–triggered phosphorylation of STAT5 and the impairment of the growth of L. monocytogenes induced by IFN-γ. No effect was seen in either case.
To test whether the secretion of TNF-α was the only sufficient mediator to cause the impairment of the intracellular growth of L. monocytogenes on treatment of the cells with GM-CSF, we examined the influence of exogenous TNF-α (10 ng/mL) maintained in the presence of the cells for 24 h at 37°C before challenging them with L. monocytogenes. As shown in figure 3B, TNF-α alone had only a modest, nonsignificant effect that did not fully reproduce the effect observed with GM-CSF.
Role of NO and hydrogen peroxide. Murine macrophages are known to produce large quantities of NO and hydrogen peroxide on stimulation with GM-CSF [10, 21]. The results of Western blot analysis (figure 4) showed that the expression of iNOS was markedly induced in THP-1 cells after 24 h of treatment with GM-CSF, without the addition of anti-TNF-α antibodies (which demonstrates that it was not primarily mediated by the release of TNF-α; we also tested whether the addition of l-NAME was without influence on this expression; data not shown). In parallel, we measured the production of NO in cells treated with GM-CSF (using the nitrite/nitrate method) and found a marked increase in its production, compared with controls. This production was abolished in the presence of l-NAME (400 µmol/L; data not shown). We then examined the influence of l-NAME, catalase, and the combination of both on the impairment of the growth of L. monocytogenes induced by GMCSF. As shown in figure 5, l-NAME and catalase, when maintained in contact with the cells for 24 h at 37°C together with GM-CSF, not only abolished the impairing effect of GM-CSF on the intracellular growth of L. monocytogenes in cells treated with GM-CSF alone (P <.05) but actually significantly stimulated this growth (∼28% more than in controls). A similar increase was seen in cells incubated with GM-CSF and catalase alone, which indicates that the production of hydrogen peroxide was probably a main factor that contributed to the impairment of the growth of L. monocytogenes in GM-CSF-stimulated THP-1 cells. l-NAME alone was also effective in abolishing the impairing effect of GM-CSF and in stimulating the growth of L. monocytogenes, but to a lesser extent. The addition of l-NAME, catalase, or both, without GM-CSF, had the same effect as that observed in the presence of GM-CSF (data not shown).
The present study has provided direct evidence that GM-CSF impairs the intracellular growth of L. monocytogenes in THP- 1 cells and that this effect is mediated by the binding of this cytokine to its receptor and by a GM-CSF-triggered release of TNF-α and oxidant species (nitrogen oxide and hydrogen peroxide) related to the expression of iNOS. THP-1 macrophagelike cells were chosen because they are largely unable to kill L. monocytogenes, in contrast to what is observed with resident macrophages or monocytes [22, 23]. This offered the possibility to quantify the influence of GM-CSF independently of the other mechanisms of cell defense, a point that may need to be taken into account to explain the discordant results concerning the action of GM-CSF on Listeria infection . Conversely, it must be stressed that THP-1 cells are undifferentiated, nonadherent cells that may, therefore, not represent fully functional macrophages as they exist in vivo. Yet we have shown here that THP-1 cells have a basal production of oxidant species that limits the inhibition of the growth of L. monocytogenes. Moreover, these cells respond to GM-CSF to further limit this growth and have successfully been used previously to examine the mechanism of the protective effect exerted by IFN-γ against L. monocytogenes infection and to analyze its cooperation with antibiotics for the eradication of L. monocytogenes from these cells [6, 14].
The impairment of intracellular growth observed in THP-1 cells pretreated with GM-CSF can reasonably be attributed to an effect toward the intracellular bacteria that is related to the activation of a specific surface receptor. First, we have shown that THP-1 cells display high-affinity receptors for GM-CSF, with binding parameters to similar those found in other macrophages [25, 26] and with evidence of functionality (as assessed by the demonstration of the ligand-triggered phosphorylation of STAT5) [17, 18]. Second, all influence of GM-CSF on the growth of L. monocytogenes is suppressed when activation takes place in the presence of an anti-receptor-blocking antibody. This antibody did not impair the IFN-γ-dependent protective effect and could, therefore, not act in an unspecific fashion or by a global toxic effect. Third, there was no evidence of a direct toxic effect of GM-CSF on bacteria, as evidenced by the lack of effect of GM-CSF on bacterial growth in broth and the additive effects of treatment with GM-CSF (before phagocytosis) and gentamicin (added after phagocytosis). The latter finding strongly suggests that GM-CSF and gentamicin act on distinct pools of bacteria—namely, the truly intracellular ones for GM-CSF and the adherent but not phagocytosed organisms, as well as those remaining in phagosomes for gentamicin . This also allows us to discount an effect of GM-CSF on bacterial adherence and an ensuing slowing down of the internalization of L. monocytogenes—in that case, GM-CSF and gentamicin should not have additive effects, because their action would be exerted against the same pool of extracellular bacteria. Fourth, we discounted a potential effect of GM-CSF on the phagocytosis of L. monocytogenes in our model. This control was important, because early studies had shown that macrophages obtained from transgenic mice that produce high levels of GM-CSF phagocytose more actively than cells from normal littermates while being no more efficient at lysing intracellular bacteria . It is therefore important to underline that the phagocytosis of L. monocytogenes and its subsequent intracellular growth were studied here in the absence of GM-CSF during this process. This experimental design was specifically meant to examine the long-term result of the activation of the cells by GMCSF on the handling of L. monocytogenes, and not a direct effect of GM-CSF on the bacteria (whether making them more fragile or enhancing their level of phagocytosis).
The initial role of the binding of GM-CSF to its receptor in the subsequent impairment of the growth of L. monocytogenes could be disputed on the basis that a significant protective effect is only seen at concentrations (⩾10 ng/mL) that are far above the saturation point of the putative receptor of [125I]-labeled GM-CSF. A plausible explanation could be that the low number of receptors for GM-CSF on THP-1 cells makes them relatively insensitive to the action of this cytokine, which therefore requires the full occupancy of the receptors to yield a significant protective effect. Future experiments with cells overexpressing the GM-CSF receptor could be useful in this context.
The role of TNF-α as a mediator of GM-CSF, for impairing the intracellular growth of L. monocytogenes appears clearly from the observations: the production of TNF-α after exposure of the cells to GM-CSF proceeds quantitatively, almost in parallel with its effect on bacterial growth, and the neutralization of TNF-α suppresses all response to treatment with GM-CSF. The simple fact, however, that the addition of TNF-α to cells not exposed to GM-CSF does not allow them to reproduce the action of GMCSF also indicates that TNF-α alone is insufficient to bring about a significant impairment of the growth of L. monocytogenes. Our current hypothesis is, therefore, that TNF-α produced by cells on stimulation with GM-CSF acts through a NO-independent pathway but that neither autocrine TNF-α nor induced free radicals are able to control the intracellular growth of L. monocytogenes by themselves in THP-1 macrophages. The necessity for TNF-α to be combined with other factors to exert a significant effect on the growth of L. monocytogenes has already been observed in IFN-γ . In parallel, TNF-α can limit the growth of L. monocytogenes in macrophages in a NO-independent manner , which suggests the existence of multiple mechanisms. Finally, although very large amount of nitrosative species may be needed to obtain a complete inhibition of the intracellular growth of L. monocytogenes, macrophages stimulated simultaneously by IFN-γ, TNF-α, and lipopolysaccharide are still able to reduce intracellular growth in the presence of iNOS inhibitors, which demonstrates the involvement of both NO-dependent and NOindependent pathways for the control of bacterial growth .
Murine macrophages are known to produce large concentrations of NO and hydrogen peroxide after stimulation with GM-CSF [10, 21], and resident peritoneal cell populations from GM-CSF−/− mice show a reduced production of NO . We have shown here that THP-1 cells also respond to GM-CSF by an increase in the production of NO, and we have presented evidence of a GM-CSF-dependent expression of iNOS in these cells. The data also show that the expression of iNOS and the production of TNF-α are 2 independent consequences of the exposure of THP-1 cells to GM-CSF, because the expression of iNOS was not affected by the presence of antibodies raised against TNF-α during exposure to GM-CSF. In parallel, we have shown that the production of hydrogen peroxide is critical in the impairment of the growth of L. monocytogenes that is brought about by GM-CSF, because this effect of GM-CSF was entirely defeated by the addition of catalase. NO, however, should also play a role, because the addition of l-NAME was also effective. Incidentally, these results also show that unstimulated THP-1 cells were capable of a basal production of NO and hydrogen peroxide in response to infection: bacterial growth was more extensive in the presence of l-NAME and catalase (without GM-CSF) than in control cells. GM-CSF appears, therefore, to be an add-on factor in this context. We do not know, however, whether the GM-CSF-triggered increased expression of iNOS and production of NO (and, most likely also, of hydrogen peroxide) is sufficient to impair the growth of L. monocytogenes or whether TNF-α must also be present. The results of previous studies have shown that TNF-α triggers macrophages to produce large amounts of reactive nitrogen intermediates but that the complete inhibition of this production does not decrease cell bactericidal activity , which suggests a direct effect of TNF-α on bacterial growth. Conversely, no additional factor besides NO and hydrogen peroxide seems to be necessary to severely limit the growth of L. monocytogenes in THP-1 cells pretreated with IFN-γ . The latter, however, not only triggers the production of nitrogen and oxygen reactive species but also prevents the migration of L. monocytogenes from phagosomes to the cytosol, where the bacteria will actively multiply . We do not know how GMCSF influences the subcellular distribution of L. monocytogenes in THP-1 cells nor how a synergy among TNF-α, NO, and hydrogen peroxide could be important in this context. These critical issues will need to be examined in future experiments.
In spite of these uncertainties, our results already provide interesting perspectives in 2 complementary directions. First, they may help us to understand how GM-CSF affords protection of the immunocompetent host against infection by L. monocytogenes, as has been demonstrated in animals [7, 8]. In a broader context, GM-CSF could therefore be viewed as participating in the general control of the L. monocytogenes infection by the immunocompetent host, making the bacteria less able to survive intracellularly, as has also been observed with IFN-γ [5, 6]. This potential role of GM-CSF, in combination with other cytokines, for increasing the antibacterial competence of macrophages has already been advocated . More specifically, GM-CSF could also cooperate for defense against Listeria infection with other proinflammatory cytokines, such as interleukin (IL)-12 [31, 32] and IL-18 [33, 34]. Intriguingly, the administration of a traditional Chinese medicine that protects against Listeria infection in mice has been shown to trigger the production of IL-1, IL-6, and GM-CSF . Conversely, it could be of interest to monitor the production of GM-CSF in response to an L. monocytogenes infection and to examine to what extent variations under specific conditions and/or among patients could be correlated to commensurate changes in risk factors. Second, our data may suggest the use of GM-CSF as a pharmacological agent for the control of Listeria infection in patients who are at risk. It could be argued that a protective effect of GM-CSF could only be observed at concentrations far above those observed physiologically in humans, even during infection and/or after the administration of endotoxin . Yet the administration of GM-CSF for postchemotherapy bone-marrow recovery at conventional doses allows it to reach concentrations up to 5 ng/mL, and higher doses appear to be well tolerated . Animal and exploratory clinical trials with GM-CSF may therefore be warranted.
Jérôme Zanon (Unité de Pharmacologie Cellulaire et Moléculaire, Université Catholique de Louvain, Brussels, Belgium) developed the assay for tumor necrosis factor-α, and Jean-Christophe Renauld and Diane Lejeune (Ludwig Institute for Cancer Research, Brussels, Belgium) provided guidance for the Western blot analyses.