Migration Inhibitory Factor Induces Killing of Leishmania major by Macrophages: Dependence on Reactive Nitrogen Intermediates and Endogenous TNF-α¹

Macrophage migration inhibitory factor (MIF)³ was originally described as a T cell-derived cytokine that inhibits the random migration of macrophages (1, 2). More recently, MIF has been found preformed in corticotropic cells of the anterior pituitary gland (3, 4, 5) and in monocytes/macrophages (6). MIF plays an important role in LPS-induced septic shock (3, 7, 8, 9, 10, 11) and has the unique property to exhibit cytokine as well as enzymatic activities. For example, MIF has been shown to catalyze the tautomerization of d-dopachrome into 5,6-dihydroxyindole-2-carbaxolic acid in vitro (12). Furthermore, MIF has recently been found to have protein-thiol oxidoreductase activity (13).

MIF acts as a profound counter-regulator of the anti-inflammatory and immunosuppressive effects of glucocorticoids (11). Recently, MIF has been demonstrated to be induced by glucocorticoids, and recombinant MIF has been shown to override glucocorticoid inhibition of proinflammatory cytokines, such as TNF-α, IL-1β, IL-6, and IL-8 in LPS-stimulated human monocytes (14). Also, a proinflammatory spectrum of action for MIF has been verified. MIF induces TNF-α and IL-8 secretion by macrophages (7) or in bronchoalveolar lavage cells, respectively (15). Recombinant MIF in concert with other proinflammatory cytokines, such as IFN-γ, has been found to promote macrophage nitric oxide (NO) release (10). In an immunologically induced kidney disease in the rat, anti-MIF Ab resulted in a marked inhibition of disease progression and iNOS expression (16). Various proinflammatory stimuli, such as LPS, TNF-α, and IFN-γ, were potent inducers of MIF secretion (6). Previous reports that a cloned species of human MIF activates macrophages to kill Leishmania donovani had to be retracted due to PHA contamination of the MIF-containing COS cell supernatants employed (17, 18).

The function of macrophages, the most important effector cells in murine leishmaniasis, is determined by a variety of cytokines, mainly produced by T cells (for review, see Refs. 19 and 20). While the Th1-specific cytokine IFN-γ has been shown to activate macrophages to synthesize iNOS and thus to enhance the killing of intracellular leishmania amastigotes, the Th2 cytokines IL-4 and IL-10 have been demonstrated to deactivate these effector functions of macrophages. Mice developing a Th1 response to cutaneous leishmaniasis are able to control the local lesion, whereas mice with a predominant Th2 answer cannot restrict the parasite replication and die of visceral leishmaniasis (19, 20). MIF has recently been identified as a cytokine produced by mitogen-stimulated Th2, but not Th1 cell clones (21). Among the MIF-producing T cell clones examined was L1/1, a T cell clone specific for Leishmania major (22). Therefore, it was the aim of this study to analyze whether MIF is able to influence the leishmanicidal effector functions of macrophages in vitro. Using murine macrophages infected with L. major, we examined 1) the effect of MIF on leishmanicidal effector functions of the host cells of this protozoan parasite, 2) the mechanism by which the leishmanicidal effect is mediated, and 3) the possible influence of known macrophage-deactivating cytokines, such as IL-10, IL-13, and TGF-β, on MIF-induced macrophage activation.

Recombinant mouse and human MIF were cloned, expressed in Escherichia coli, and purified by anion exchange and reverse phase chromatography as described previously (23, 24). The LPS content of purified recombinant MIF was <20 pg/μg protein as measured by the QCL-1000 Limulus amebocyte lysate assay obtained from BioWhittaker/Serva (Heidelberg, Germany), performed according to the manufacturer’s instructions. Recombinant mouse IFN-γ, recombinant murine IL-10, and recombinant murine IL-13 (LPS content of each cytokine, <0.1 ng/μg protein) were purchased from R&D Systems (Wiesbaden, Germany). Recombinant murine TNF-α (LPS level, <0.1 ng/μg protein) was obtained from IC Chemicals (Munich, Germany). TGF-β (PHA, Hannover, Germany) had an LPS content of <20 pg/μg protein. The reagents for TNF-α ELISA were obtained from PharMingen (Hamburg, Germany). Cells were cultured, and all cytokines were diluted to the concentrations used in the experiments in C-RPMI culture medium 1640 (Biochrom, Berlin, Germany; supplemented with 2 mM glutamine, 10 mM HEPES buffer, 7.5% NaHCO₃, 0.05 mM 2-ME, 100 μg/ml penicillin, 160 μg/ml gentamicin, and 10% selected FCS with a total LPS content <100 pg/ml).

LPS (E. coli serotype O127:B8) was purchased from Sigma (Munich, Germany). l-NIL was obtained from Calbiochem (San Diego, CA). Murine anti-mouse MIF mAb was generated as described previously (16, 25). Murine IgG (Sigma) was used as an isotype control. Rabbit antiserum raised against recombinant mouse TNF-α was provided by Dr. H. U. Beuscher (Erlangen, Germany). Rabbit antiserum was generated against recombinant mouse IL-10 as described previously (26). Rabbit preimmune serum served as control serum.

Female mice of the inbred strains BALB/c, C57BL/6, and C3H/HeJ were obtained from Charles River Breeding Laboratories (Sulzfeld, Germany). TNFR55^−/− mice (27) as well as sex- and age-matched control animals were gifts from Dr. Klaus Pfeffer (Munich, Germany); iNOS^−/− mice (28) as well as sex- and age-matched control animals were provided by Dr. Christian Bogdan (Erlangen, Germany).

L. major promastigotes of the strain MHOM/IL/81/FEBNI were grown in vitro in blood agar cultures as described previously (29). Stationary phase promastigotes were washed in PBS and were added to the macrophage cultures as indicated below.

T cell-free macrophage monolayers were prepared and tested for purity as described in detail previously (30, 31). Briefly, thioglycolate-elicited peritoneal exudate cells (PEC) were washed twice and resuspended in C-RPMI culture medium, seeded into Lab-Tek tissue culture chamberslides (2 × 10⁵ cells/chamber; Nunc, Wiesbaden, Germany), and allowed to adhere for 3 to 4 h (37°C, 5% CO₂, 95% air humidity); thereafter, nonadherent cells were removed by three extensive washings with culture medium.

Bone marrow-derived macrophages (BMM) were prepared essentially as described previously (30, 32). Cells were prepared from femurs of mice, and after 3 days in culture, nonadherent progenitor cells were taken and cultured for an additional 7 days in culture medium supplemented with 30% (v/v) L cell-conditioned medium as a source of CSF-1. Adherent BMM were harvested with a rubber policemen and seeded into Lab-Tek tissue chamberslides.

Before infection, macrophage cultures were incubated in culture medium or in culture medium containing cytokine(s), anti-cytokine Abs, or reagents for a standard time of 4 h (in some experiments a different incubation time was used, as indicated in the text). Macrophages were then infected with L. major promastigotes (parasite/cell ratio, 10:1) for 4 h. Thereafter, nonphagocytosed parasites were washed off, and the cultures were further incubated in the presence or the absence of cytokines and Abs for 72 to 96 h. Intracellular live amastigotes were assessed after staining with ethidium bromide (50 μg/ml) and acridine orange (5 μg/ml) in PBS by fluorescence photography as described in detail previously (30). The percentage of infected macrophages (mean ± SD) of three or four different cultures were tested for statistical significance by Student’s t test for unpaired samples (two-tailed).

The nitrite concentration was measured by a microplate assay method with Griess reagent (1% sulfanilamide/0.1% naphthylethylene diamine dihydrochloride/2.5% H₃PO₄) as previously described (33).

Female mice of the inbred strain BALB/c and C57BL/6 at 6 to 12 wk of age were infected with 2 × 10⁶ L. major promastigotes in volume of 50 μl intradermally. After RNA extraction by use of acidic guanidinium thiocyanate (34) from lymph nodes or skin at the time points indicated, the levels of MIF mRNA were quantified applying the RiboQuant multiprobe RNase protection assay kit mCK-2b (PharMingen, San Diego, CA) according to the manufacturer’s instruction. Ten micrograms of total RNA from tissues of two mice per group and time point were analyzed in each reaction, and the intensities of the resulting bands were quantified with a BAS 2000 Bioimager (Fuji, Tokyo, Japan) and TINA 2.0 software (Raytest, Straubenhardt, Germany).

Total lymph node cells of leishmania-infected C57BL/6 or BALB/c mice were cultured for 72 h in vitro at a density of 2 × 10⁶ cells/ml in the presence or the absence of L. major freeze/thawed promastigote Ag (Lsh Ag; 5 × 10⁶ cell equivalents/ml) or concanavalin A at a concentration of 7.5 μg/ml (Sigma). For measurements of MIF, protein-conditioned media were analyzed by sandwich ELISA employing a monoclonal anti-MIF capture Ab, a polyclonal rabbit anti-mouse MIF detector, and purified mouse rMIF as standard as described previously (14).

Recombinant human as well as mouse MIF significantly reduced the percentage of infected cells and the number of parasites per 100 macrophages (BALB/c PEC) in a dose-dependent manner (Fig. 1). The fact that MIF stimulates PEC of LPS-insensitive C3H/HeJ mice to the same extent (70 ± 2.4% infected PEC in the medium control compared with 48 ± 1.7% infected PEC cultured with 2.5 μg/ml murine MIF) as PEC of BALB/c and C57/BL6 mice excludes macrophage activation mediated by low levels of LPS (<20 pg/μg MIF). The MIF dose necessary for a significant reduction exceeded 1 μg/ml. This is about 100 times higher than that of all cytokines identified to date to enhance the leishmanicidal effector functions of macrophages, but approximately the same MIF dose (0.1–10 μg/ml) is required to induce TNF-α and NO secretion (10) and within the range of concentrations (e.g., up to 340 ng/ml in sera of mice) observed during inflammation in vivo (35). Addition of anti-MIF mAb, but not control mAb, completely reversed the stimulatory effect of murine MIF, demonstrating the specificity of the observed macrophage activation (Fig. 2, A and B). Furthermore, MIF enhanced the killing of L. major in BMM of BALB/c and C57BL/6 mice, although in these cells the overall infection rate was markedly reduced (Fig. 2,E). As demonstrated in Figure 2, C and D, experiments with kinetic assessment (24, 48, 72, and 96 h post-L. major infection) revealed that the MIF effect was particularly evident late after infection (e.g., 72 and 96 h). MIF also induced killing of parasites in the presence of the IFN-γ-neutralizing Ab R4.6A2, which was capable of blocking the activity of 10 ng/ml of rIFN-γ. Furthermore, there was no significant difference in the effect of an isotype-matched control mAb and that of the R4.6A2 mAb on MIF-induced Leishmania killing (data not shown). Thus, we could exclude a functionally significant contribution of IFN-γ produced by (very low) numbers of contaminating T or NK cells on the MIF effects on cultured macrophages. The finding that the simultaneous addition of IFN-γ together with MIF resulted in a much more pronounced activation of macrophages (11 ± 3% of L. major-infected cells) than optimal concentrations of either of the cytokines alone further argues for an IFN-γ-independent effect of MIF.

MIF has previously been demonstrated to promote the release of TNF-α from RAW 264.7 macrophages and in concert with IFN-γ to induce the release of NO (10). Therefore, we next asked whether the MIF-mediated activation of macrophages to kill L. major amastigotes was dependent on endogenously produced TNF-α and/or NO. MIF induced the production of detectable amounts of TNF-α (60 ± 18 pg/ml) in cultures of L. major-infected PEC. As demonstrated in Figure 3,B, administration of anti-TNF-α antiserum, but not of a control serum, markedly reduced MIF-mediated macrophage activation. For TNF-α, two different receptor molecules (TNF-R; molecular masses, 55 and 75 kDa) have been described in mice (36). The roles of these two receptors in TNF-α-mediated cytotoxity and antimicrobial effector function are not yet fully understood and remain a topic of controversial discussion (37, 38, 39, 40). To determine the role of the smaller TNF-R molecule, we studied the PEC of TNF-R55^−/− mice. Unlike the PEC of wild-type control mice, L. major killing was not enhanced after stimulation with MIF in TNF-R55^−/− macrophages. Consistent with this, in the receptor-deficient cells there was no synergistic effect of exogenously added TNF-α and IFN-γ, in contrast to wild-type PEC (Fig. 4). These findings demonstrate the crucial role of endogenously produced TNF-α and the requirement of the 55-kDa TNF-R molecule for the MIF-mediated activation.

We next investigated the question of whether MIF-mediated killing of leishmania is mediated by reactive nitrogen metabolites produced endogenously by macrophages. Wild-type PEC were incubated with MIF together with the highly specific iNOS inhibitor l-NIL or with MIF alone as a control before and throughout infection with L. major. l-NIL completely blocked MIF-mediated killing of parasites (Fig. 3 A). In line with this observation, MIF does not enhance leishmania killing in PEC obtained from iNOS-deficient mice (28) (provided by Dr. C. Bogdan at our institute; data not shown). Thus, we deduce that reactive nitrogen metabolites generated endogenously by macrophages play an essential role in MIF-mediated enhancement of the leishmanicidal effector function. However, in agreement with findings reported by Bernhagen et al. (10), no significant elevation of the NO concentration was measured in supernatants from MIF-stimulated macrophages compared with unstimulated macrophages. By contrast, stimulation with IFN-γ and MIF resulted in a marked increase in nitrite levels (32.1 ± 3.3 μM) compared with those in supernatants stimulated with IFN-γ alone (17.2 ± 2.0 μM).

Following infection with L. major, a complex network of different cytokines become expressed that has only been partially defined (19, 20). Thus, we examined whether MIF-mediated macrophage activation could be inhibited by cytokines known to inhibit the effect of stimulatory cytokines, e.g., IL-10, IL-13, and TGF-β, since these mediators have been shown to impair the intracellular killing of leishmania in vitro (26, 41, 42). In addition, these cytokines are secreted by Th2 cells (IL-10, IL-13) and macrophages (IL-10, TGF-β), cells that are also important sources of MIF (1, 2, 6, 43, 44, 45, 46, 47). TGF-β, IL-13, and IL-10 completely inhibited MIF-mediated activation of the leishmanicidal effector function (Fig. 5). Secretion of IL-10 by T cells occurs relatively late after stimulation (47), in contrast to MIF which is released as early as 2 h after T cell activation (21). Therefore, in a further set of experiments, the time dependence of the IL-10-mediated macrophage deactivation was examined. IL-10 was most effective when added 4 h before MIF. When added simultaneously with or 4 h after MIF, the deactivating effect of IL-10 was markedly decreased. IL-10 did not exhibit any measurable inhibition of leishmania killing when PEC had been activated previously with MIF for 12 h (Fig. 5, A and B). To evaluate the role of autocrine IL-10 produced by PEC, endotoxin-free anti-IL-10 antiserum was added simultaneously with MIF to the leishmania-infected macrophages. In contrast to control preimmune serum, a significant further reduction of infected macrophages (18 ± 6%) was observed, suggesting an inhibitory influence of endogenously produced IL-10 as previously reported for other macrophage stimulatory cytokines (26).

To evaluate whether MIF is expressed after L. major infection in vivo, thereby contributing to the control of the parasites in tissues of mice, we first analyzed the expression of MIF mRNA by RNase protection assays. As depicted in Figure 6 there was a measurable up-regulation of MIF mRNA in the lymph nodes draining the site of infection, which was not detected in contralateral or lymph nodes of mice injected with PBS only (data not shown). Compared with the mRNAs of other cytokines measured simultaneously in the same RNase protection assays (e.g., IL-12 p35, IL-12 p40, IL-10, IL-1α, IL-1β, and the IL-1R antagonist), the expression of MIF mRNA was at least 7 to 10 times higher. In addition, constitutive and similar high levels of MIF mRNA were detected in the skin of mice, which reached approximately 50% of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA concentrations and did not change significantly during the course of infection (data not shown). MIF protein was also expressed by lymph node cells stimulated in vitro with concanavalin A and leishmania Ag on day 7 after infection, while earlier (e.g., days 0 and 2) or later (days 14, 21, and 35) after infection the concentrations of MIF after antigenic stimulation were below the sensitivity of the ELISA (1 ng/ml). Summarizing these findings, the expression of MIF mRNA and protein is transiently up-regulated after infection with L. major and may thus contribute to the control of the parasite in vivo.

This study provides the first evidence that MIF, originally described as a factor inhibiting the migration of leukocytes, is able to induce killing of L. major amastigotes by murine macrophages. Activation of macrophages, mediated by cytokines, is required for overcoming infection of a host with intracellular pathogens such as toxoplasma, listeria, mycobacteria, and leishmania. IFN-γ appears to be one of the principal cytokines that activate macrophages for enhanced killing of a variety of intracellular parasites, including leishmania (19). Additional cytokines reported to enhance macrophage antimicrobial activity include TNF-α (31, 48), IL-4 (49, 50), IL-3 (48), and IL-7 (30). While some of these factors act without further stimuli, the activity of others, such as IL-4 (49) and in some studies TNF-α (31), is dependent on the presence of synergistically stimulating factors, such as IFN-γ.

Only recently has pure MIF become available for study (24), and this has been shown to promote TNF-α and NO production in human murine monocytes (10). We thus became interested in studying the capacity of MIF to induce the killing of intracellular L. major amatigotes. It was shown that MIF activates murine bone marrow and peritoneally derived macrophages to inhibit the growth of and/or kill L. major amatigotes. At its peak effective concentrations of 1.25 to 2.5 μg/ml, which are similar to levels observed in vivo (35), MIF alone reduced the parasite burden of infected macrophage cultures nearly as effectively as IFN-γ, and there was an additive effect when MIF was administered together with suboptimal concentrations of IFN-γ. As in the case of other well-studied cytokines or cytokine combinations, MIF must be added to macrophages before or together with the infection (30, 31, 49). This together with our observation that NO metabolites that are produced after stimulation with MIF are crucial for the leishmanicidal effects, argues for similar antileishmanial effector pathways stimulated in the same way as by other cytokines such as IFN-γ and IL-7 (15, 30, 49, 50, 51).

As described previously for MIF-mediated TNF-α release, we found a dose-response relationship for the leishmanicidal effect of MIF that is unusual for a cytokine. Compared with its migration inhibitory effect (24) as well as the glucocorticoid antagonistic properties (14, 35), the stimulation of leishmanicidal functions required markedly higher concentrations of MIF. To date, little is known about cellular receptor(s) for MIF, and it remains to be determined whether different receptor types displaying different binding affinities are responsible for distinct thresholds for the biologic effects of MIF. Alternatively, MIF could exert some of its biologic effects in the high dose range by modifying target proteins at the cell surface via its described enzymatic activities, i.e., its functioning as tautomerase (12) or oxidoreductase (13).

Calandra et al. (6) demonstrated that MIF promotes TNF-α protein synthesis and secretion by mouse macrophages. TNF-α has been characterized in the past as a cytokine capable of enhancing the microbicidal activity of macrophages in combination with other cytokines (31, 48, 52). However, mice deficient for the p55 TNF-R, although they fail to resolve lesions caused by infection with L. major, are able to control parasite replication in vivo (39). Thus, we wanted to test the hypothesis that endogenously produced TNF-α stimulated by MIF is involved in the macrophage killing of leishmania. The antiparasitic effect of MIF could indeed be inhibited significantly by the neutralization of TNF-α as has been demonstrated previously for IL-4, IL-7, IFN-γ, and IL-3 (30, 50, 51, 53, 54), while TNF-α alone did not enhance the leishmanicidal activity of the PEC in our experimental setting (our unpublished observation). The fact that there was no detectable enhancement of leishmanicidal effector functions by MIF using cells of TNF-R p55-deficient mice confirms the experiments using anti-TNF-α Abs. In addition, these findings demonstrate the requirement of the 55-kDa TNF-R molecule for autocrine TNF-α signaling in this system. Since TNF-α added exogenously together with MIF did not further enhance the killing of leishmania amastigotes (data not shown), endogenously produced TNF-α does not appear to be limiting.

The severity of the disease in mice experimentally infected with L. major is genetically determined (19, 55). BALB/c mice are susceptible to L. major and develop inexorably progressive lesions with a uniform fatal outcome, while resistant C57BL/6 mice control the infection via the generation of protective CD4-positive Th1 cells (56, 57). To establish whether there are differences between the two strains of mice with regard to the responsiveness of their macrophages, MIF effects were compared on macrophages of BALB/c and C57BL/6 origins. As described for other stimulatory cytokines (31, 49), no differences between these two strains of mice were found. MIF exerted stimulatory effects on macrophages of both bone marrow and peritoneal origins with a comparable efficiency in both inbred strains. Despite the similar MIF responsiveness in the two mouse strains analyzed, there might very well be genetically determined differences in the production of MIF, as has recently been shown in different inbred strains of mice intratracheally injected with viable bacillus Calmette-Guérin (58) as well as in the case of irritant contact dermatitis (59). In this study we found MIF mRNA and protein to be up-regulated in the lymph nodes draining the site of L. major inoculation during the first days of infection in both C57BL/6 and BALB/c mice. However, measurements of MIF protein in the infected tissue in vivo, especially in the skin of mice, have to be performed to test whether there is a differential expression of MIF in the different strains of inbred mice. Shimizu et al. (60) and Bacher et al. (61) detected MIF mRNA and protein in the basal layer of normal uninflamed human epidermis. In line with, this we found MIF mRNA to be abundantly expressed in the skin of mice. Moreover, the release of MIF protein from the epidermis can be induced by proinflammatory stimuli such as LPS (61). Thus, MIF represents a factor preformed not only by T cells, macrophages, and pituitary cells but also by cells of the skin, the primary location of infections with parasites such as leishmania (19). Thus, the immediate release of preformed MIF without prior induction of synthesis could have important implications for the early, innate immune response after cutaneous infection with L. major. Firstly, as suggested by this study, MIF could directly and without other cytokines suppress the early local replication of the parasite. Secondly, MIF may inhibit the migration of monocytes that enter sites of inflammation as proposed from a recent study of MIF in a rat model of glomerulonephritis (16). It is therefore tempting to speculate that this may be one mechanism by which parasite-infected host cells are prevented from leaving the skin. By these two mechanisms MIF could contribute to the containment of parasites at the site of inoculation and prevent the systemic spread of L. major. Of special interest in this context, Laskay et al. have recently shown that the early containment of L. major is decisive for resistance to the infection and that the local restriction of parasites mediated by the innate immune system plays an important role in the development of a protective T cell response (62).

The stimulation of leishmanicidal effector functions described in this study together with other recently described immunomodulatory capacities of MIF, e.g., its T cell stimulatory effect (21) and its critical involvement in the delayed-type hypersensitivity reaction (1, 2, 63), will prompt us to evaluate the local expression of MIF in the skin of infected mice as well as to test the potential use of this cytokine in the experimental therapy of leishmaniasis in vivo.

We thank Drs. K. Pfeffer, H. U. Beuscher, and C. Bogdan for providing mice and antisera, respectively. The excellent technical assistance of Ms. Carmen Bauer is gratefully acknowledged.