|Home | About | Journals | Submit | Contact Us | Français|
Engagement of surface receptors contributes to the antimicrobial activity of human immune cells. We show here that infection of human monocyte-derived macrophages (MDM) with live Mycobacterium avium induced the expression of CD23 on their membrane. Subsequent cross-linking of surface CD23 by appropriate ligands induced a dose-dependent antibacterial activity of MDM and the elimination of most infected cells. The stimulation of inducible nitric oxide synthase-dependent generation of NO from MDM after CD23 activation played a major role during their anti-M. avium activity. CD23 activation also induced tumor necrosis factor alpha (TNF-α) production from MDM. Mycobacteria reduction was partially inhibited by the addition of neutralizing anti-TNF-α antibody to cell cultures without affecting NO levels, which suggested the role of this cytokine for optimal antimicrobial activity. Finally, interleukin-10, a Th2 cytokine known to downregulate CD23 pathway, is shown to decrease NO generation and mycobacteria elimination by macrophages. Therefore, (i) infection with M. avium promotes functional surface CD23 expression on human macrophages and (ii) subsequent signaling of this molecule contributes to the antimicrobial activity of these cells through an NO- and TNF-α-dependent pathway. This study reveals a new human immune response mechanism to counter mycobacterial infection involving CD23 and its related ligands.
Disseminated Mycobacterium avium complex infection is a severe complication encountered among immunocompromised patients such as patients with AIDS (3, 22). Like other pathogenic mycobacteria in humans (e.g., M. tuberculosis and M. leprae), M. avium is phagocytosed by macrophages after entry into the body (10, 23). Uptake of pathogenic mycobacteria occurs via the complement receptors CR3b and CR1 (7, 42), mannosyl-fucosyl receptor (7), fibronectin receptor (36), and/or αvβ3 integrin receptor (7). The phagosomal compartments of nonactivated macrophages containing mycobacteria failed to mature to phagolysosomes but retain the cellular markers associated with early endosomes (18, 39). This feature explains why M. avium, like M. tuberculosis, is not destroyed by nonactivated macrophages but is able to survive and multiply within their phagosomes. The mechanism(s) whereby pathogenic mycobacteria are able to preclude phagosomal maturation is not well understood, but the ability to survive and multiply within macrophages is a key component of mycobacterial pathogenesis.
Signaling through Toll-like-receptor 2 (TLR-2) appears to be essential for the induction of costimulatory molecules and the secretion of cytokines which are essential for the macrophage-mediated antimicrobial activity and antigen-presentation to T lymphocytes (41, 44). Most TLRs recruit the MyD88 adaptor molecule, which leads to the activation of NF-κB and mitogen-activated protein kinases, resulting in the induction of inflammatory mediators (45, 50). After M. avium infection, macrophages produce various effector molecules essential for infection control including tumor necrosis factor alpha (TNF-α), interleukin-12 (IL-12), and reactive oxygen intermediates (2, 46). Inducible nitric oxide synthase (iNOS)-derived NO was also reported as a powerful and necessary antimycobacterial defense mechanism in mouse models (9), and compromised handling of infection was observed in mice lacking the iNOS gene (1). Although iNOS has been detected in macrophages from disease lesions (31), it has not been possible to consistently demonstrate an NO-mediated mycobacterial killing phenomenon in human macrophages (38, 44). TLR-induced antimicrobial activity in human monocytes/macrophages was not found to be dependent on iNOS activity, nor could NO be detected in these cells (44). These data suggested that human macrophages have an NO-independent antimicrobial mechanism without excluding a role for NO in human host defense.
Upon infection, monocyte-derived macrophages (MDM) fulfill their maturation and acquire or increase their surface differentiation markers such as CD14 and CD23 (21, 44). However, whereas CD14/TLR-4 expression has been well documented (44), no data were available on CD23 expression after mycobacterial infection. In the present study, we found that infection with live M. avium subsp. avium induced CD23 expression on the surface of MDM. The role of this molecule during antimicrobial activity of macrophages was then investigated by using various CD23 agonistic ligands (35, 48). Our data indicate that surface CD23 mediates antimicrobial activity of human macrophages, in part, through the induction of NO and TNF-α production.
M. avium subsp. avium strains were obtained from the American Type Culture Collection (ATCC 35713 and ATCC 35716) and grown at 37°C in 5% CO2 in Middlebrook 7H9 medium with 10% oleic acid-albumin-dextrose complex (OADC; Difco, Sparks, MD). M. avium subsp. avium samples were serially diluted and counted by determining the optical density (OD) absorbance at 600 nm or on a bacterial hemacytometer. Before macrophage infection, bacterial suspensions were vigorously vortexed to disperse clumps and obtain a consistent OD reading, followed by dilution in phosphate-buffered saline. Heat-inactivated mycobacteria were obtained after incubation at 90°C for 10 min and quantified by determining the OD.
Peripheral blood samples pretested for the absence of human immunodeficiency virus or hepatitis B and C virus infection were obtained from healthy volunteers (Immunology Department, Pitié-Salpêtrière Hospital, Paris, France). The present study was carried out in compliance with the Helsinki Declaration. Peripheral blood-derived mononuclear leukocytes were obtained by Ficoll gradient separation and suspended in McCoy 5A modified culture medium supplemented with 100 U of penicillin/ml, 100 μg of streptomycin/ml, 25 mM HEPES, 0.1 mM 2-mercaptoethanol, 2 mM sodium pyruvate, 0.2 mM l-cysteine, 5 μg of polymyxin B/ml, and 10% fetal calf serum (FCS; all from Invitrogen, Paisley, United Kingdom). All of the above culture media, chemicals, and FCS were endotoxin-free and tested for the absence of direct activation effects on human monocytes (CD23 expression and TNF-α production as activation markers). Monocytes were subsequently separated from other leukocytes by adherence to FCS-coated culture flasks and reincubated for an additional 5 days in the same culture conditions. After these procedures, >95% of cells expressed CD14 antigen and displayed the cytochemical characteristics of MDM (48). Cells were collected and reincubated (2 × 105/ml) 24 h with live or heat-inactivated M. avium subsp. avium at a ratio of 10 bacilli per cell. Extracellular bacteria were eliminated by washing with culture medium, and cells were reincubated for an additional 24 h. This led to the infection of 50 to 70% MDM, as counted after Ziehl-Neelsen carbolfuchsin staining and microscopic examination of acid-fast bacteria (AFB). To measure MDM viability and apoptosis, externalization of membrane phosphatidylserine was analyzed by using annexin V-fluorescein isothiocyanate (FITC) and a propidium iodide kit (Immunotech, Marseille, France). For simultaneous bacteria and apoptosis staining, cells were cultured on four-well chamber slides (Lab-Tek; Nunc, Brumath, France). After exposure to various culture conditions, the slides were labeled with annexin-propidium iodide (Immunotech) and intracellular AFB staining. The percentage of apoptotic and/or infected cells was obtained after counting 200 cells in each well under fluorescence microscopy.
Cell surface expression of CD23 was measured by flow cytometry, using FITC-labeled anti-CD23 (Beckman-Coulter, Fullerton, CA). Infected MDM were directly activated through CD23 pathway, while CD23 cell surface expression was induced in uninfected MDM in the presence of 10 ng of recombinant IL-4/ml, as described elsewhere (48). CD23-positive cells were incubated in the presence of cross-linking anti-CD23 monoclonal antibody (anti-CD23 MAb; clone 135, immunoglobulin G1κ [IgG1κ], 20 μg/ml; Novartis, Geneva, Switzerland) or human IgE/anti-IgE (10 μg of each/ml; Nordic, Tilburg, The Netherlands) immune complexes as detailed elsewhere (6, 48). MDM cultures were supplemented with the chemical NO donor, SNAP (S-nitroso-N-acetylpenicillamine; Coger, Paris, France), recombinant human IL-10, TNF-α, or neutralizing anti-human TNF-α (Genzyme, Cambridge, MA).
Cell supernatants were harvested 72 h after cell activation for the quantification of TNF-α by a specific enzyme-linked immunosorbent assay (R&D, Paris, France). They were also tested for the levels of nitrites (NO2−), the stable end product of NO, using the Griess reaction modified as detailed elsewhere (25). This method gave a sensitivity limit of 0.2 μM if McCoy 5A modified culture medium was used. An inhibitory analog of l-arginine, l-NMMA (NG-monomethyl-l-arginine; Coger), was used to inhibit iNOS activity, while d-NMMA was used as a negative control.
To collect intracellular bacteria, macrophages were lysed inside culture wells by the addition of saponin (1 mg/ml, final concentration; Sigma-Aldrich, St. Louis, MO), washed, and cultured on 7H9 agar plates supplemented with OADC. This procedure took into account both the intracellular and the extracellular CFU in each well. Colony counts were performed after 3 weeks as described previously (24). The data are expressed as the mean percentage ± the standard deviation. The differences in the patterns between uninfected and M. avium subsp. avium-infected cultures were assessed by analysis of variance. The differences between other paired data were examined by the Student t test or correlation test. P values of <0.04 were considered statistically significant.
We investigated the expression of CD23 in in vitro MDM cultures with live or heat-inactivated M. avium subsp. avium. We first showed that a significant percentage of macrophages acquired CD23 surface expression after 48-h cell incubation with live M. avium subsp. avium (from 5% ± 3% in uninfected cells to 34% ± 8% after infection, P < 0.0002, Fig. Fig.1).1). In contrast, MDM incubation with heat-inactivated M. avium subsp. avium failed to promote CD23 expression (Fig. (Fig.1).1). Since CD23 mediates MDM activation, we sought to determine whether CD23 engagement led to the activation of antimicrobial activity of these cells. To induce CD23 related signaling, IgE/anti-IgE immune complexes or a cross-linking anti-CD23 antibody were used (48). After 3 days of incubation, we counted the percentage of AFB-containing infected cells and lysed macrophages to evaluate the total CFU. Upon infection, engagement of CD23 by both ligands reduced the number of infected cells in a dose-dependent manner (Fig. (Fig.2).2). Isotype-matched control (anti-CD8, IgG1κ) or anti-FcRI antibodies failed to induce antimicrobial effects. Interestingly, Fab fragments of anti-CD23, which bound CD23 without cross-linking, had no effect in this respect (Fig. (Fig.2).2). We also determined the intra- and extramacrophagic growth of the M. avium subsp. avium after CD23 ligation, and our results confirmed a reduction of bacterial CFU by this pathway (Fig. (Fig.2,2, lower panel). In addition, good correlation was found between the percentage of infected cells and the number of CFU in these cultures (correlation coefficient of 97%). Relationship between M. avium subsp. avium infection and MDM survival was also analyzed with or without CD23 cross-linking. Our data (Fig. (Fig.3)3) showed that M. avium subsp. avium infection induced clear elimination of both infected and uninfected macrophages in the same cultures. Upon CD23 activation, infected cell death was accelerated, while the number of uninfected cells remained stable (Fig. (Fig.3)3) and could be maintained in culture for 10 to 14 days, the usual survival delay for activated macrophages in vitro (48). Therefore, in addition to microbiostasis, the CD23 pathway rescues uninfected cells from M. avium subsp. avium-induced mortality.
In human macrophages, CD23 ligation is well known to promote gene expression and the secretion of various inflammatory cytokines, including TNF-α, IL-1, IL-6, and IL-8, among others (29, 35). The CD23 pathway also promotes the generation of superoxide, the transcription of iNOS gene, and subsequent production of NO from human macrophages, eosinophils, and epithelial cells (4, 6, 48). Since NO mediates microbiostatic effect, we sought to determine whether the NO pathway had a role in the inhibition of M. avium CFU after CD23 activation. The levels of nitrites, the final metabolites of NO, were quantified in cell supernatants. The data in Fig. Fig.44 indicate that CD23-mediated CFU inhibition correlated with increased NO levels produced by MDM. The addition of an iNOS inhibitor, l-NMMA, significantly reversed microbiostatic activity (P < 0.007), whereas d-NMMA had no effect. SNAP, a chemical NO donor (28), mimicked the CD23-mediated effect. The results also indicate that, upon activation, infected MDM produced higher NO levels than uninfected cells (P < 0.008, Fig. Fig.4),4), which suggests that infection promoted CD23 signaling, as has been observed after parasitic infection of human macrophages (48). These data clearly support the critical role for NO during CD23-mediated antimycobacterial activity.
We next investigated the role of TNF-α during CD23-mediated M. avium subsp. avium killing because this cytokine has been suggested as an antimycobacterial mediator (2, 17). After CD23 activation, a significant increase in TNF-α levels was observed in activated MDM cell supernatants compared to controls (from 32 ± 14 pg/ml to 588 ± 154 pg/ml, P < 0.0001, means ± the standard deviations of five distinct MDM preparations). In nonactivated MDM, exogenous TNF-α induced low but significant CFU inhibition (Fig. (Fig.5),5), whereas the simultaneous addition of neutralizing anti-TNF-α antibody during CD23 activation reduced bacteriostatic activity (P < 0.006). This suggests a role for this cytokine during CD23 engagement and may explain the inability of iNOS inhibitor to completely inhibit CD23 effect. However, anti-TNF-α failed to significantly inhibit NO generation from human macrophages, which suggested that, in infected cells, NO pathway is independent of TNF-α production. Finally, the regulation of microbiostatic effects of CD23 pathway by IL-10 was investigated. We have previously shown that this Th2 cytokine, produced during various inflammatory responses (32), downregulated CD23 pathway in human cells (5, 47). Accordingly, the addition of recombinant IL-10 during CD23 engagement significantly inhibited antimicrobial activity and reduced NO generation from human macrophages (Fig. (Fig.55).
CD23 is an activation/differentiation antigen expressed by most mature macrophages. During various parasitic infections, increased expression of this molecule has been observed. Its role in the generation of microbicidal mediators from macrophages and in intracellular and extracellular killing of pathogens has been described (29, 40, 48). Since both gamma interferon (IFN-γ) and IL-4 are able to promote CD23 gene expression, this pathway could be involved during both Th1 and Th2 responses (20, 35). We clearly show here that infection with live M. avium subsp. avium promoted surface CD23 expression in human MDM. With heat-inactivated M. avium subsp. avium, we failed to detect a CD23 increase. CD23 expression may be due to the ability of mycobacteria to activate mitogen-activated protein kinase and NF-κB, potent factors for the activation of macrophage gene expression (21). Intracellular cell signaling and cell activation via CD23 require the presence of appropriate physiologic ligands. In addition to its ability to bind IgE (FcRII), CD23 displays various other physiologic ligands such as CD21, CD11b/c, CD47-vitronectin, and mannose-containing proteins (20, 29). CD23 cross-linking by its ligands promotes the secretion of IL-1, IL-6, TNF-α, superoxide, and NO from various human cells (4, 6, 29). In the present study, we show that CD23 ligation promotes antimicrobial activity of human macrophages, in part through the generation of NO and TNF-α, two potent antimycobacterial mediators. CD23 also accelerated infected cell death (Fig. (Fig.3),3), while uninfected MDM survived and may continue their antimicrobial activity through the release of various mediators. The mechanism of cell death remains to be clarified and may be due to autophagy-dependent antimicrobial activity of macrophages, recently reported in a similar model (43). Fab fragments of anti-CD23 MAb (clone 135), which bind CD23 without cross-linkage, blocked cell activation and reversed antimicrobial effect. Therefore, the acquisition of surface CD23 and the availability of its physiological ligands may play an important role during human immune response to M. avium subsp. avium infections.
CD23-mediated NO generation seems to play a major role during the antimicrobial activity of MDM. NO is a diffusible free radical, soluble in both lipids and water, and able to highly react with oxygen and oxygen intermediates to form NO2, NO2−, NO3−, N2O3, and the highly mycobactericidal ONOO− (30). iNOS is a tightly regulated enzyme and the major generator of NO in immune cells (49). Studies investigating defense against NO derivatives in mycobacteria have demonstrated that several mycobacterial proteins are produced against reactive nitrogen intermediates in order to rescue bacteria from the immune response (11, 27). Recently, it was reported that M. tuberculosis inhibits trafficking of iNOS to the mycobacterial phagosome to escape immune response (12). In the present study we show that CD23 engagement, which mediates increased transcription of iNOS in human macrophages, induced NO generation after M. avium subsp. avium infection in a dose-dependent manner (Fig. (Fig.2).2). We consistently failed to detect spontaneous NO production after M. avium subsp. avium infection as reported previously (14). In addition to the generation of highly mycobactericidal ONOO−, NO may inhibit M. avium subsp. avium growth through its ability to reduce iron levels. We have previously shown that NO-mediated antitumoral activity is in part due to its ability to decrease intracellular ferric iron levels in target cells (15). This may explain why intracellular iron elevation increased mycobacterial resistance and human susceptibility to these infections (8).
Among cytokines, TNF-α plays a critical role in host defense against intracellular mycobacterial infections (2, 17). Researchers have reported that M. avium subsp. avium infection produces lower levels of TNF-α and NO compared to infections with nonpathogenic mycobacteria (37), contributing to the persistence of infection. CD23 engagement significantly induced TNF-α generation by infected MDM and the addition of anti-TNF-α during CD23 activation partially decreased antimicrobial activity, supporting a role for this cytokine. However, whereas TNF-α was previously suggested to be necessary for optimal NO production from human infected macrophages (13), CD23-mediated NO generation was not significantly affected after the addition of anti-TNF-α to our cultures, which suggests that early TNF-α generation following CD23 ligation (33) was not required for iNOS induction in infected macrophages. Meanwhile, the presence of both factors seems to be required for optimal antimicrobial activity. Therefore, human CD23 pathway differs from TLR-mediated antimycobacterial activity in mice (19), where the role of oxidative radicals and TNF-α is marginal.
During antigen-specific immune response, activation of infected macrophages and the control of mycobacteria multiplication is essentially dependent on IFN-γ produced by Th1 lymphocytes (26). We have previously shown that IFN-γ is a potent inducer of CD23 gene transcription and surface expression (34) and may therefore potentiate antimicrobial activity via CD23 pathway without IgE requirement. As with most Th1-like responses, IL-10 inhibits CD23-mediated activities (5, 47), and transgenic mice expressing human IL-10 showed increased susceptibility to infection with M. avium subsp. avium associated with decreased macrophage effector function and apoptosis (16). Similarly, we here show that CD23 activation and NO generation were significantly decreased after the addition of exogenous IL-10 to human macrophages. In conclusion, CD23 expression after mycobacterial infection, its engagement, and the production of antimicrobial mediators may contribute to our understanding of human immune response to these infections.
We thank M. Nacher, F. Ouaaz, and V. Jarlier for helpful scientific advice.
This study was supported by grants from the Conseil Régional d'Aquitaine and the Ministère de l'Industrie. The fund providers had no role in study design, data collection and analysis, the decision to publish, or the preparation of the manuscript.
Editor: J. L. Flynn
Published ahead of print on 5 October 2009.