The Mannose Receptor Mediates Uptake of Pathogenic and Nonpathogenic Mycobacteria and Bypasses Bactericidal Responses in Human Macrophages

Abstract

The mannose receptor (MR) is involved in the phagocytosis of pathogenic microorganisms. Here we investigated its role in the bactericidal functions of human monocyte-derived macrophages (MDMs), using (i) trimannoside-bovine serum albumin (BSA)-coated latex beads and zymosan as particulate ligands of the MR, and (ii) mannan and mannose-BSA as soluble ligands. We show that phagocytosis of mannosylated latex beads did not elicit the production of O₂⁻. Zymosan, which is composed of α-mannan and β-glucan, was internalized by the MR and a β-glucan receptor, but the production of O₂⁻ was triggered only by phagocytosis through the β-glucan receptor. Activation and translocation of Hck, a Src family tyrosine kinase located on lysosomes, has previously been used as a marker of fusion between lysosomes and phagosomes in human neutrophils. In MDMs, Hck was activated and recruited to phagosomes containing zymosan later than LAMP-1 and CD63. Phagosomes containing mannosylated latex beads fused with LAMP-1 and CD63 vesicles but not with the Hck compartment, and the kinase was not activated. We also demonstrate that the MR was unable to distinguish between nonpathogenic and pathogenic mycobacteria, as they were internalized at similar rates by this receptor, indicating that this route of entry cannot be considered as a differential determinant of the intracellular fate of mycobacteria. In conclusion, MR-dependent phagocytosis is coupled neither to the activation of NADPH oxidase nor to the maturation of phagosomes until fusion with the Hck compartment and therefore constitutes a safe portal of entry for microorganisms.

Pathogenic microorganisms have developed several strategies to circumvent microbicidal responses of host cells. It has been postulated that phagocytosis through receptors that by-pass the bactericidal activity of macrophages may provide an opportunity for pathogens to manipulate the host environment to their own advantage (12, 15). For example, the entry through CR1 prevents both the fusion of lysosomes with phagosomes containing Salmonella typhimurium (21) and the production of O₂⁻ in macrophages infected by Leishmania major (8). Furthermore, when human macrophages directly recognize and internalize Mycobacterium tuberculosis, the organism resides in a phagosomal compartment that resists fusion with lysosomes. Interestingly, the fusogenic properties of these phagosomes are restored when M. tuberculosis are serum opsonized (1).

Among the macrophage receptors used by microorganisms, considerable attention has focused on the mannose receptor (MR). This receptor recognizes glycosylated molecules with terminal mannose, fucose, or N-acetylglucosamine moieties and efficiently internalizes soluble and particulate ligands through the endocytic and phagocytic pathways, respectively (38). Recognition of glycoconjugates on the surface of microorganisms by the MR leads to the phagocytosis of several pathogens such as Candida albicans (14), Leishmania donovani (53), and Pneumocystis carinii (13). In human monocyte-derived macrophages (MDMs), the MR is involved in the nonopsonic phagocytosis of the virulent M. tuberculosis strains H37Rv and Erdman but not of the avirulent strain H37Ra (45). This observation suggests that the MR expressed in human MDMs may constitute a safe route of entry for pathogenic mycobacteria. However, the ability of the MR to influence the microbicidal responses of these cells remains to be elucidated. In murine or rabbit macrophages, several studies support the idea that the MR is coupled to bactericidal functions, as it triggers the secretion of lysosomal enzymes (5, 35) and the production of O₂⁻ (4, 23) and cytokines (48, 56). However, it is well known that mycobacterial virulence differs from one animal to another, complicating the extrapolation of results from mice to humans (36).

The present study was therefore undertaken to examine the role of the MR in the regulation of bactericidal functions in human MDMs. Macrophages participate in host defense by several mechanisms which lead to the destruction of invading microorganisms: (i) attachment of the particles mediated by cell surface receptors and internalization into phagosomes, (ii) fusion of phagosomes with cytoplasmic vesicles (endosomes and lysosomes) containing proteases, and (iii) activation of the respiratory burst enzyme NADPH oxidase, which generates superoxides (O₂⁻), the source of secondary reactive oxygen products. In the absence of particulate ligands binding specifically to the MR, it is difficult to study the role of the MR in the bactericidal responses. Therefore, we have synthesized a polymer of mannose which has a higher affinity for the MR than monomers (22). The polymer was covalently linked to bovine serum albumin (BSA) adsorbed onto latex beads. Since one of our objectives was to determine if the human MR is also able to distinguish between pathogenic and nonpathogenic strains, we used two nonpathogenic species, M. phlei and M. smegmatis, and an opportunistic pathogen, M. kansasii (20). The results presented in this paper support a role of the human MR in the process of phagocytosis that by-passes the production of O₂⁻ and delays the maturation of phagosome along the endosomal-lysosomal pathway. Furthermore, the MR efficiently participates in the internalization of both pathogenic and nonpathogenic strains, indicating that it is unable to distinguish between mycobacteria on the basis of their pathogenicity.

MATERIALS AND METHODS

Materials.

RPMI 1640 with Glutamax, Ficoll-Hypaque, phenylmethylsulfonyl fluoride, penicillin, and streptomycin were purchased from Eurobio (Les Ulis, France). Phenylarsine oxide was obtained from Aldrich-Chemie (Steinheim, Germany). Minimum essential medium (MEM), fetal calf serum, and HEPES were from GIBCO (Cergy Pontoise, France). Polystyrene microspheres were from Polysciences Inc. (Warrington, Pa.). A glycoconjugate of 20 to 24 mol of mannose/mol of albumin (Man-BSA) and other chemicals were obtained from Sigma (St. Louis, Mo.). All chemicals used for electron microscopy were obtained from Roth (Karlsruhe, Germany).

The following antibodies (Abs) were used: Goat polyclonal Abs against the human MR, a generous gift of P. D. Stahl (St. Louis); mouse monoclonal Ab (MAb) RC20, directed against phosphotyrosine and linked to horseradish peroxidase, purchased from Transduction Laboratories (Lexington, Ky.); mouse MAb anti-human LAMP-1 H4A3, kindly provided by T. Levade (Toulouse, France); mouse anti-human CD63 MAb, purchased from CLB (Amsterdam, The Netherlands); affinity-purified rabbit anti-Hck Abs, obtained from Santa Cruz Biotechnology, Inc. (Tebu, France); and rabbit anti-Hck antiserum generated against a peptide corresponding to the N-terminal amino acid residues 38 to 52, previously characterized (32). Secondary Abs were purchased from Sigma.

Human MDMs.

Leukocytes from healthy donors were isolated by dextran sedimentation and centrifugation through Ficoll-Hypaque as previously described (28). The interface band containing mononuclear cells was removed and washed twice with ice-cold phosphate-buffered saline (PBS), and the cell pellet was resuspended in cold RPMI supplemented with penicillin-streptomycin (100 U/ml). Cells were then counted in the presence of trypan blue, adjusted to 5 × 10⁶ viable cells/ml, distributed in 1-ml aliquots in 24-well tissue culture plates, and incubated for 1 h at 37°C to allow adherence of monocytes. For phagocytosis experiments, monocytes were plated on sterile glass coverslips. The nonadherent cells were removed, and the cell monolayer was washed vigorously twice with warm RPMI. The cells were then placed in RPMI containing 10% heat-inactivated fetal calf serum and antibiotics and maintained in culture in 5% CO₂ for 6 to 7 days. The culture medium was renewed on the third day. The number of adherent cells per well averaged (5.53 ± 1.1) × 10⁵ (n = 14). Before stimulation, MDMs were washed twice with MEM containing 10 mM HEPES (MEM-HEPES, pH 7.4) and equilibrated for 20 min at 37°C in the same medium.

Bacterial strains and growth conditions.

M. phlei (ATCC 11758), M. smegmatis (ATCC 607), and M. kansasii (ATCC 124478) were grown and isolated as previously described (34). Briefly, after 3 days (for M. phlei and M. smegmatis) or 3 weeks (for M. kansasii) of growth, the culture medium was discarded, and the bacterial pellicle was disrupted by gentle shaking with glass beads (4-mm diameter) in PBS (pH 7.4), centrifuged, and resuspended in PBS (pH 7.4). Large clumps were sedimented by centrifugation for 10 min at 200 × g. The number of bacilli in the supernatant was counted in a Thoma chamber and adjusted to the required concentration. Up to 90% of mycobacteria in the supernatant were individualized; the remaining formed small aggregates containing two or three bacilli. The number of viable mycobacteria assessed by serial dilutions and plating on culture medium averaged 85%.

Synthesis of trimannoside-BSA.

The trimannoside (methoxycarbonyl octyl 2-O-[(α-d-mannopyranosyl)-2-O-(α-d-mannopyranosyl]-(α-d-mannopyranoside) was produced by successive additions of mannopyranosidic units with complete stereocontrol of the glycosylation steps (37) and then conjugated to BSA by the acyl azide method (6). The neoglycoprotein was then dialyzed three times against water and lyophilized. Carbohydrate analysis was carried out on Dionex after acid hydrolysis and gave a ratio of 11 to 12 mol of trimannoside per mol of protein.

Immunoblots.

For protein tyrosine phosphorylation, MDMs were treated for 5 min with 100 μM vanadate before addition of the stimulus. Then the reaction was stopped by addition of 1 volume of ice-cold MEM-HEPES containing 1 mM vanadate, 1 mM phenylmethylsulfonyl fluoride, and 20 μM phenylarsine oxide. The supernatant was discarded, and adherent cells were lysed in boiling Laemmli sample buffer. Proteins were solubilized in boiling Laemmli sample buffer, separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to nitrocellulose membranes as previously described (32). For detection of the MR, proteins from MDMs were solubilized at different times of cell culture, separated, and transferred to nitrocellulose membranes as described above. Nitrocellulose membranes were immunoblotted with Abs against the MR (1:200 dilution) or against phosphotyrosine (1:10,000) and visualized by using an enhanced chemiluminescence system.

Infection of adherent macrophages and phagocytosis assay.

First, particles used for phagocytic assays were prepared. Mycobacteria were stained with fluorescein isothiocyanate (FITC) as previously described (34). Briefly, 10⁹ bacteria were added to 1 ml of 0.01% FITC in 0.2 M Na₂CO3–NaHCO₃ buffer containing 150 mM NaCl (pH 9.2) for 10 min. The bacteria were then washed twice in PBS (pH 7.4). In some experiments, mycobacteria and zymosan were serum opsonized by incubation in pooled human sera as previously described (34). Trimannoside-BSA was adsorbed on microparticles as recommended by the manufacturer. Briefly, 2 × 10⁹ 1-μm-diameter polystyrene microspheres were washed twice in 0.1 M borate buffer (pH 8.5) and incubated overnight at room temperature with 1 mg of trimannoside-BSA per ml. The beads were then washed twice, incubated for 30 min in borate buffer containing 10 mg of BSA per ml, resuspended in PBS containing 10 mg of BSA per ml, 0.1% NaN₃, and 5% glycerol, and stored at 4°C for less than 1 month. Using FITC-Man-BSA and fluorescence-activated cell sorting analysis, we observed that beads were homogeneously coated.

Second, MDMs that had engulfed particles were washed twice with MEM-HEPES to remove unbound particles and fixed. To visualize phagosomes containing zymosan or latex beads, MDMs were fixed and permeabilized in methanol for 6 min at −20°C, washed in PBS containing 0.1% Tween 20, and stained successively with MAbs against human LAMP-1 (1:50 dilution) (58) and rhodamine-conjugated secondary Abs. In some experiments, double staining was performed with LAMP-1 or CD63 Abs (1:100 dilution) and affinity-purified rabbit anti-human Hck Abs (1:200) (32). Phagocytosis of FITC-stained bacteria was determined as previously described (34). Briefly, after cell fixation with 3.7% paraformaldehyde for 20 min, MDMs were viewed in the presence of trypan blue (5 mg/ml) to quench the fluorescence of bacteria remaining in the extracellular medium (34). Intracellular fluorescent mycobacteria were then counted by alternatively viewing the cells by phase-contrast and fluorescence microscopy. MDMs containing at least one fluorescent phagosome or one FITC-stained mycobacterium were counted among approximately 100 cells in duplicate samples.

Electron microscopy.

After the standard phagocytosis assay was carried out, MDMs were washed and fixed with cold fixative medium at 4°C and prepared for electron microscopy as previously described (41).

Assay for O₂⁻ generation.

The production of O₂⁻ was determined spectrophotometrically by measuring the superoxide dismutase-inhibitable reduction of ferricytochrome c as previously described (26).

Assay for β-glucuronidase release in the extracellular medium.

Lysosomal exocytosis was evaluated by measuring the release of β-glucuronidase as previously described (19). Extracellular media of stimulated and nonstimulated MDMs (2 × 10⁶/ml) were centrifuged to eliminate the few floating cells. The cell pellet and the adherent cells were pooled and lysed overnight in 0.5% Triton X-100. Enzyme activity was measured at 405 nm in solubilized pellets and supernatants. The release of β-glucuronidase was expressed as the percentage of the total cell content.

Hck tyrosine kinase assay.

Hck solubilization, immunoprecipitation, and kinase assay were performed as previously described (32, 52). Briefly, Hck was solubilized from 2.5 × 10⁶ MDMs with a buffer containing 2% Nonidet P-40, conventional protease inhibitors, and cytosol from NB4 cells to further avoid proteolysis (32, 52). The kinase was then immunoprecipitated and assayed for in vitro protein tyrosine kinase activity in the presence of acid-denatured rabbit muscle enolase (as an exogenous substrate) and 10 μCi of [³²P]ATP (6,000 mCi/mmol) (32). Proteins were then separated by SDS-PAGE (8% gel), and Hck-dependent phosphorylation of enolase was quantified by using the Image QuaNT program on a Molecular Dynamics Storm840 imager (34).

RESULTS

MR-dependent phagocytosis does not trigger the O₂⁻ generation in MDMs.

First, we verified that the MR was expressed during the differentiation of human monocytes into macrophages. As shown in Fig. 1A, the receptor was detected by day 1 and increased thereafter. Zymosan, which is commonly used as a particulate ligand of the MR (38, 49), was efficiently internalized by MDMs and consistently induced the production of O₂⁻ (Fig. 1). To analyze more precisely the role of the MR in these responses, soluble forms of mannosylated molecules, mannan and Man-BSA, were used as competitive inhibitors (46). When cells were preincubated for 10 min with mannose-BSA (200 μg/ml) or mannan (3 mg/ml), internalization of zymosan was inhibited by more than 40% (Fig. 1B). Increasing the concentration of soluble ligands did not further inhibit phagocytosis (data not shown). However, none of these compounds modified the O₂⁻ generation associated with the ingestion of zymosan (Fig. 1C), suggesting that zymosan did not trigger the generation of O₂⁻ through the MR.

FIG. 1.

The MR and a laminarin-sensitive β-glucan site are both involved in the phagocytosis of zymosan by human MDMs, but only phagocytosis through the β-glucan site triggers O₂⁻ production. (A) Expression of the MR increases during differentiation of human monocytes into macrophages. At different days of culture, the presence of the MR in the cells (5 × 10⁵ cells/lane) was determined by immunoblotting. Positions of M_r standards are shown on the left in thousands. (B and C) MDMs were incubated for 10 min in the presence or the absence of inhibitory concentrations of soluble ligands (200 μg/ml for Man-BSA, 3 mg/ml for mannan, and 800 μg/ml for laminarin) and then challenged with zymosan (particle/cell ratio of 50:1) for 30 min at 37°C. MDMs were fixed with methanol, and phagosomal membranes were stained with mouse anti-human LAMP-1 and FITC-conjugated anti-mouse immunoglobulin G. The percentage of cells containing at least one labeled phagosome was determined by fluorescence microscopy (B); O₂⁻ production was measured by the superoxide-inhibitable reduction of ferricytochrome c. Data are corrected for the basal O₂⁻ production, which never exceeded 10 nmol/10⁶ cells (C). None of the soluble ligands added to MDMs affected O₂⁻ production by themselves. Results are expressed as means ± SEM of three to eight experiments. ∗, P < 0.05; ∗∗, P < 0.01 compared with the control by unpaired Student’s t test.

Since zymosan is composed of β-glucan in addition to α-mannan (10), the effect of laminarin, a β-glucan-soluble polymer, was studied. When laminarin was added to MDMs 10 min before zymosan, both phagocytosis and production of O₂⁻ decreased (Fig. 1B and C). Increasing laminarin concentrations up to 1.5 mg/ml or prolonging the time of incubation did not improve the inhibitory effect (data not shown). When used in combination, Man-BSA and laminarin had additive inhibitory effects on the phagocytosis of zymosan (Fig. 1B): 41 and 46% inhibition, respectively, when used alone, and 68% inhibition when used in combination. In contrast, this did not potentiate the inhibitory effect of laminarin on the production of O₂⁻ (Fig. 1C). Therefore, it is likely that zymosan was internalized into MDMs at least through the MR and a laminarin-sensitive β-glucan site, but only phagocytosis through the β-glucan receptor triggered the production of O₂⁻.

To confirm that the MR did not trigger the production of O₂⁻, we used another particulate ligand of the receptor, latex beads coated with trimannoside-BSA. The internalization of these beads (Fig. 2D) was 78% inhibited by preincubating MDMs with 1 mg of trimannoside-BSA per ml (from 33% ± 13% to 7% ± 2% [mean ± standard error of the mean {SEM}] of cells having engulfed particles; n = 2), thereby indicating that mannosylated latex beads are a convenient particulate ligand of the MR. Under these conditions, no generation of O₂⁻ was detected (Table 1), even when the number of particles per cell was increased to 200 (data not shown).

FIG. 2.

MR-dependent phagocytosis induces the translocation of LAMP-1 but not Hck to the phagosomal membrane. Indirect immunofluorescence microscopy was performed on doubly stained MDMs with anti-LAMP-1 and anti-Hck antibodies. (A) Control cells. Adherent MDMs were incubated at 4°C for 30 min with zymosan (50 particles per cell), then washed free of nonadherent zymosan, and incubated at 37°C for 15 min (B) or 30 min (C). MDMs were incubated for 60 min at 37°C with trimannoside-BSA latex beads (50 particles per cell (D). The arrow points to a Hck-positive, LAMP-1-negative phagosome. (E and F) Double immunostaining with anti-CD63 and anti-Hck antibodies of control cells (E) and cells with zymosan internalized for 30 min (F). When second antibodies were used alone, no fluorescence was observed.

TABLE 1.

The MR does not regulate O₂⁻ production in human MDMs^a

Treatmenta	Conditions	Mean O2− production (nmol/106 cells) ± SEMb
BSA latex beads	PBS	0.34 ± 0.12 (3c)
Trimannoside-BSA latex beads	PBS	0.92 ± 0.21 (2)
Opsonized zymosan	PBS	19.75 ± 2.10 (3)
Opsonized zymosan	Man-BSA	22.07 ± 3.15 (3)
PMA	PBS	10.70 ± 3.00 (4)
PHA	Mannan	12.07 ± 3.12 (4)

^aMDMs were exposed to BSA-latex beads or trimannoside-BSA latex beads (50 particles per cell) for 60 min, to opsonized zymosan (50 particles per cell) for 30 min, or to PMA (1 μg/ml) for 30 min. In some experiments, MDMs were preincubated with mannan (3 mg/ml) or Man-BSA (200 μg/ml). 

^bO₂⁻ generation over the basal production. 

^cNumber of independent experiments. 

Next, we examined if the MR could regulate the production of O₂⁻ elicited by well-known particulate or soluble NADPH oxidase stimuli such as opsonized zymosan or phorbol myristate acetate (PMA) (26, 30). As shown in Table 1, neither Man-BSA nor mannan affected the net O₂⁻ production evoked by opsonized zymosan or PMA, respectively. Taken together, these data show that MR-dependent phagocytosis neither stimulated nor regulated the generation of O₂⁻ in human MDMs.

Exocytosis of the lysosomal enzyme β-glucuronidase is not triggered during phagocytosis.

In addition to the production of O₂⁻, secretion of lysosomal enzymes constitutes a major bactericidal response of phagocytes. Although exocytosis of lysosomal enzymes during phagocytosis has been found in murine macrophages (5), no data on the release of lysosomal enzymes by human macrophages are available. Therefore, we first determined whether such a process occurred in human MDMs by measuring the release of the lysosomal enzyme β-glucuronidase in the extracellular medium. We showed that phagocytosis of zymosan at a particle/cell ratio of 50:1 did not enhance the release of β-glucuronidase into the medium compared to control cells (22.2% ± 4.5% versus 20.4% ± 3.9% of cell content; n = 4). Similar results were obtained with trimannoside-BSA latex beads. Also, the release of β-glucuronidase was not triggered by serum-opsonized zymosan (data not shown), a potent lysosomal secretagogue in human neutrophils (34).

MR-dependent phagocytosis does not induce translocation of Hck on phagosomes.

Another approach was undertaken to determine the role of the MR in the process of phagosomal maturation by following the acquisition of endosomal-lysosomal markers by phagosomes. In the absence of an exclusive marker of lysosomes, we investigated whether Hck could be a reliable marker of this compartment. Hck is a tyrosine kinase of the Src family specifically expressed in phagocytes (39, 57). It is associated with the membrane of lysosomal granules in human neutrophils and recruited to the phagosomal membrane during the biogenesis of phagolysosomes (32, 34, 51). Similar subcellular localization and recruitment to the phagosomal membranes have been observed in HL60 cells differentiated into macrophages (32). In MDMs, Hck is associated with vesicular structures in the cytoplasm (Fig. 2A and E), recruited to phagosomal membranes in cells having engulfed zymosan particles (Fig. 2C), and activated (Fig. 3A). To determine the identity of the Hck-positive vesicular compartment, double immunostaining of Hck and LAMP-1, a protein mainly associated with the membrane of late endosomes, was performed. As shown in Fig. 2A, Hck and LAMP-1 poorly colocalized in resting cells. Similar experiments using CD63, a marker of late endosomes and lysosomes (31), showed that Hck was not on this compartment either (Fig. 2E). Synchronized phagocytosis of zymosan was then performed to examine whether vesicles bearing Hck or LAMP-1 fuse with phagosomes at different stages of maturation. After 15 min, phagosomes containing zymosan were stained poorly or not at all by anti-Hck antibodies whereas LAMP-1 was clearly present (Fig. 2B). After 30 min, both markers were located on phagosomes (Fig. 2C) and occasionally did not colocalize (Fig. 2C). Similar results were obtained at 60 min (data not shown). Therefore, Hck is associated with vesicles which fused with phagosomes containing zymosan after LAMP-1 vesicles. Similarly, CD63 and Hck were not always delivered to the same phagosomes (Fig. 2F). Taken together, these data indicate that (i) the Hck-positive compartment is probably distinct from LAMP-1 and CD63 compartments and (ii) Hck may serve as a marker of a phagosomal maturation event which occurs later than LAMP-1. Phagosome maturation stops at the early endososomal stage when latex beads are internalized by MDMs (9), whereas it proceeds to the late endosomal-early lysosomal stage if beads are coated with proteins (9, 40). To determine the stage of phagosome maturation when the MR is involved in phagocytosis, beads were coated with trimannoside-BSA. We observed that Hck did not translocate to phagosomes containing mannosylated latex beads (Fig. 2D) whereas LAMP-1 (Fig. 2D) and CD63 (not shown) stained the phagocytic vacuoles, indicating that maturation of these phagosomes stops before fusion with Hck-positive vesicles.

FIG. 3.

MR-dependent phagocytosis does not stimulate Hck tyrosine kinase activity. MDMs were stimulated with mannosylated particles (50 particles per cell) for the indicated periods of time: zymosan in the presence or the absence of Man-BSA (200 μg/ml) (A) or latex beads coated with trimannoside-BSA (B). Hck was then solubilized (2.5 × 10⁶ cell equivalents per assay) and immunoprecipitated, and phosphorylation of the exogenous substrate enolase was quantified. The histogram shows the amount of ³²P incorporated into enolase. Data are means ± SEM of four experiments. ∗, P < 0.05 compared with control by paired Student’s t test.

In neutrophils, activation of Hck has been correlated with its translocation toward phagosomes (32, 34, 51). In MDMs, zymosan internalization markedly increased the Hck kinase activity after 5 min, reaching a maximal value after 10 min and decreasing thereafter (Fig. 3A). Treatment of MDMs by Man-BSA prior to ingestion of zymosan did not interfere with the activation of Hck. Thus, zymosan triggered activation of Hck in an MR-independent pathway. In contrast, preincubation of MDMs with 800 μg of laminarin per ml for 10 min before addition of zymosan significantly decreased the kinase activity tested at time points coinciding with peaks of Hck activation (1.57 ± 0.58-fold versus 2.24 ± 0.68-fold over basal activity; n = 4, P = 0.0091 [paired Student’s t test]), indicating that activation of Hck is mediated by a laminarin-sensitive receptor. Consistent with these results, trimannoside-BSA coated latex beads did not trigger the activation of Hck (Fig. 3B).

The MR does not trigger phosphorylation of tyrosine residues.

Several data indicate that tyrosine kinases are involved in the MR signal transduction: binding of mannosylated proteins to the MR increases protein tyrosine phosphorylation in murine macrophages (33), and mutagenesis of the single tyrosine residue located in the cytoplasmic domain of the MR decreases the endocytosis of soluble ligands in transfected Cos cells (25). However, in human MDMs, Man-BSA did not affect the phosphorylation profile compared to control cells (Fig. 4, lanes 1 and 2). In addition, it has been reported that phagocytosis of zymosan increases the level of protein tyrosine phosphorylation in murine macrophages (18). Human MDMs incubated with zymosan for 30 min showed a marked increase in the pattern of tyrosine phosphorylation (lane 3) compared to control cells (lane 1). Pretreatment with soluble Man-BSA before the addition of zymosan (lane 4) did not change the phosphorylation profile induced by zymosan alone (lane 3), while phagocytosis was inhibited (Fig. 1B). Furthermore, ingestion of trimannoside-BSA latex beads after 10 or 30 min (data not shown) or 60 min (lane 6) did not modify the pattern of protein tyrosine phosphorylation compared to control MDMs (lane 1). Taken together, these results indicate that phagocytosis through the human MR is not associated with protein phosphorylation of tyrosine residues. In contrast, preincubation with laminarin markedly decreased zymosan-stimulated tyrosine phosphorylation (lane 5), thereby indicating that signal transduction of the laminarin-sensitive β-glucan receptor involved activation of tyrosine kinases.

FIG. 4.

Phagocytosis through the MR does not stimulate protein tyrosine phosphorylation in MDMs. Cells were preincubated for 10 min with medium (lanes 1, 3, and 6), Man-BSA (200 μg/ml; lanes 2, 4) or laminarin (800 μg/ml; lane 5) and then exposed to particles (50 per cell): zymosan for 30 min (lanes 3 to 5) or latex beads coated with trimannoside-BSA for 60 min (lane 6). Proteins (5 × 10⁵ cell equivalents/lane) were immunoblotted with antiphosphotyrosine as described in Materials and Methods. Positions of molecular markers are shown on the left in kilodaltons. Results of one representative experiment out of three are shown.

The MR mediates phagocytosis of both pathogenic and nonpathogenic mycobacteria.

Human MDMs that had ingested pathogenic (M. kansasii) and nonpathogenic (M. smegmatis and M. phlei) mycobacteria were examined by electron microscopy. As previously observed with nonpathogenic mycobacteria (2) and zymosan (42), M. smegmatis was inside either single-bacterium phagosomes (tight fitting or spacious) or multibacterial phagosomes (Fig. 5). M. phlei and M. kansasii were located in the same types of vacuoles (data not shown). When we explored the role of the MR in the phagocytosis of mycobacteria, especially its ability to internalize nonpathogenic mycobacteria, we observed that mannosylated ligands such as Man-BSA decreased the uptake of M. kansasii, M. phlei, and M. smegmatis to similar extents (43, 45, and 34%, respectively) (Fig. 6). The possibility that Man-BSA bound to mycobacteria, thereby blocking their entry by a mechanism other than the MR, was ruled out by showing that pretreatment of mycobacteria with Man-BSA did not affect their uptake compared to untreated mycobacteria (data not shown). These results indicate that the MR is involved in the recognition and phagocytosis of both pathogenic and nonpathogenic mycobacteria. In contrast, pretreatment of MDMs with laminarin (800 μg/ml) did not affect the ingestion of either M. kansasii, M. phlei, or M. smegmatis (Fig. 6), suggesting that the laminarin-sensitive β-glucan site did not contribute to the entry of these mycobacteria.

FIG. 5.

Electron microscopy of human MDMs after infection with M. smegmatis. MDMs were incubated with M. smegmatis (50 bacilli per cell) for 60 min, washed, and fixed. Representative cross sections of MDMs show bacilli in separate (spacious and tight) and joint phagosomes. Black arrows in panels B and D indicate multibacterial vacuoles; the arrow in panel A indicates a spacious single-bacterium phagosome; white arrows in panels B and C show tight single-bacterium phagosomes. Magnifications: A, ×14,000 (1 cm = 0.7 μm); B to D, ×17,000 (1 cm = 0.6 μm).

FIG. 6.

Man-BSA but not laminarin markedly decreases the phagocytosis of both pathogenic and nonpathogenic mycobacteria by MDMs. Cells were preincubated for 10 min with either the medium or soluble ligands (200 μg of Man-BSA or 800 μg of laminarin per ml) and then exposed to FITC-stained mycobacteria (50 per cell) for 60 min. The percentage of MDMs containing at least one bacillus was determined by fluorescence microscopy. Data are expressed as mean ± standard deviation of three to six independent experiments. ∗, P < 0.05; ∗∗, P < 0.01 for MDMs preincubated with Man-BSA compared to nontreated cells by paired Student’s t test.

Finally, we studied whether mycobacteria elicit the production of O₂⁻. Compared to zymosan, neither M. phlei, M. smegmatis, nor M. kansasii affected the basal generation of O₂⁻ (Table 2) even when the bacterium/cell ratio was increased from 50 to 200 (data not shown). Phagocytosis of mycobacteria opsonized in pooled human sera activated the generation of O₂⁻ (Fig. 7), suggesting that the failure of mycobacteria to stimulate the production of O₂⁻ might be associated with a serum-independent phagocytic process.

TABLE 2.

Nonopsonic phagocytosis of pathogenic and nonpathogenic mycobacteria does not stimulate the production of O₂⁻ in MDMs

Treatmenta	Mean ± SEM (no. of independent experiments)
	Phagocytic cells (%)b	O2− production (nmol/106 cells)c
Zymosan	36 ± 3 (11)	17.77 ± 5.44 (8)
M. phlei	36 ± 3 (8)	−1.75 ± 2.63 (8)
M. smegmatis	33 ± 4 (11)	0.33 ± 2.12 (6)
M. kansasii	32 ± 3 (11)	1.66 ± 3.19 (3)

^aMDMs were exposed to zymosan for 30 min or mycobacteria for 60 min at a particle-to-cell ratio of 50:1. 

^bPercentage of cells having internalized at least one particle. 

^cO₂⁻ generation over the basal production. 

FIG. 7.

Serum opsonization of mycobacteria stimulates the generation of O₂⁻ in MDMs. Cells were incubated in the presence or absence (control) of serum-opsonized (+) and nonopsonized (−) zymosan for 30 min or with mycobacteria for 60 min at a particle-to-cell ratio of 50:1. O₂⁻ production was measured by the superoxide-inhibitable reduction of ferricytochrome c. Data are means ± SEM of three independent experiments. ∗, P < 0.05, ∗∗, P < 0.01 compared to control by paired Student’s t test.

DISCUSSION

The objective of this work was to determine whether the entry through the MR elicits bactericidal responses in human MDMs and the extent to which the MR can distinguish between pathogenic and nonpathogenic mycobacteria.

Cell responses triggered by the MR are difficult to evaluate because of the contamination of soluble ligands by other carbohydrates; e.g., traces of β-glucan have been detected in mannan (17). Furthermore, zymosan, which is commonly used as a particulate ligand of the MR, also binds to a β-glucan receptor. We show here that the bactericidal responses triggered by zymosan result from its interaction with a β-glucan receptor rather than the MR. Therefore, to avoid potential misinterpretation, we used several soluble and particulate ligands of the MR.

We show that phagocytosis through the MR does not trigger the generation of O₂⁻ and that soluble ligands do not modulate the O₂⁻ production elicited by PMA or opsonized zymosan in human MDMs. Therefore, internalization of particles through the MR represents a new example of dissociation between phagocytosis and activation of the NADPH oxidase (54, 55). Comparison of these results to those previously obtained for murine or rabbit macrophages reveals interspecies differences since binding of mannosylated proteins has been shown to cause a significant increase in the oxygen consumption in rat macrophages (23) and modulate the O₂⁻ production induced by PMA in mouse macrophages (16). Furthermore, in rabbit macrophages, soluble ligands of the MR have been reported to stimulate the release of β-glucuronidase and hexosaminidase (35), and in murine macrophages, a correlation has been observed between lysosomal enzyme secretion and MR expression (11, 47). Although the release of β-glucuronidase in the extracellular medium of control human MDMs was of the same magnitude as in rat macrophages (47), no increase in the secretion of β-glucuronidase in response to the conventional stimulating agent, serum-opsonized zymosan, or to MR particulate ligands was observed. These results, together with the lack of published data showing regulated exocytosis of lysosomal enzymes in human macrophages, led us to question the existence of a fusion process between lysosomes and the plasma membrane in human MDMs. Consequently, another approach based on analysis of phagosome maturation along the endocytic pathway was undertaken. For this purpose, two markers of the late endosomal-lysosomal compartments, LAMP-1 and CD63 (24, 31), were used in addition to Hck, a marker of lysosomes in human neutrophils (32, 51). In resting MDMs, CD63 and LAMP-1 did not colocalize with Hck. When MDMs internalized zymosan, Hck translocated to the membrane of phagosomes and was activated. Under these conditions, Hck frequently colocalized with LAMP-1 and CD63. The kinetics of translocation of Hck differed from that of LAMP-1 which became associated with phagosomes earlier than Hck, indicating that LAMP-1 and Hck are localized on distinct organelles. Most likely LAMP-1 is mainly associated with the late endocytic compartment (24, 44) and Hck is found primarily in a compartment which might be the lysosomes (32), although we could not definitively identify it. Reevaluation of the localization of vesicle markers will be required, as previously suggested by Russell et al. (44), for unequivocal identification of the vesicles fusing with phagosomes.

In light of our results, we propose that Hck might serve as a useful marker of a phagosomal maturation stage occurring after fusion with LAMP-1 vesicles. We demonstrate that (i) MR-dependent phagocytosis allowed the fusion of phagosomes with LAMP-1- but not with Hck-carrying vesicles and (ii) the MR did not trigger activation of Hck. Therefore, entry of particles via the MR did not allow the maturation of phagosomes to proceed to fusion with Hck-positive vesicles.

Although it was thought that a closely juxtaposed membrane of the phagosome around mycobacteria is associated with virulence, it has been observed that M. smegmatis enclosed in tight phagosomes can be killed (2). We observed that both nonpathogenic and pathogenic mycobacteria are enclosed in tight phagosomes containing a single bacterium and also in spacious phagosomes and multibacterial vacuoles, indicating that the morphology of the phagosomes does not reflect the pathogenicity of the bacilli.

We also show that the MR is involved in the uptake both of pathogenic mycobacteria and of nonpathogenic mycobacteria such as M. phlei and M. smegmatis and, generally speaking, in the internalization of mannosylated particles (i.e., zymosan and trimannoside-BSA latex beads). In light of these results, we propose that the human MR internalized mycobacteria without discrimination between pathogenic and nonpathogenic species and cannot be considered a differential determinant of the intracellular fate of mycobacteria, at least for the earliest steps of infection. The finding that the avirulent substrain M. tuberculosis H37Ra is not recognized by the MR (45) may suggest that in contrast to most other mycobacterial species (7, 12), H37Ra does not present mannose-containing glycoconjugates at the surface. Alveolar macrophages which are targeted in vivo by mycobacteria are, to some extent, different from MDMs (50). Therefore, the inability of the MR to distinguish between nonpathogenic and pathogenic mycobacteria must also be examined in these cells.

In addition, we demonstrate that nonopsonic phagocytosis of pathogenic and nonpathogenic mycobacteria did not trigger the O₂⁻ production even when the number of ingested particles was increased. This could not be attributed to an O₂⁻ scavenger effect since M. phlei and M. kansasii did not affect the O₂⁻ production elicited by PMA in the macrophage cell line U937 (27). Interestingly, the laminarin-sensitive β-glucan site, which participates in the internalization of zymosan and triggers the production of O₂⁻, is not used by the mycobacterial species examined by us. Phagocytosis of mycobacteria by human MDMs does not trigger the production of O₂⁻ and is only 40% inhibited by Man-BSA, which suggests that, in addition to the MR, other receptors not coupled to the O₂⁻-generating enzyme NADPH oxidase participate in the internalization of mycobacteria. It is possible that nonopsonic internalization of mycobacteria would also by-pass lysosomal enzyme release, although published data concerning the fusion of lysosomes with phagosomes containing nonpathogenic mycobacteria are lacking for human MDMs. In support of this hypothesis, we have recently observed that in human neutrophils, lysosome-like azurophil granules do not fuse with phagosomes containing either pathogenic or nonpathogenic mycobacteria (34). The identity of other receptors which contribute to the silent phagocytosis of nonopsonized mycobacteria remains to be established. It seems reasonable to look for these receptors among those already implicated in the nonopsonic binding of mycobacteria (12).

In conclusion, MR-dependent phagocytosis is coupled to neither the activation of NADPH oxidase nor the maturation of phagosomes until fusion with the Hck compartment. We propose that the MR which is unable to distinguish between pathogenic and nonpathogenic mycobacteria constitutes a safe portal of entry for microorganisms.