TEXT

Top Abstract Text References

Leukocyte adhesion molecules are used by respiratory pathogens to bind and invade macrophages via different mechanisms. Recognition of Mycobacterium tuberculosis by macrophages involves opsonin-dependent interactions as well as direct binding of bacterial structure to host cell receptors (14). Opsonic uptake of M. tuberculosis is mediated by cell surface receptor CR1 (CD35) and the 2 integrin molecules CR3 (CD11b or CD18, Mac1, or _M2) and CR4 (CD11c or CD18, gp150 or -95, or _X2) (4, 7). However, M. tuberculosis uptake by human monocytes is less dependent on serum than that of other mycobacteria, such as Mycobacterium avium (10). This serum-independent mechanism is especially important in the early stages of infection (11) or in an environment poor in opsonins like the lungs. Nonopsonic uptake of M. tuberculosis is mediated by the mannose receptor, CD14, and CR3 (2, 5, 8). CR4 has been shown to participate in serum-independent recognition of Histoplasma capsulatum by human neutrophils (9) and Escherichia coli by human macrophages (13). The role of CR4 in nonopsonic uptake of M. tuberculosis by macrophages has not been assessed. 2 integrins also transmit signal to macrophages leading to phagocytosis (15). Study of this signal transduction pathway may be crucial in elucidating the ability of M. tuberculosis to evade intracellular killing responsible for development of infection.

In the present work, we show that CR4 is the predominant integrin receptor on the macrophage and plays an important role in M. tuberculosis-macrophage binding in the absence of serum, as well as in its subsequent signal transduction.

We first studied nonopsonic binding of M. tuberculosis to monocyte-derived macrophages (MDMs). Mononuclear cells were isolated from peripheral blood of healthy donors with Ficoll-Paque (Pharmacia, Uppsala, Sweden) by adhesion for 60 min at 37°C with removal of nonadherent cells (15). Monocytes were allowed to differentiate into macrophages by further culture on plastic for 5 days in Iscove's modified Dulbecco's medium (BioWhittaker, Walkersville, Md.) supplemented with 2 mM L-glutamine, 100 U of penicillin per ml, 100 µg of streptomycin per ml, and 10% autologous serum. Twelve hours before being harvested, cells were incubated in serum-free medium.

Monocytes and MDMs were infected at a ratio of 100 bacteria per cell in RPMI 1640 (BioWhittaker) for different lengths of time at 37°C. The H37Ra strain of M. tuberculosis used in this study was purchased from the American Type Culture Collection (ATCC) (Rockville, Md.) and grown as previously described (4). After infection, cells were fixed and mycobacteria were stained by Kinyoun's modified acid-fast bacillus staining method. The association index was calculated by multiplying the average number of bacteria per cell with the percentage of infected cells.

The mean association index after 60 min of monocyte infection was 5.4 ± 2.3 (56% of cells infected with at least 1 bacterium, 9.7 bacteria per infected cell). The mean association index for MDMs increased dramatically to 28.3 ± 11 (81% of infected cells, 34.4 bacteria per cell) (Fig. 1A). M. tuberculosis binding to MDMs was significantly greater than that of monocytes at all time points (5 min, P = 0.04, Student's t test; 15 min, P < 0.001; 60 min, P = 0.02; and 120 min, P = 0.01).

View larger version (18K): [in this window] [in a new window]   FIG. 1.   M. tuberculosis binding and receptor expression. (A) Monocytes or MDMs from the same donors were infected with M. tuberculosis at a ratio of 100 bacteria per cell. Bacteria were detected by Kinyoun's method. Results are expressed as the association index ± standard deviation averaged from three experiments. (B) Monocytes or MDMs were incubated with MAb and labeled with fluorescein isothiocyanate-conjugated antimouse IgG. The fluorescence values determined by flow cytometry are the mean ± standard deviation from three experiments.

The structural basis for this enhanced uptake of M. tuberculosis by MDMs was studied. We measured surface CR1, CR3, CR4, and CD14. Monoclonal antibodies (MAbs) against CD11b (clone M1/70 [ATCC]), CD11c (clone SHCL-3 [Becton Dickinson, San Jose, Calif.]), CR1 (clone 3D9 [kindly provided by J. O'Shea]), and CD14 (Leu-M3 [Becton Dickinson]), followed by fluorescein isothiocyanate-conjugated antimouse immunoglobulin G (IgG), were used to study receptor expression. Fluorescence was analyzed by flow cytometry of 10,000 cells with a Becton Dickinson fluorescence-activated cell sorting analyzer. There was no significant difference between surface expression of CR1, CR3, and CD14 on monocytes and that on MDMs. In contrast, a dramatic increase in CR4 expression on MDMs was observed compared to that on monocytes (Fig. 1B).

Since MDMs efficiently bind M. tuberculosis and express high levels of CR4, we tested the hypothesis that CR4 directly interacts with M. tuberculosis by two different methods. First, we used a blocking antibody to CR4 to examine its effect on M. tuberculosis-MDM binding. Preincubation of MDMs for 30 min at 4°C with an anti-CD11c MAb (10 µg/ml; clone 3.9 [Serotec, Raleigh, N.C.]) inhibited M. tuberculosis-MDM binding by 36% (P < 0.025), while isotypic control IgG had no effect. Second, we studied the ability of CR4 to bind M. tuberculosis by using transfected cells. CHO cells expressing CR4 (clone RI 3.7) and mock-transfected CHO cells (CHO/Neo) were kindly provided by D. T. Golenbock (Boston, Mass.) and grown in Ham's F-12 medium supplemented with 10% fetal bovine serum and 2 mM L-glutamine. After 60 min of incubation at 37°C in the absence of serum, M. tuberculosis was found to bind specifically the CR4-transfected cells (Fig. 2). The association index was 3.5 ± 0.4 (31% of infected cells, 12 bacteria per infected cell) for M. tuberculosis binding to RI 3.7 versus 0.5 ± 0.2 to CHO/Neo (P < 0.001). M. tuberculosis binding to CR4-transfected cells was lower than that observed with MDMs. This was probably explained by lower expression of CR4 (about 15 times) on transfected cells compared to that of MDMs (data not shown) and led us to use a higher bacterium/cell ratio (500/1) for studies with the CR4-transfected cells. Cooperativity with other receptors present exclusively on MDMs (CR3, mannose receptor, and CD14) can also explain the difference between MDMs and CHO/CR4 cells in M. tuberculosis binding. Moreover, macrophages are known to produce complement proteins, and use of transfected cells allowed us to confirm nonopsonic binding of M. tuberculosis to CR4. Preincubation of RI 3.7 cells with anti-CR4 MAb (clone 3.9) inhibited M. tuberculosis binding in a dose-dependent manner (Fig. 2). Control antibody recognizing major histocompatibility complex class I molecules (clone W6.32 [ATCC]) failed to inhibit M. tuberculosis binding to CR4-transfected cells.

View larger version (33K): [in this window] [in a new window]   FIG. 2.   M. tuberculosis binding to transfected cells. CR4-transfected cells were preincubated with anti-CR4 antibody (clone 3.9) or control antibody and then infected with 500 bacteria/cell. The association index averaged from three experiments is shown. Preincubation with 30 or 10 µg of anti-CR4 antibody per ml significantly inhibited M. tuberculosis binding to the transfected cells (*, P = 0.002; **, P < 0.001). Binding of M. tuberculosis to control cells transfected with the vector only (CHO/Neo) was performed in the absence of antibody (Ab).

From these data, we concluded that CR4 directly binds M. tuberculosis and confers the enhanced M. tuberculosis binding activities to MDMs.

The next series of experiments was aimed to determine if binding of M. tuberculosis to CR4 triggered signal transduction. It is known that CR4 binds sCD23, the soluble low-affinity receptor for IgE, on human monocytes, resulting in NO production (1). Tyrosine phosphorylation is a major early event induced by the engagement of integrins in phagocytes (6). Thus, involvement of CR4 in uptake of M. tuberculosis might be accompanied by such an activation event. Incubation of MDMs with the bacteria was carried out in RPMI 1640 and stopped in cold Hanks' balanced salt solution (BioWhittaker) containing protease inhibitors and sodium orthovanadate, a tyrosine phosphatase inhibitor (15). Cell lysates were electrophoresed by sodium dodecyl sulfate-10% polyacrylamide gel electrophoresis. After transfer onto a nitrocellulose sheet, blots were probed with an antiphosphotyrosine antibody (UBI, Lake Placid, N.Y.). The reaction was completed with an antibody coupled to peroxidase and a chemiluminescent substrate (ECL [enhanced chemiluminescence] substrate; Amersham, Arlington Heights, Ill.).

As shown in Fig. 3A, M. tuberculosis induced tyrosine phosphorylation of several MDM proteins. One major phosphorylated protein migrating at 60 kDa was detected after 5 min of exposure to the bacteria. Phosphorylation increased further between 5 and 30 min and then decreased at 60 min. Tyrosine phosphorylation was observed at 5 min after infection for 101-, 75-, 61-, and 45-kDa proteins; 15 min for the 97-, 83-, 41- and 37-kDa proteins; and 30 min for the 34- and 29-kDa proteins. At 60 min, the phosphorylation pattern was less intense than that in unstimulated cells. We speculate that this pattern is due to activation of tyrosine phosphatases. Antibody 3.9, directed against CR4 (used at 20 µg/ml), inhibited the tyrosine phosphorylation of the 60-kDa protein (Fig. 3B), suggesting that this protein may play an important role in MDM response to M. tuberculosis binding. This inhibition was specific, since irrelevant antibodies (clone W6.32) had no effect on phosphorylation of the 60-kDa protein. Figure 3C shows densitometer measurements averaged from three inhibition experiments. Anti-CD11c antibody significantly inhibited the 60-kDa protein tyrosine phosphorylation compared to the control antibody (P = 0.01; Student's t test). Incubation of the MDMs with the antibodies alone (3.9 and W6.32) did not induce tyrosine phosphorylation (data not shown).

View larger version (23K): [in this window] [in a new window]   FIG. 3.   Tyrosine phosphorylations induced by M. tuberculosis infection. (A) MDMs were infected with 100 bacteria per cell for different periods of time. Blots with an antiphosphotyrosine MAb and antimouse antibody coupled to horseradish peroxidase are shown. (B) After preincubation with antibody for 30 min at 4°C, MDMs were infected for 30 min at 37°C. Blots are representative of three experiments. (C) Densitometer measurements of the 60-kDa band intensity were averaged from three experiments. Ab, antibody.

In summary, we demonstrate a role for CR4 in nonopsonic M. tuberculosis binding to human macrophages. This interaction led to rapid tyrosine phosphorylation of a major 60-kDa protein in host cells, and this phenomenon was sensitive to blocking by an antibody to CR4. This 60-kDa molecule, possibly a member of the Src family, such as p60^src, may play an important role in M. tuberculosis phagocytosis, since tyrosine phosphorylations are involved in the ingestion process of inert particles (3). The association between p60^src and cytoskeletal proteins such as paxillin (12) could lead to cytoskeleton remodeling during the phagocytic process. The early phosphorylation observed for the 60-kDa protein after M. tuberculosis infection of MDMs is evidence in favor of its role in the phagocytic process.

Given that high levels of CR4 are present on macrophage cell surfaces, we propose that binding of a constituent of M. tuberculosis cell wall to CR4 constitutes a major route for the entry of bacteria into macrophages, especially at early stages of infection when opsonins are not available. This interaction leads to macrophage activation and to phosphorylation of proteins that may be critical to the macrophage response to M. tuberculosis.