A Macrophage Invasion Mechanism of Pathogenic Mycobacteria

See allHide authors and affiliations

Science  22 Aug 1997:
Vol. 277, Issue 5329, pp. 1091-1093
DOI: 10.1126/science.277.5329.1091


Tuberculosis is the leading cause of death due to an infectious organism, killing an estimated 3 million people annually.Mycobacterium tuberculosis, the causative agent of tuberculosis, and other pathogenic mycobacteria require entry into host macrophages to initiate infection. An invasion mechanism was defined that was shared among pathogenic mycobacteria including M. tuberculosis, M. leprae, and M. avium but not by nonpathogenic mycobacteria or nonmycobacterial intramacrophage pathogens. This pathway required the association of the complement cleavage product C2a with mycobacteria resulting in the formation of a C3 convertase. The mycobacteria-associated C2a cleaved C3, resulting in C3b opsonization of the mycobacteria and recognition by macrophages.

It is estimated that a third of the world's population is infected withMycobacterium tuberculosis, the causative agent of tuberculosis, and an additional 8 million new cases of pulmonary tuberculosis occur each year (1). Much of the resurgence of tuberculosis can be attributed to the human immunodeficiency virus epidemic (2), which also predisposes infected hosts to other mycobacterial infections, especially M. avium(3). These mycobacteria are intracellular pathogens that reside almost exclusively within macrophages of infected individuals. In vitro studies have demonstrated a number of mechanisms by which mycobacteria can invade host macrophages including opsonization of mycobacteria with C3 through activation of the alternative pathway of complement, followed by invasion of macrophages by way of the complement receptors CR1, CR3, and CR4 (4). Consistent with this mechanism, heat treatment of fresh normal human serum, which inactivates the complement pathways, abolishes its ability to support mycobacteria invasion of macrophages (4, 5). However, heat-treated commercial equine serum or human serum from a mixed lymphocyte reaction retains the ability to markedly enhanceM. avium uptake by macrophages (6). This indicates that the latter serum fractions contain an additional factor, not present in freshly isolated normal human serum, that can facilitate M. avium invasion of macrophages.

We have purified the active component from heat-treated equine serum (7) and tested its ability to mediate macrophage invasion for different mycobacterial species. The purified component markedly enhanced macrophage invasion by pathogenic mycobacteria [M. leprae, bacillus Calmette-Guerin (BCG), and M. tuberculosis] but not by fast-growing nonpathogenic mycobacteria (M. vaccae, M. smegmatis, and M. phlei) (Fig.1A). Other pathogens such asLeishmania mexicana, Listeria monocytogenes, Staphlococcus sp. (Fig. 1B), andNocardia asteroides (8) could not use this serum factor to mediate macrophage entry. Thus, this method of invasion appears unique to pathogenic mycobacteria and suggests a virulence mechanism for mycobacterial infection.

Figure 1

Use of a heat-stable serum component for macrophage infection by pathogenic mycobacteria. (A) The various mycobacterial species [5 × 105colony-forming units (CFU)] were incubated with human macrophages (19) in the presence of infection media (RPMI, 0.5% human serum albumin) or infection media with Con A–fractionated Δ56°C equine serum, or 10% fresh normal human serum, and the percentage of infected macrophages was determined (20). (B) The various organisms (5 × 105 CFU) were incubated with human macrophages for 2 hours in the presence of infection media with or without fractionated Δ56°C equine serum, and the percentage of infected macrophages was determined. The ratio of macrophages infected in the presence of serum proteins compared to infection with media alone is shown. Percentage of infected macrophages in media alone: M. avium, 3; L. mono., 41;L. mex., 35; S. cerevisiae., 40; S.aureus, 10. Abbreviations: MTB, M. tuberculosis; M. smeg., M. smegmatis;L. mono, Listeria monocytogenes; L. mex., Leishmania mexicana; S. cerevisiae,Saccharomyces cerevisiae; S. aureus,Staphylococus aureus, coagulase negative.

The purified active component from equine serum is a 70-kD protein (7). The NH2-terminal sequence of this protein was identical to amino acids 244 to 263 of human complement component C2 in 19 of 20 amino acids (9). This sequence corresponds to the NH2-terminus of the C2a fragment of C2, which is generated during classical complement pathway activation (10). This 70-kD protein was of the correct molecular size for C2a and was recognized by a polyclonal antibody to human C2 (8). These active fractions had no other detectable proteins on silver staining and no intact C2 by protein immunoblot. Our findings were unexpected because C2a has been shown to function only as part of a complement C3-cleaving enzyme (convertase) and in the absence of C4b has no known activity (11).

To confirm the role of C2a in mycobacteria invasion of macrophages, we immunodepleted C2 and its activation fragments C2a and C2b from fractionated heat-treated equine serum and human serum (12). Depletion of C2 and its fragments abolished facilitation of M. avium uptake by macrophages (Fig.2A), whereas treatment of the serum samples with antibodies to human fibronectin had no effect (8). To demonstrate that C2a was sufficient to stimulate the invasion process, we incubated purified complement components C1 and C4 with antibody-coated erythrocytes (EA) to form a C2 convertase. Upon prolonged incubation with commercially obtained pure human C2, the C4b2a complex forms and then dissociates, leaving C2a free in solution. This C2a enhanced M. avium ingestion by macrophages, whereas intact C2 did not enhance M. avium invasion (Fig. 2B). These data demonstrate that carry-over of trace amounts of intact C2 to form a classical pathway C3 convertase is not the mechanism for this mycobacterial invasion of macrophages. Therefore, pathogenic mycobacteria have evolved a mechanism to salvage C2a and use it as a means of invading macrophages. C2a does not reassociate with C4b and thus in the absence of these pathogenic mycobacteria, has no ability to cleave C3 (11).

Figure 2

C2a is the active component from heat-treated serum that enhances M. avium entry into macrophages. (A) Partially purified C2a from heat-treated human and equine serum was tested for its ability to support M. aviuminvasion of human monocyte-derived macrophages after incubation with anti-C2 sepharose or control sepharose (12). (B) Human macrophages and M. avium were incubated with supernatant from immunoglobulin M–coated sheep erythrocytes (EA) treated sequentially with complement proteins C1, C4, and C2 or with C2 alone (21). The conditions that generate C2a (that is, EAC1, 4, 2) produce invasion-promoting activity.

Prior treatment of C2a with the irreversible serine protease inhibitor diisopropyl-fluorophosphate (DFP) markedly decreased C2a enhancement ofM. avium entry into macrophages (Fig.3A), suggesting a role for C3, the only known C2a substrate, in C2a-dependent macrophage invasion. Addition of DFP to M. avium preopsonized with C3 by the alternative pathway did not block its invasion of macrophages, confirming that DFP acted on C2a and not on the macrophages (8). Prior incubation of M. avium with both C2a and C3, but not with either component individually, markedly increased its ingestion by human macrophages (Fig. 3B). Furthermore, the combination of C2a- and C3-treated mycobacteria stained positively with an antibody to human C3 (anti-C3) (8), indicating that incubation of mycobacteria with C2a resulted in deposition of C3 on the mycobacteria. C3 cleavage was measured directly by enzyme-linked immunosorbent assay (ELISA) with antibodies specific for C3a, a cleavage product of C3. Incubation of pathogenic mycobacteria with C2a and C3, but not C2 and C3, resulted in C3 cleavage (Fig. 3C). Thus, pathogenic mycobacteria can use C2a but not intact hemolytically active C2 to form a C3 convertase.

Figure 3

Mycobacterium avium and C2a function as a C3 convertase. (A) Partially purified C2a from heat-treated human and equine serum was incubated for 1 hour at 25°C with or without 5 mM DFP, a serine protease irreversible inhibitor. The C2a samples were dialyzed at 4°C against PBS to remove free DFP and then incubated with M. avium and human monocyte-derived macrophages to test the C2a fractions for invasion-enhancing activity. (B) After preincubation of M. avium with either partially purified C2a or 50 U of C3 or both, the organisms were washed and resuspended in infection media and tested for their ability to invade human macrophages. (C) Mycobacterium avium BCG, and M. smegmatis (2 × 107), were incubated with intact C2, C2a, and C3 as indicated for 1.5 hours at 37°C. Cleavage of human C3) was assessed by enzyme-linked immunosorbent assay for C3a (Quidel, San Diego, California). (D) Mycobacterium avium invasion of mouse bone marrow macrophages (22) was tested in the presence of infection media only or infection media containing 10% immune complex–treated serum (Δ56°C) from C3−/− mice (23). (E) Same as (C) except that 10% heat-treated equine serum was used. (F)Mycobacterium avium invasion of human macrophages in the presence of partially purified C2a was assessed after preincubation of the macrophages with monoclonal antibodies (30 μg/ml): 1B4 (24) and OKM1 (25) blocking antibodies to CR3; 543 (26) nonblocking and 3D9 (27) blocking antibody to CR1; B6H12 irrelevant binding control antibody against integrin-associated protein (28). Media: M. avium added to macrophages in the absence of C2a.

The initial experiments showed that C2a could enhance mycobacteria invasion of macrophages without an exogenous source of C3, suggesting that macrophages could be a source of C3. To address this possibility, we incubated macrophages from C3−/− mice (13) or control mice with heat-treated serum from the C3−/−mice as a source of murine C2a. Under these conditions, M. avium was not ingested by the C3−/− macrophages but was able to invade wild-type macrophages (Fig. 3D). Addition of heat-treated equine serum as a source of C2a and C3 restored M. avium invasion of C3−/− macrophages, demonstrating that these cells were not abnormally resistant to infection (Fig. 3E). Additional in vitro experiments indicated that the predominant M. avium opsonin is C3b, the ligand for CR1, and not C3bi, the ligand for CR3 (Fig. 3F).

Macrophage invasion by mycobacteria can proceed by opsonization with C3 by the alternative complement pathway or by interactions of mannose-containing cell wall glycolipids with macrophage lectins (14). However, neither of these mechanisms is specific to pathogenic mycobacteria, which have evolved as extremely successful intracellular pathogens. Here we describe a mycobacterial macrophage invasion mechanism that is not shared by other organisms including other nonmycobacterial intracellular pathogens and other mycobacteria that are not intracellular pathogens. This C2a-dependent entry pathway thus has the characteristics of a virulence mechanism for pathogenic intracellular mycobacteria. The invasion mechanism is very sensitive to C2a since 1 to 10 nM C2a is sufficient to mediate a maximum rate of invasion in vitro (8). This requires cleavage of <1% of the serum concentration of C2, suggesting that local concentrations of C2a in a biologically significant range could be obtained at the site of a mycobacterial infection. Subsequent steps after this mycobacteria-dependent mechanism for C2a cleavage of C3 are identical to what has already been shown for mycobacteria opsonized with C3b by the alternative complement pathway, including invasion of the macrophage by way of complement receptors and survival and replication within macrophage phagosomes (14).

This C2a-mediated invasion mechanism may be involved in both the establishment of an infection and in its propagation. Previous work has shown that lung exposure to M. tuberculosis results in an inflammatory response that leads to complement activation and C2a production (15). Additional studies have shown that patients with active pulmonary tuberculosis have high levels of circulating immune complexes (16) with activation of the classical complement pathway resulting in increased serum levels of C2a (17). The presence of C2a during several stages of a mycobacterial infection, along with conservation of this C2a-dependent uptake pathway among pathogenic mycobacteria, suggests that this invasion mechanism has a crucial role in mycobacterial pathogenesis. The mycobacterial cell wall component required for this invasion process could provide a new target for therapeutic intervention.


View Abstract

Navigate This Article