Inhibition of Ca²⁺ Signaling by Mycobacterium tuberculosisIs Associated with Reduced Phagosome–Lysosome Fusion and Increased Survival within Human Macrophages

Tuberculosis is a global health problem with enormous impact on human morbidity and mortality 1. Approximately one-third of the world's population is infected with Mycobacterium tuberculosis, and three million people die of active disease each year. An essential virulence characteristic of M. tuberculosis is its ability to successfully parasitize monocytes and macrophages, despite the presence of multiple microbicidal mechanisms within these cells 2. The molecular mechanisms responsible for the intracellular survival of M. tuberculosis are unknown.

Multiple host–pathogen interactions may impact the fate of M. tuberculosis within human monocytes and macrophages and, consequently, the presence or absence of disease in infected individuals. The earliest interaction between M. tuberculosis and mononuclear phagocytes is the binding and uptake of the bacilli by plasma membrane phagocytic receptors 3. Phagocytosis of M. tuberculosis, in either the presence or absence of serum, is predominantly mediated by the complement receptor (CR)1, CR3, and CR4 4,5,6. In human monocytes and monocyte-derived macrophages (MDMs), the β₂-integrin CR3 is the major phagocytic receptor for M. tuberculosis, and anti-CR3 Abs inhibit ingestion of tubercle bacilli by ∼80% 5. In serum-free conditions, the macrophage mannose receptor also mediates mycobacterial phagocytosis, although its contribution to ingestion of M. tuberculosis is much less in the presence of complement proteins 7.

The ability of M. tuberculosis to enter macrophages via the CR-mediated phagocytic pathway may contribute to its intracellular survival, as, in many cases, CR ligation does not trigger phagocyte microbicidal responses 8,9. Studies with murine macrophages demonstrate that the class of phagocytic receptor that mediates ingestion of M. tuberculosis has a strong influence on the extent of phagosomal maturation. CR-mediated phagocytosis of M. tuberculosis results in a phagosome that is unable to fuse with lysosomes 10. Conversely, if the bacillus is opsonized with M. tuberculosis–specific Abs, its ingestion is mediated by macrophage FcγRs, and the mycobacterial phagosome undergoes full maturation to a phagolysosome 11. These results suggest that FcγR-mediated ingestion of M. tuberculosis must mobilize signaling pathways that are distinct from those that are activated by CRs, which are responsible for the difference in phagosome maturation. The relevance of these observations to human disease has been questioned, because the antimycobacterial activity of murine macrophages is much more easily demonstrated in vitro than that of human macrophages. Although multiple investigators have demonstrated that CR-dependent ingestion of M. tuberculosis by human macrophages is also followed by defective phagosomal maturation 12, to our knowledge, no data is available on the effects of Ab opsonization on survival of M. tuberculosis within human macrophages. Furthermore, the biochemical mechanisms responsible for incomplete maturation of M. tuberculosis–containing phagosomes are unknown.

Many distinct signal transduction pathways contribute to the activation of phagocyte antimicrobial defenses, but their integrative function and relative priority in the killing of specific pathogens is unknown. Stimulation-induced increases in cytosolic Ca²⁺ concentration ([Ca²⁺]_c) are essential for activation of the phagocyte respiratory burst, production of nitric oxide, secretion of microbicidal granule constituents, and synthesis of proinflammatory mediators, including TNF-α 13,14,15,16,17. Based on these considerations, three questions of specific relevance to the pathogenesis of tuberculosis were investigated in this study: (a) Does virulent M. tuberculosis alter Ca²⁺-mediated signal transduction in human macrophages? If so, (b) Do these alterations in macrophage Ca²⁺ signaling contribute to incomplete phagosomal maturation and intracellular survival of M. tuberculosis, and (c) Does the route of entry into human macrophages, i.e., via CR- versus FcγR-mediated phagocytosis, affect the intracellular viability of M. tuberculosis?

Hepes, zymosan, and collagen were obtained from Sigma Chemical Co. RPMI 1640 medium with l-glutamine and PBS was purchased from GIBCO BRL. Middlebrook 7H9 broth was obtained from BBL Microbiology Systems, and 7H11 agar, oleic acid-albumin-dextrose-catalase enrichment medium, and auramine-rhodamine stain were from Difco Labs., Inc. Teflon wells were obtained from Savillex Corp. Tissue culture plates were purchased from Linbro Flow Labs., and Fura2 and A23187 were from Molecular Probes, Inc. Bis-(2-amino-S-methylphenoxy) ethane-N,N,N′,N′,-tetraacetic acid tetraacetoxymethyl ester (MAPTAM) was obtained from Calbiochem Corp., and dipalmitoylphosphatidylcholine (DPPC) was from Avanti, Inc.

Polyclonal (A-188) and monoclonal Abs (CS-40, CS-35) to lipoarabinomannan (LAM) from M. tuberculosis were provided by Drs. Patrick Brennan and John Belisle (Colorado State University, Fort Collins, CO; National Institutes of Health grant AI-75320). A-188 and CS40 are specific to LAM from the virulent Erdman strain of M. tuberculosis, whereas CS35 recognizes an epitope common to LAMs from several strains of M. tuberculosis. mAbs to CD18 (H52) and lysosome-associated membrane protein (LAMP)1 were obtained from the Developmental Studies Hybridoma Bank (University of Iowa, Iowa City, IA). F(ab′)₂ fragments of α-CD18 were prepared by digestion with pepsin as previously described 20 and partially purified by protein G–Sepharose chromatography. Goat anti–human C3 IgG was obtained from Atlantic Antibodies, Inc.

PBMCs were isolated from healthy, purified protein derivative (PPD)-negative, adult volunteers and cultured in Teflon wells for 5 d in RPMI 1640 with 20% fresh autologous serum as previously described 21. Macrophages were purified by adherence to chromic acid–cleaned, collagen-coated glass coverslips for 2 h at 37°C in 5% CO₂. Monolayers were washed repeatedly and incubated in RPMI, 20 mM Hepes (RH), pH 7.4, 2.5% serum for use in experiments. Effects of experimental manipulations on macrophage viability were assessed by exclusion of trypan blue, and monolayer density was determined by nuclei counting with naphthol blue-black stain 22.

The Erdman, H37Rv, and H37Ra strains of M. tuberculosis were obtained from the American Tissue Type Culture Collection and were cultured and prepared for use in experiments as noted previously 5,7,21. In brief, aliquots of frozen M. tuberculosis stocks in 7H9 broth were thawed, cultured for 9 d on 7H11 agar at 37°C in 5% CO₂/95% air, scraped from agar plates, and suspended in RH by vortexing briefly in an Eppendorf tube containing two glass beads. After settling, the supernatant was transferred to a new tube and allowed to settle once again. Heat killing was accomplished by incubating this final suspension at 100°C for 10 min and confirmed by absence of CFUs 23,24. Gamma-irradiated (killed) M. tuberculosis was provided by Drs. Patrick Brennan and John Belisle (Colorado State University). For experiments requiring complement-opsonized (C-op) bacilli, aliquots of M. tuberculosis (live, heat-killed, or gamma-irradiated) were preopsonized in 50% human serum for 30 min at 37°C and then washed three times in PBS. Ab opsonization of M. tuberculosis was achieved by incubating the bacilli with 10 μg/ml CS-40 or CS-35 or 10 μl of A-188 for 30 min, followed by washing in PBS. After opsonization, M. tuberculosis preparations were resuspended in HBSS using glass beads, and clumped organisms were allowed to settle, as described above. M. tuberculosis suspensions were counted in a Petroff-Hauser chamber, and the concentration of bacteria was adjusted for use in experiments. Final M. tuberculosis preparations contained >95% single bacteria, with ≥75% viability by determination of CFUs 5,21. The effects of various experimental manipulations on the viability of M. tuberculosis were also determined by analysis of CFUs.

Phagocytosis of M. tuberculosis was determined as previously described 5,21. In brief, macrophage monolayers adherent to glass coverslips (∼2 × 10⁵ MDMs per coverslip) in 24-well tissue culture plates were incubated with M. tuberculosis (multiplicity of infection [MOI] of 10:1) in RH, 2.5% autologous nonimmune serum. After incubation for various intervals, monolayers were washed repeatedly to remove nonadherent bacteria, fixed in 10% formalin, and stained with auramine-rhodamine for 20 min 5,21. Coverslips were washed with distilled water and incubated with acid alcohol for 3 min, washed, and incubated in KMnO₄ for 2 min. Adherent bacteria were quantitated by fluorescence microscopy of triplicate coverslips for each experimental condition (50–200 MDMs per coverslip), and results of a set of experiments were expressed as the mean (± SEM) number of adherent M. tuberculosis per 100 macrophages (phagocytic index). Previous electron microscopic studies of this assay have indicated that all adherent mycobacteria are phagocytosed, both under control conditions and in experiments in which phagocytosis is inhibited or augmented 5,21.

Heat-killed or live M. tuberculosis was incubated in 50% human serum for 30 min at 37°C. The bacteria were recovered by centrifugation at 12,000 g for 10 min, washed twice, and solubilized in SDS sample buffer (62.5 mM Tris/HCl, pH 6.8, 2% SDS, 10% glycerol, 75 mM β-ME, 0.0025% bromphenol blue). After SDS-PAGE on 10% gels, proteins were transferred to polyvinylidene difluoride membranes, Western blotted with goat anti–human C3 IgG, and detected by enhanced chemiluminescence as described 5,21.

Calcium measurements were performed at the Cell Fluorescence Core Facility (Veterans Affairs Medical Center, Iowa City, IA). MDMs were adhered to collagen-coated glass coverslips and incubated with 10 μM Fura2-AM in HBSS for 30 min at 37°C. [Ca²⁺]_c in single MDMs, or the mean [Ca²⁺]_c of groups of 10–20 cells, was determined using a Photoscan II spectrofluorometer (Photon Technology Intl.) with a Nikon microscope (Nikon, Inc.). [Ca²⁺]_c was determined from the ratio of fluorescence emission intensities at 510 nm after excitation at 340 and 380 nm, respectively. Background fluorescence intensities at each excitation wavelength were subtracted from each data point. The ratios of the corrected fluorescence intensities (R) were then converted to the actual calcium concentration using the formula, [Ca²⁺] = K_d(R − R_min)/(R_max − R) (reference 25), where the maximum and minimum ratios, as well as the dissociation constant, were empirically derived from [Ca²⁺] curves generated with the instrument. In certain experiments, the effects of the absence of extracellular Ca²⁺ were determined by incubation of MDMs in Ca²⁺-free HBSS with 3 mM EGTA. To chelate cytosolic Ca²⁺, MDMs were preincubated with MAPTAM (15–25 μM) for 30 min at 37°C 26,27.

MDMs, adherent to collagen-coated glass coverslips, were infected at an MOI of 1:1 with Erdman M. tuberculosis (preopsonized with complement or anti-LAM Abs) in RH, 2.5% heat-inactivated autologous serum. After 1 h, the monolayers were washed and repleted with buffer containing 1% heat-inactivated serum. 24, 48, and 96 h after infection, supernatants were transferred to sterile microfuge tubes, monolayers were lysed with ice cold sterile water, and SDS was added to a final concentration of 0.25%. Lysates were combined with their corresponding supernatants and resuspended in 7H9, and serial dilutions were plated in duplicate on 7H11 agar. Colonies were counted 2 wk after plating. To determine the effect of elevation of MDM intracellular [Ca²⁺] on mycobacterial survival, monolayers were infected at a 1:1 ratio with C-op M. tuberculosis in HBSS containing the Ca²⁺ ionophore A23187 (1 μM) or an equivalent volume of ethanol solvent (0.1%). After 20 min, monolayers were washed and repleted with 20 μg/ml phosphatidylcholine vesicles, 1% autologous serum in RH, to reverse the A23187-mediated influx of extracellular Ca²⁺ 28. DPPC vesicles were prepared by evaporation of a chloroform/methanol (2:1) solution under N₂ and resuspension in HBSS by sonication for 10 min at 25°C 21. CFUs were counted as described above.

The acidophilic dye LysoTracker Red (Molecular Probes, Inc.) was incubated at a 1:10,000 dilution with MDM monolayers in RH, 2.5% autologous serum, for 2 h at 37°C. Unincorporated dye was removed by washing, followed by infection with M. tuberculosis for 1 h. After removal of nonadherent bacilli, LysoTracker Red was added to each well at the same concentration used for initial labeling. 24 h after infection, MDMs were fixed in 3.75% paraformaldehyde for 15 min and permeabilized with ice cold methanol/acetone (1:1). The localization of M. tuberculosis was ascertained by incubating monolayers with auramine for 20 min at 25°C, followed by a 3-min incubation in acid alcohol. After thorough washing, monolayers were blocked with a PBS, 5% BSA, 10% goat serum for 1 h. In parallel experiments using Abs to the lysosomal protein markers LAMP-1, cathepsin D, and CD63, coverslips were incubated with the appropriate primary Abs (diluted in blocking solution) for 1 h at 25°C, washed, and then incubated with the corresponding fluorophore-conjugated secondary anti-IgG Ab for 1 h. After repeated washings, coverslips were mounted with buffered glycerol solution and sealed with nail polish. Confocal microscopy was performed on a Zeiss Laser Scan Inverted 510 microscope (Carl Zeiss, Inc.). An argon/krypton laser (excitation, 488 nm; emission, 505–530 nm) was used for detection of auramine fluorescence, and a helium/neon laser (excitation, 543 nm; emission, >585 nm) for detection of Texas Red and LysoTracker Red. The percentage of M. tuberculosis phagosomes colocalizing with the marker of interest was determined by counting >25 phagosomes from at least 10 different fields per condition.

Data from each experimental group were subjected to an analysis of normality and variance. Differences between experimental groups composed of normally distributed data were analyzed for statistical significance using Student's t test. Nonparametric evaluation of other data sets was performed with the Wilcoxon Rank Sum test 29.

As binding and phagocytosis of M. tuberculosis by macrophages in the presence or absence of serum is primarily mediated by CRs 4,5,30, we first characterized macrophage Ca²⁺ signaling induced by the model particulate CR–ligand, C-op zymosan (COZ) 31,32. Previous studies in neutrophils have demonstrated that COZ induces a significant increase in [Ca²⁺]_c due to stimulation of CR3 and, to a lesser extent, CR1 33,34,35,36,37. In addition to CR3 and CR1, macrophages, unlike neutrophils, also express high levels of CR4 38. Therefore, it was necessary to characterize in detail the effects of COZ on [Ca²⁺]_c in human macrophages to serve as a control for subsequent experiments with M. tuberculosis.

MDMs were purified from PBMCs of healthy, PPD-negative adult donors after 5-d culture in RPMI, 20% autologous serum, by adherence to collagen-coated glass coverslips 21. After loading of MDMs with the Ca²⁺-sensitive dye Fura2 (10 μM), monolayers were washed and placed in Ca²⁺, Mg⁺²-containing HBSS (CHBSS). Levels of [Ca²⁺]_c in single MDMs were determined by fluorescence ratio imaging of Fura2 25. The basal level of [Ca²⁺]_c in resting MDMs ranged from ∼50 to 150 nM (Fig. 1 A). Incubation with COZ at a particle/cell ratio of 10:1 resulted in a rapid increase in macrophage [Ca²⁺]_c, which peaked in the 300–800 nM range and gradually returned to basal levels over the next 8–10 min (Fig. 1 A; fold-increase in [Ca²⁺]_c = 4.6; range, 2.4–6.5-fold; n = 20). Subsequent addition of thapsigargin (1 μM), which inhibits the Ca²⁺/ATPase responsible for reaccumulation of [Ca²⁺]_c into endoplasmic reticulum stores 39, resulted in a further increase in macrophage [Ca²⁺]_c. This thapsigargin-induced increase in [Ca²⁺]_c provided verification of the intact capacity of the intracellular Ca²⁺ storage pool and the functional integrity of the capacitative Ca²⁺ entry mechanism 40,41.

The COZ-induced increase in macrophage [Ca²⁺]_c was due to both release of Ca²⁺ from intracellular stores and influx of extracellular Ca²⁺, as the average magnitude and duration of the elevated [Ca²⁺]_c was significantly attenuated, but not abolished, by incubation of MDMs in Ca²⁺, Mg⁺²-free HBSS containing 3 mM EGTA (data not shown). Under these conditions, the residual elevation in [Ca²⁺]_c is due to release from intracellular Ca²⁺ stores in the endoplasmic reticulum, as evidenced by the increase in Fura2 fluorescence upon addition of thapsigargin. Preincubation of MDMs with the intracellular Ca²⁺ chelator MAPTAM (12.5 μM), followed by placement in Ca²⁺-free HBSS, 3 mM EGTA (EHBSS), completely inhibited the increase in [Ca²⁺]_c due to COZ (data not shown). To test the hypothesis that COZ-induced [Ca²⁺]_c elevations were dependent on stimulation of the β₂-integrins CR3 (CD11b/CD18) and CR4 (CD11c/CD18), MDMs were preincubated with F(ab′)₂ fragments of α-CD18 mAb (H52). Subsequent addition of COZ did not cause a significant change in [Ca²⁺]_c (Fig. 1 B). Inhibition of the increase in [Ca²⁺]_c by α-CD18 F(ab′)₂ fragments was specific for CR-dependent stimuli, as there was no effect on the [Ca²⁺]_c elevation stimulated by platelet activating factor (PAF; Fig. 1 B). These experiments demonstrate that, similar to neutrophils 35,36,42, CR-dependent stimulation of human macrophages with COZ results in a marked increase in [Ca²⁺]_c, which is derived from both intracellular and extracellular Ca²⁺ pools. Furthermore, the β₂-integrins, CR3 and CR4, are responsible for the majority of macrophage CR-stimulated Ca²⁺ signaling.

In the presence of serum, M. tuberculosis is opsonized with C3bi and, to a lesser extent, C3b via the alternative pathway of complement (5, 43, and data not shown). The effect of binding and phagocytosis of the virulent Erdman strain of M. tuberculosis on MDM [Ca²⁺]_c levels was determined exactly as noted above for COZ. Incubation of MDM with live, C-op M. tuberculosis at an MOI of 10:1 did not result in any significant change in macrophage [Ca²⁺]_c (Fig. 2 A; fold-increase in [Ca²⁺]_c = 1.01; range, 1.0–1.25-fold; n = 18). To determine whether the failure of M. tuberculosis to elicit an increase in MDM [Ca²⁺]_c was due to a defect in either intracellular Ca²⁺ stores or capacitative Ca²⁺ entry via macrophage plasma membrane Ca²⁺ channels, we analyzed the effect of thapsigargin on M. tuberculosis–infected MDMs. Addition of 1 μM thapsigargin resulted in a prompt rise in MDM [Ca²⁺]_c (Fig. 2 A), the magnitude and duration of which were comparable to that of uninfected MDMs. This response to thapsigargin confirmed the adequacy of both intracellular Ca²⁺ stores and the capacitative coupling of store depletion to the influx of extracellular Ca²⁺ in M. tuberculosis–infected MDMs.

The lack of an increase in macrophage [Ca²⁺]_c was not due to a failure to bind or ingest M. tuberculosis. At an MOI of 10:1, the mean (± SEM) number of ingested bacilli per macrophage was 5.36 ± 0.41, and 73 ± 4% of MDMs ingested at least one tubercle bacillus 21. To ensure that each MDM phagocytosed at least one tubercle bacillus, select single-cell [Ca²⁺]_c determinations were conducted with increased MOIs of 30:1 and 100:1. At these higher levels of infection, each MDM phagocytosed at least one bacillus, as determined by subsequent staining of cell monolayers with auramine-rhodamine (data not shown). However, even at MOIs of 30:1 (data not shown) and 100:1 (Fig. 2 B), M. tuberculosis did not result in any change in macrophage [Ca²⁺]_c. To complement the single-cell determinations of [Ca²⁺]_c, the aperture of the spectrofluorometer was adjusted to encompass a population of 15–20 MDMs per sample to determine the average [Ca²⁺]_c in resting and M. tuberculosis–infected macrophages. Similar to the results of the single-cell analysis, the mean [Ca²⁺]_c of a population of macrophages did not exhibit a significant change in [Ca²⁺]_c upon incubation with M. tuberculosis (Fig. 2 C).

The defect in macrophage Ca²⁺ signaling was not restricted to the Erdman strain of M. tuberculosis. Infection of MDMs with a second, well characterized virulent strain, H37Rv M. tuberculosis, also occurred without a significant alteration of [Ca²⁺]_c (Fig. 2 D; fold-increase in [Ca²⁺]_c = 1.0; n = 4). These results suggest that the lack of initiation of Ca²⁺ signaling may be a general property of pathogenic M. tuberculosis. The ability to test this hypothesis is limited somewhat by the lack of an avirulent strain of M. tuberculosis. The attenuated H37Ra strain of M. tuberculosis exhibits decreased virulence in animal models of tuberculosis 44. However, its interactions with human mononuclear phagocytes in vitro have variably been reported as either highly similar 45,46 or very distinct 47,48 from those of virulent M. tuberculosis strains, both in terms of its intracellular survival and ability to modify immune responses. Incubation of Fura2-loaded human macrophages with H37Ra M. tuberculosis did not result in a detectable change in [Ca²⁺]_c (data not shown; fold-increase over basal [Ca²⁺]_c = 1.0; n = 7). Thus, the three strains of M. tuberculosis, Erdman, H37Rv, and H37Ra, were similar in terms of their lack of effect on basal levels of [Ca²⁺]_c in human macrophages.

Ca²⁺-mediated signal transduction is characterized by a complex series of positive and negative regulatory circuits, as well as distinct temporal and spatial determinants of signal propagation 49,50,51,52. To determine whether the defect in Ca²⁺ signaling accompanying infection with M. tuberculosis resulted in a global depression of macrophage Ca²⁺-dependent signal transduction, we tested the response of infected MDMs to PAF, a potent Ca²⁺-mobilizing ligand that binds to a G protein–coupled receptor 53. 10 min after infection with Erdman M. tuberculosis, macrophages were incubated with 100 nM PAF, and levels of [Ca²⁺]_c were determined via fluorescence of Fura2. As demonstrated in Fig. 3 A, PAF induced a rapid and significant rise in [Ca²⁺]_c in infected macrophages that did not differ in onset, amplitude, or duration from the response of uninfected cells to PAF (data not shown). Similarly, addition of PAF concurrent with Erdman M. tuberculosis also resulted in an intact [Ca²⁺]_c response (Fig. 3 B). These PAF-induced elevations in [Ca²⁺]_c indicate that infection with M. tuberculosis does not render the macrophage refractory to Ca²⁺-mediated signal transduction. To exclude a potential inhibitory effect of M. tuberculosis that could be specific to CR-induced elevations in [Ca²⁺]_c, we determined whether infected macrophages exhibited an altered [Ca²⁺]_c response to COZ. Similar to the PAF-stimulated MDMs noted above, COZ-induced elevations in macrophage [Ca²⁺]_c occurred normally in the presence of prior (Fig. 3 C) or concurrent (data not shown) infection with M. tuberculosis. Therefore, the lack of increase in [Ca²⁺]_c levels during M. tuberculosis infection is not accompanied by alterations in [Ca²⁺]_c signaling by either particulate or soluble stimuli, which utilize the same or different classes of macrophage cell-surface receptors for their initiation. Our results do not exclude the possibility that M. tuberculosis may alter more distal aspects of Ca²⁺-mediated signal transduction by these or other stimuli.

The specific virulence determinants that enable M. tuberculosis to survive within the phagosomes of human macrophages are unknown. In addition, there are no avirulent strains of M. tuberculosis that may be used to define the molecular mechanisms that regulate essential pathogenic interactions between tubercle bacilli and mononuclear phagocytes. Despite these limitations, considerable evidence indicates that the failure of M. tuberculosis–containing phagosomes to mature into acidic microbicidal phagolysosomes is an important component of tuberculous pathogenesis 24,54,55. Clemens and Horwitz have demonstrated that this inhibition of phagosomal maturation is dependent on the viability of M. tuberculosis, as phagosomes containing heat-killed M. tuberculosis develop into mature phagolysosomes 23,24. We tested the hypothesis that the M. tuberculosis–induced inhibition of macrophage Ca²⁺ signaling would demonstrate a similar requirement for bacterial viability. Erdman M. tuberculosis was killed by heating to 100°C for 10 min, followed by opsonization in autologous, nonimmune serum as described above for live bacilli 23,24. Particular care was taken to ensure that the preparation of heat-killed M. tuberculosis consisted of >95% single bacilli, as noted in Materials and Methods. The loss of viability of heat-killed M. tuberculosis was verified by absence of growth on 7H11 agar. Heat-killed Erdman M. tuberculosis induced a rapid and significant rise in macrophage [Ca²⁺]_c (Fig. 4 A; fold-increase in [Ca²⁺]_c = 3.8; range, 2.1–6.5-fold; n = 16), which closely resembled that induced by COZ. Utilization of an alternate protocol for heat killing (80°C, 60 min; reference 5) resulted in similar stimulation of increased [Ca²⁺]_c by dead, C-op M. tuberculosis (data not shown). The increase in levels of macrophage [Ca²⁺]_c induced by heat-killed M. tuberculosis was completely inhibited by preincubation of these cells with F(ab′)₂ fragments of α-CD18 mAb (Fig. 4 B), indicating a major role for CR3 and/or CR4 in the initiation of this response. Studies with Ca²⁺-free media (Fig. 4 C) and intracellular Ca²⁺ buffering (Fig. 4 D) indicated that the increase in [Ca²⁺]_c stimulated by heat-killed M. tuberculosis resulted from both release of Ca²⁺ from intracellular stores as well as influx of extracellular Ca²⁺.

As heat killing of M. tuberculosis may induce changes in mycobacterial surface structures that could alter MDM Ca²⁺ signaling by mechanisms other than the loss of bacterial viability, similar studies were conducted with M. tuberculosis that had been killed by gamma irradiation. Incubation of MDMs with gamma-irradiated M. tuberculosis resulted in a prompt increase in [Ca²⁺]_c (Fig. 4 E; fold-increase in [Ca²⁺]_c = 2.9; range, 2.6–3.4; n = 6) that was indistinguishable from that induced by heat-killed tubercle bacilli. These studies support the hypothesis that mycobacteria-induced inhibition of macrophage Ca²⁺ signaling requires viability of M. tuberculosis.

To test the hypothesis that the receptors that mediate phagocytosis of M. tuberculosis are determinants of macrophage Ca²⁺ signaling, Erdman M. tuberculosis was incubated with polyclonal rabbit Ab to its cell wall glycolipid, LAM, to confer phagocytosis by macrophage FcγRs. After opsonization with anti-LAM Ab, mycobacteria were washed in PBS, resuspended in CHBSS, and counted to ensure that single bacilli comprised at least 95% of the preparation. Addition of Ab-op M. tuberculosis induced a prompt and significant increase in macrophage [Ca²⁺]_c (Fig. 5 A, fold-increase in [Ca²⁺]_c = 3.3; range, 2.1–5.0-fold; n = 9). In control experiments conducted with MDMs from the same donors, M. tuberculosis incubated with preimmune serum or irrelevant rabbit polyclonal Ab (antimyeloperoxidase Ab) did not induce a change in macrophage [Ca²⁺]_c (data not shown). Opsonization of M. tuberculosis with either of two mAbs to LAM (CS-35 or CS-40) resulted in similar elevations in macrophage [Ca²⁺]_c when added to Fura2-loaded macrophages (Fig. 5 B and data not shown). Parallel control experiments conducted with M. tuberculosis that had been incubated with isotype-matched irrelevant mAb (mouse myeloma IgG₁) did not result in stimulation of MDM Ca²⁺ signaling (data not shown). The difference in macrophage Ca²⁺ signaling induced by Ab-op versus C-op M. tuberculosis was not due to differences in mycobacterial adherence or phagocytosis. As all adherent bacilli were phagocytosed over the 30-min course of the experiment, the comparative interactions of different preparations of M. tuberculosis could be assessed with two parameters: (a) the phagocytic index (the number of bacilli phagocytosed per macrophage) and (b) the number of macrophages that ingest at least one bacillus. At an MOI of 10:1, neither of these parameters differed between the Ab- and C-op (live or dead) M. tuberculosis: mean phagocytic index = 5.46 ± 0.67 bacilli/MDM, and 78 ± 5% of MDMs ingested at least one tubercle bacillus.

Taken together, these results demonstrated that both the viability of M. tuberculosis and the receptors that mediate its phagocytosis were significant determinants of macrophage Ca²⁺-mediated signal transduction. Ingestion of live M. tuberculosis via CRs was not accompanied by detectable changes in levels of macrophage [Ca²⁺]_c, whereas phagocytosis of either dead bacilli via CRs or live, Ab-op M. tuberculosis by FcγR was associated with significant and prolonged increases in [Ca²⁺]_c.

To evaluate the hypothesis that mycobacteria-induced inhibition of CR-dependent Ca²⁺ signal transduction contributes to the intracellular survival of M. tuberculosis, we compared the viability of C- and Ab-op bacilli at serial time points after infection of human macrophages. MDMs adherent to collagen-coated coverslips were infected at an MOI of 1:1 with either C- or Ab-op Erdman M. tuberculosis at 37°C for 1 h. The level of phagocytosis did not differ between these two groups (data not shown). 1 h after infection, the monolayers were washed repeatedly to remove nonadherent bacilli and repleted with fresh media. The viability of intracellular M. tuberculosis was assessed 1, 48, and 96 h after infection by determination of CFUs. As prolonged in vitro culture of MDMs is associated with detachment of a minor fraction of the cells from the monolayer, lysates of both adherent macrophages and detached cells in the supernatant were combined for each sample, as per Paul et al. 46. As early as 1 h after infection, Ab-op M. tuberculosis exhibited significantly decreased viability compared with C-op bacilli (reduction in CFUs of 43%; range, 31–58% reduction; P < 0.01; n = 6; Fig. 5 C and Table). As there was wide variability between macrophages from different individuals in the absolute number of mycobacterial CFUs at all time points tested (including time = 0), the data on intracellular viability of M. tuberculosis have been presented both as (a) a percentage of the specific control value for C-op M. tuberculosis for each donor ([CFU of Ab-op bacilli/CFU of C-op bacilli] × 100%; Fig. 5 C) and (b) the raw data for CFU of both M. tuberculosis preparations for each individual set of donor macrophages (Table). This difference in intramacrophage survival between Ab- and C-op M. tuberculosis increased progressively with increasing duration of infection. 96 h after infection, there was a 78% decrease in CFUs derived from macrophages infected with Ab-op M. tuberculosis compared with MDMs infected with C-op bacilli (range, 64–86% reduction in CFUs; P < 0.001; Fig. 5 C and Table). These results demonstrate that the lack of increase in macrophage [Ca²⁺]_c during phagocytosis of C-op M. tuberculosis is associated with increased intracellular survival of mycobacteria.

The previous set of experiments demonstrated that the opsonin on M. tuberculosis and, consequently, the class of phagocytic receptor that primarily mediated its ingestion were major determinants of the extent of mycobacterial survival within human MDMs. Although these observations correlated with the difference in Ca²⁺ mobilization between Ab- and C-op M. tuberculosis, a causal relationship between [Ca²⁺]_c and intracellular mycobacterial viability cannot be inferred, as multiple differences exist between FcγR- and CR-mediated phagocytosis 8,9. Therefore, to directly test our hypothesis, we used the Ca²⁺ ionophore, A23187, to modulate the cytosolic Ca²⁺ levels in human MDMs during phagocytosis of C-op M. tuberculosis. After loading with Fura2 and washing to remove unincorporated dye, MDMs were placed in HBSS solutions in which the concentration of extracellular free Ca²⁺ was buffered in the range of 225–700 nM with EGTA. These levels of Ca²⁺ were chosen to approximate the [Ca²⁺]_c that occurred in MDMs stimulated by COZ, C-op dead M. tuberculosis, and Ab-op mycobacteria. Addition of 1 μM A23187 resulted in a rapid equilibration of the intracellular and extracellular Ca²⁺ concentrations (Fig. 6 A). To mimic the temporally restricted elevation of [Ca²⁺]_c initiated by the particulate stimuli noted above, the effects of A23187 were reversed after 20 min by addition of phosphatidylcholine vesicles (20 μg/ml) 28, which resulted in a rapid return of [Ca²⁺]_c to a level approximating that of resting macrophages (Fig. 6 A). To examine the effects of cytosolic Ca²⁺ levels on survival of intracellular M. tuberculosis, parallel sets of infected MDM monolayers were lysed and viable mycobacteria quantitated 24 and 48 h after infection by analysis of CFUs. Compared with untreated, M. tuberculosis–infected macrophages, MDMs incubated with A23187 during infection contained ∼50% less viable M. tuberculosis at the 24- and 48-h time points (Fig. 6 B and Table). These results were not due to a direct bactericidal effect of A23187, as incubation of M. tuberculosis suspensions in the calcium ionophore, followed by addition of phosphatidylcholine vesicles, under the exact conditions applied to infected macrophages did not result in alteration of mycobacterial viability (data not shown). These results indicated that ionophore-induced elevation of macrophage [Ca²⁺]_c during phagocytosis of C-op M. tuberculosis was associated with decreased intracellular survival of the bacilli.

A key aspect of tuberculous pathogenesis is the ability of M. tuberculosis to limit the maturation of its phagosome, thereby preventing the development of microbicidal phagolysosomes 10,11,12,23,24,54,55. We tested the hypothesis that mycobacterial inhibition of macrophage Ca²⁺ signaling contributes to retardation of phagosomal maturation (inhibition of phagosome–lysosome [P–L] fusion) by (a) characterizing the degree of maturation of phagosomes containing either live or killed C-op M. tuberculosis and (b) determining the effects of modulation of [Ca²⁺]_c on P–L fusion. The extent of maturation of M. tuberculosis–containing phagosomes 24 h after infection was characterized by confocal microscopy, using three lysosomal protein markers (cathepsin D, LAMP-1, CD63), combined with the determination of phagosomal pH with the acidophilic fluorophore, LysoTracker Red. The three protein markers were used in combination, because use of a single marker can provide ambiguous results. For example, LAMP-1 localizes to both late endosomes and lysosomes 24. LysoTracker Red was employed for assessment of phagosomal acidification, as this fluorophore is stable to fixation, ensuring that biosafety conditions are maintained during confocal microscopy.

24 h after infection of human MDMs, live, C-op M. tuberculosis was located in immature phagosomes that exhibited low amounts of the lysosomal protein markers. The percentage of phagosomes positive for cathepsin D, LAMP-1, and CD63 were 32, 37, and 25%, respectively (Fig. 7). Additionally, only 41% of phagosomes containing live M. tuberculosis colocalized with LysoTracker Red. These results are in agreement with previous characterizations of the maturational state of M. tuberculosis–containing phagosomes in macrophages, as determined by epifluorescence, confocal immunofluorescence, and cryoimmunoelectron microscopy 10,11,12,23,24,54,55.

To further evaluate the potential causal role of macrophage cytosolic Ca²⁺ in P–L fusion, the maturation of phagosomes containing live, C-op M. tuberculosis was determined after transient elevation of [Ca²⁺]_c with A23187, followed by quenching with phosphatidylcholine vesicles. Ionophore-induced elevation of [Ca²⁺]_c to ∼500 nM for 20 min during phagocytosis of live, C-op M. tuberculosis resulted in a striking reversal of the block in phagosomal maturation. The percentage of phagosomes positive for cathepsin D increased from 32 (control) to 92%, LAMP-1 positivity increased from 37 to 82%, and CD63 positivity increased from 25 to 83% (Fig. 7). Elevation of [Ca²⁺]_c also promoted increased phagosomal localization of LysoTracker Red, from a control value of 41 to 89% in MDMs treated with A23187. Elevation of [Ca²⁺]_c was required for the A23187-induced increase in P–L fusion, as incubation of macrophages in EHBSS during ionophore treatment resulted in a profile of phagosomal staining for the lysosomal protein markers and LysoTracker Red that was indistinguishable from values for control, untreated MDMs (Fig. 7).

In marked contrast to the intracellular compartmentation of live tubercle bacilli, phagosomes containing dead (gamma-irradiated) M. tuberculosis progressed to fully mature phagolysosomes, as determined by high levels of all three lysosomal protein markers (Fig. 8). 88% of phagosomes containing killed M. tuberculosis were positive for cathepsin D, whereas the corresponding values for LAMP-1 and CD63 were 77 and 76%, respectively. 88% of these phagosomes accumulated LysoTracker Red, consistent with their acidification. Incubation of macrophages in EHBSS or chelation of cytosolic Ca²⁺ with MAPTAM resulted in failure of phagosomes containing dead M. tuberculosis to accumulate lysosomal protein markers (Fig. 8). Compared with the percentage of phagosomes positive for cathepsin D, LAMP-1, and CD63 in Ca²⁺-containing media noted above, removal of extracellular Ca²⁺ resulted in significantly less colocalization with all three lysosomal protein markers: 66, 30, and 47%, respectively. Chelation of intracellular Ca²⁺ with 12.5 μM MAPTAM resulted in even more pronounced reductions in phagosomal accumulation of lysosomal markers: cathepsin D, 37%; LAMP-1, 24%; and CD63, 38%. As MAPTAM produces more significant reductions in basal and stimulated [Ca²⁺]_c compared with EGTA (Fig. 4), these results are fully consistent with the hypothesis that [Ca²⁺]_c regulates the maturation of phagosomes containing dead M. tuberculosis. Interestingly, MAPTAM but not EGTA produced significant decreases in accumulation of LysoTracker Red: untreated control, 88%; MAPTAM, 49%; and EGTA, 86% (Fig. 8). As removal of extracellular Ca²⁺ reduces but does not eliminate the increase in [Ca²⁺]_c induced by killed M. tuberculosis, these results are consistent with the hypothesis that a lesser increase in [Ca²⁺]_c is required for phagosomal acidification than for accumulation of lysosomal protein markers (especially LAMP-1 and CD63).

In summary, the results of characterization of phagosome maturation via confocal microscopy strongly support the hypothesis that levels of cytosolic Ca²⁺ regulate P–L fusion in M. tuberculosis–infected human macrophages. In all cases, elevation of macrophage [Ca²⁺]_c correlated with maturation of M. tuberculosis–containing phagosomes to phagolysosomes, and lack of elevation of [Ca²⁺]_c correlated with incomplete phagosomal maturation. Furthermore, ionophore-induced increases in [Ca²⁺]_c and the accompanying maturation of phagosomes containing live C-op M. tuberculosis correlated with decreased survival of mycobacteria within human macrophages.

Macrophages possess multiple microbicidal mechanisms to eliminate phagocytosed microorganisms and, consequently, represent a strategic target for inactivation by potential pathogens 56. The molecular mechanisms that allow M. tuberculosis to successfully survive and replicate within mononuclear phagocytes are unknown. Our overall hypothesis is that Ca²⁺-dependent signaling mechanisms are potential targets for inhibition of macrophage activation by M. tuberculosis, as [Ca²⁺]_c is a critical regulator of several antimicrobial responses, including generation of reactive oxygen and nitrogen intermediates, secretion of microbicidal proteins and peptides, and synthesis of antimycobacterial cytokines, such as TNF-α 13,14,57.

This study demonstrates that multiple strains of pathogenic M. tuberculosis inhibit Ca²⁺-mediated signal transduction during infection of human macrophages. Inhibition of macrophage Ca²⁺ signaling is tightly coupled to the failure of mycobacterial phagosomes to mature into acidic, microbicidal phagolysosomes and to successful intracellular survival of M. tuberculosis. Two determinants of mycobacteria-induced inhibition of macrophage Ca²⁺ signaling have been defined. First, the bacilli must be viable, as killing of M. tuberculosis by heat or gamma irradiation reverses the inhibition of Ca²⁺-mediated signal transduction. Although the basis of this requirement is unknown, the dependence on mycobacterial viability has previously been demonstrated for the inhibition of P–L fusion in M. tuberculosis–infected human macrophages 12,23,24. Second, infection of macrophages in the absence of increased [Ca²⁺]_c is specific for phagocytosis via CRs. Redirecting the phagocytosis of M. tuberculosis to FcγRs, via opsonization with specific polyclonal or monoclonal Abs, reverses mycobacteria-induced impairment of macrophage Ca²⁺ signaling, and, more importantly, reduces the intracellular survival of M. tuberculosis within human MDMs. The reduction in the intracellular survival of Ab-op bacilli was not due to a difference in phagocytosis, as both the phagocytic index (the number of bacilli ingested per macrophage) and the percentage of MDMs that phagocytosed at least one bacillus did not differ between the two groups.

Although mechanisms other than induction of elevated [Ca²⁺]_c may contribute to the decreased viability of Ab-op bacilli, direct evidence for a causal role of [Ca²⁺]_c in regulating the survival of M. tuberculosis within human macrophages was obtained with the calcium ionophore, A23187. Thus, in this in vitro model of primary infection of human macrophages, the lack of an increase in [Ca²⁺]_c during CR-mediated phagocytosis correlated with inhibition of P–L fusion and increased intracellular survival of M. tuberculosis. Conversely, elevation of [Ca²⁺]_c was associated with increased P–L fusion and reduced intramacrophage viability. The essential role of [Ca²⁺]_c in triggering multiple phagocyte antimicrobial defenses suggests that the inhibition of phagocytosis-initiated Ca²⁺ signaling confers a survival advantage on M. tuberculosis at the time of its entry into macrophages. Immunoelectron microscopy of M. tuberculosis–containing phagosomes supports the hypothesis that the bacilli's protected “intracellular niche” is established at a relatively early time point during infection of macrophages 24,54. The large number of Ca²⁺-dependent biochemical reactions and cellular functions suggests that M. tuberculosis–induced inhibition of changes in [Ca²⁺]_c may compromise several components of macrophage activation and antimicrobial function.

The use of the term “inhibition” to characterize the lack of increase in [Ca²⁺]_c during phagocytosis of M. tuberculosis is meant in an operational sense, as the mechanism remains unknown. Lack of initiation of a Ca²⁺ signaling pathway or its rapid termination could both yield the observed results. As CRs, especially CR3, are the primary mediators of phagocytosis of M. tuberculosis in human MDMs 5 and because other C3b/bi-opsonized particles, including dead M. tuberculosis, stimulate a rise in [Ca²⁺]_c, our working model is best summarized by the question, How does live M. tuberculosis inhibit CR-mediated increases in [Ca²⁺]_c? Evidence in favor of this model, particularly the comparison between live and dead C-op M. tuberculosis as a means to understand the pathogenesis of tuberculosis, include: (a) live and dead C-op M. tuberculosis are phagocytosed to the same extent by human MDMs 5, (b) the extent of phagocytosis of live and dead M. tuberculosis is inhibited to the same extent by anti-CR3 Abs 5, (c) the level of C3 deposition does not differ between live and heat-killed bacilli (data not shown), and (d) anti-CD18 F(ab′)₂ fragments eliminate the increase in [Ca²⁺]_c stimulated by dead M. tuberculosis or COZ. This model encompasses the possibility that additional interactions between live M. tuberculosis and human MDMs other than ligation of mycobacterial surface-bound C3b/bi by CRs may contribute to the inhibition of Ca²⁺ signaling.

As the fluorescent detection of [Ca²⁺]_c is highly sensitive, our hypothesis is that no Ca²⁺ signal is initiated during phagocytosis of live M. tuberculosis. However, the biochemical signals that normally link CRs to increases in [Ca²⁺]_c are unknown, and, therefore, we cannot ascertain whether these intermediate steps are “not initiated” or “initiated but inhibited.” These mechanistic uncertainties are an additional reason that we have used the more general phrase, “inhibition of Ca²⁺ signaling.” However, we recognize that further definition of the mechanism(s) by which CR-induced phagocytosis of C-op M. tuberculosis occurs in the absence of a change in macrophage [Ca²⁺]_c may necessitate a revision of our current model and terminology.

Comparison of our results with those recently reported by Majeed et al. 19 illustrates both the similarities and differences in the interactions of M. tuberculosis with mononuclear phagocytes versus neutrophils (PMNs). Although neutrophil ingestion of the attenuated H37Ra strain of M. tuberculosis also occurred in the absence of a rise in [Ca²⁺]_c, PMNs killed 73% of phagocytosed tubercle bacilli in 2 h 19. Whether induction of a rise in Ca²⁺ via physiologic or pharmacologic intervention would augment PMN P–L fusion or bactericidal activity toward M. tuberculosis was not reported, and no virulent strains of M. tuberculosis were used 19. Furthermore, M. tuberculosis does not successfully parasitize human neutrophils, and several studies have demonstrated that PMNs kill intracellular M. tuberculosis by both oxygen-dependent and -independent mechanisms 60,61,62. Finally, caution is required in comparing Ca²⁺-mediated signal transduction of neutrophils with that of macrophages, as these two classes of phagocytes have been reported to differ in the Ca²⁺ dependence of antimicrobial functions 26,59.

Zimmerli et al. have reported that human macrophages do not require an increase in [Ca²⁺]_c for fusion of lysosomes with phagosomes containing COZ, coagulase-negative staphylococci, or the vaccine strain M. bovis BCG 27. The differences between their results and ours may be due, at least in part, to differences in both the methods of measuring phagosomal maturation and the characteristics of the phagocytosed particles. Zimmerli et al. used colocalization of the particles with LAMP-1 and endocytosed rhodamine dextran to define maturation of phagolysosomes. However, neither LAMP-1 nor dextran localize specifically to lysosomes, as both markers label late endosomes as well 24,27. Perhaps the Ca²⁺ requirement for P–L fusion is influenced by characteristics of the particle, e.g., its virulence; M. tuberculosis is a highly virulent intracellular pathogen, whereas coagulase-negative staphylococci are extracellular pathogens of low virulence, BCG is nonpathogenic, and zymosan is a cell wall preparation from the nonpathogenic yeast, Saccharomyces cerevisiae. Finally, the lack of statistically significant differences in P–L fusion between control and Ca²⁺-buffered MDMs in their study 27 may have been influenced by the small sample size, as the average decrease in P–L fusion in MAPTAM-treated MDMs compared with control MDMs was as great as 23%, and the standard deviations ranged from 30 to 100% of the mean values. Further studies will be required to clarify the variables that affect the Ca²⁺ dependence of phagosomal maturation in human macrophages.

A fascinating aspect of M. tuberculosis–induced inhibition of Ca²⁺ signaling is its particle specificity during concurrent or subsequent addition of a Ca²⁺-mobilizing stimulus. Although MDMs did not generate an increase in [Ca²⁺]_c during infection by live, C-op bacilli, these same cells maintained the capacity to respond to other phagocytosed particles (COZ, killed or Ab-op M. tuberculosis) or soluble stimuli (PAF) by increasing their levels of [Ca²⁺]_c. Therefore, viable, C-op M. tuberculosis does not introduce a generalized defect in the Ca²⁺ signaling pathways of human MDMs. These results suggest that CR-induced increases in [Ca²⁺]_c are spatially restricted to each specific phagocytic event, i.e., each forming phagosome, although further studies will be required to directly test this hypothesis. This proposed focal nature of phagocytosis-associated Ca²⁺ signaling or, in the case of M. tuberculosis, the lack thereof, is consistent with previous studies demonstrating the tightly regulated spatial constraints of Ca²⁺-mediated signal transduction 36,50,51,63. In fact, Stendahl et al. 36 have recently demonstrated that during phagocytosis of COZ, [Ca²⁺]_c levels are highest in the periphagosomal region 34.

Whereas our data demonstrating increased P–L fusion after Ab opsonization of M. tuberculosis are in agreement with those of Armstrong and Hart 11, our studies differ with respect to the consequences for mycobacterial survival. In this study, opsonization of M. tuberculosis with specific monoclonal or polyclonal Abs resulted in significant decreases in intracellular viability within human macrophages. In contrast, Armstrong and Hart demonstrated that survival within murine macrophages was similar for serum- and Ab-op M. tuberculosis 11. As numerous investigators have documented significant differences between the tuberculocidal capacities of human versus murine macrophages (for review see references 3, 6, 46–48), we hypothesize that this species specificity is a major factor contributing to the contrasting effects of Ab opsonization on mycobacterial survival noted in our two studies. Zimmerli et al. recently demonstrated that Ab-mediated inhibition of individual CRs or the mannose receptor did not alter the intracellular survival of M. tuberculosis within human MDMs 58. This study differs from ours in two respects. First, the effect of FcγR-mediated phagocytosis on mycobacterial survival was not determined, and second, receptor-blocking reagents were used to direct phagocytosis of M. tuberculosis to unblocked receptor 58. However, blocking reagents may introduce confounding effects by stimulating the receptors to which they bind, and it is often difficult to block multiple receptor classes ligated by complex particles, such as M. tuberculosis. In contrast, opsonization of M. tuberculosis with specific ligands provides a direct, physiologically relevant analysis of the impact of individual receptor classes on the intracellular survival of M. tuberculosis.

The lack of Ca²⁺ mobilization during ingestion of M. tuberculosis may represent an important mechanism of immune evasion that contributes to its survival within human macrophages. As the specific mechanism(s) by which human macrophages kill intracellular M. tuberculosis is unknown, it is difficult at present to define the means by which inhibition of Ca²⁺ signaling promotes mycobacterial survival, although inhibition of P–L fusion is likely to contribute. Despite these challenges, characterization of the molecular mechanisms responsible for M. tuberculosis–induced alterations in macrophage Ca²⁺ signaling and its specific contribution to intracellular survival will provide important insights into the pathogenesis of tuberculosis and may contribute to the development of novel therapies to treat this formidable disease.

We thank our colleagues in the Inflammation Program at the University of Iowa, particularly William M. Nauseef, Jerrold P. Weiss, and Lee-Ann Allen, for their many helpful discussions and thoughtful critiques. We are especially grateful to Lee-Ann Allen for her guidance with confocal microscopy. We also thank Larry S. Schlesinger and Thomas Kaufman for assistance with the determination of C3 deposition on M. tuberculosis, Patrick J. Brennan and John T. Belisle for their generous provision of several mycobacterial reagents, and William M. Nauseef for his kind gift of polyclonal Ab to myeloperoxidase.

Support for these studies was provided by a Veterans Affairs (VA) Merit Review Grant, VA Career Development Award, and National Institutes of Health grant (NIH) AI18571-17 to D.J. Kusner, a VA Merit Review Grant to G. Denning, and an institutional NIH National Research Service Award to the Medical Scientist Training Program at the University of Iowa to Z.A. Malik.