CFP-10 from Mycobacterium tuberculosis Selectively Activates Human Neutrophils through a Pertussis Toxin-Sensitive Chemotactic Receptor

Abstract

Upon infection with Mycobacterium tuberculosis, neutrophils are massively recruited to the lungs, but the role of these cells in combating the infection is poorly understood. Through a type VII secretion system, M. tuberculosis releases a heterodimeric protein complex, containing a 6-kDa early secreted antigenic target (ESAT-6) and a 10-kDa culture filtrate protein (CFP-10), that is essential for virulence. Whereas the ESAT-6 component possesses multiple virulence-related activities, no direct biological activity of CFP-10 has been shown, and CFP-10 has been described as a chaperone protein for ESAT-6. We here show that the ESAT-6:CFP-10 complex induces a transient release of Ca²⁺ from intracellular stores in human neutrophils. Surprisingly, CFP-10 rather than ESAT-6 was responsible for triggering the Ca²⁺ response, in a pertussis toxin-sensitive manner, suggesting the involvement of a G-protein-coupled receptor. In line with this, the response was accompanied by neutrophil chemotaxis and activation of the superoxide-producing NADPH-oxidase. Neutrophils were unique among leukocytes in responding to CFP-10, as monocytes and lymphocytes failed to produce a Ca²⁺ signal upon stimulation with the M. tuberculosis protein. Hence, CFP-10 may contribute specifically to neutrophil recruitment and activation during M. tuberculosis infection, representing a novel biological role for CFP-10 in the ESAT-6:CFP-10 complex, beyond the previously described chaperone function.

INTRODUCTION

Human tuberculosis is caused by the bacterium Mycobacterium tuberculosis, which enters its host through the airways and can establish infection in the lungs. Inhalation of M. tuberculosis-containing aerosols leads to one of three outcomes—clearance of the infection by the innate immune system, establishment of a latent disease state where viable bacteria are contained within granulomas by the combined action of the innate and adaptive immune systems, or development of active tuberculosis (1, 2). The importance of innate immunity in overcoming an M. tuberculosis infection is illustrated by the fact that up to 50% of exposed individuals are believed to clear the infection without involvement of the adaptive immune system (3). The alveolar macrophage is considered the main host cell of M. tuberculosis, and a lot of effort has been put into elucidating how bacteria manage to survive inside these cells.

The short-lived and abundant neutrophils are the first leukocytes to respond to inflammatory stimuli. Although it is clear that massive recruitment of neutrophils occurs upon M. tuberculosis infection as a result of chemokine production at the site of infection and that neutrophils phagocytose great numbers of M. tuberculosis bacilli, much less is understood about their role in combating M. tuberculosis infection. Conflicting data exist, but it is becoming clear that neutrophils can have both protective and immunopathological effects in tuberculosis (4).

M. tuberculosis has a complex, lipid-rich cell wall in which five known type VII secretion systems are expressed, and these systems are optimized to effectively export material from the bacterial cytoplasm to the extracellular space. The best studied of the secretion systems is the so-called 6-kDa early secreted antigenic target (ESAT-6) secretion system 1 (ESX-1), which is encoded by region of difference 1 (RD1) in the mycobacterial genome and which is essential for M. tuberculosis virulence. The ESX-1 is responsible for the transfer to the surrounding milieu of a heterodimeric protein complex containing ESAT-6 (also known as EsxA) and a 10-kDa culture filtrate protein (CFP-10, also known as EsxB or M. tuberculosis-specific antigen 10 [MTSA-10]) (5–9). The CFP-10 contains a C-terminal sequence that enables secretion of the complex from the bacterial cytoplasm (10), and the complex is believed to dissociate under acidic conditions, e.g., inside a phagolysosome (11). The ESAT-6:CFP-10 complex is an essential virulence factor of both M. tuberculosis and M. marinum. It appears to have membrane-lysing activity (11–13), contributing to microbial escape from the phagosome into the macrophage cytoplasm (13–15), and to induce host cell necrosis and spread of bacteria to adjacent cells (12, 16). The complex has furthermore been implicated in inflammasome activation (17) and inhibition of Toll-like receptor (TLR) signaling (18, 19) in monocytes as well as in induction of interleukin-8 (IL-8) production in lung epithelial cells (20). The CFP-10 part of the complex has been described as a chaperone protein for ESAT-6, responsible for delivering its more biologically active binding partner to the site of action (11).

A recent study by Corleis et al. showed that upon phagocytosis by neutrophils, M. tuberculosis activates the antimicrobial artillery of these cells but then escapes by inducing neutrophil necrosis through an RD1-dependent mechanism (21). Very little is, however, known about the interaction between the RD1 gene product ESAT-6:CFP-10 and neutrophils. This prompted us to investigate the direct interaction between ESAT-6:CFP-10 and human neutrophils. We found that neutrophils were able to recognize the ESAT-6:CFP-10 complex and that CFP-10 rather than ESAT-6 was the component recognized by the cells. CFP-10 stimulation of neutrophils resulted in a transient release of Ca²⁺ from intracellular stores, accompanied by neutrophil chemotaxis and production of reactive oxygen species (ROS). The CFP-10-induced Ca²⁺ and ROS responses were sensitive to pertussis toxin (PtX), suggesting the involvement of a G-protein-coupled receptor (GPCR). Neutrophils specifically recognized CFP-10, and no Ca²⁺ signal was induced in monocytes or lymphocytes. Thus, the results of this study show that the CFP-10 component of ESAT-6:CFP-10 activates human neutrophils, suggesting direct proinflammatory activity that may be of importance for the M. tuberculosis-host interplay.

MATERIALS AND METHODS

Cell separation.

Peripheral blood neutrophils were separated from 1-day-old buffy coats from healthy blood donors by dextran sedimentation and Ficoll-Paque centrifugation according to the method of Boyum (22, 23). After separation and lysis of erythrocytes, the neutrophils were washed and resuspended in Krebs-Ringer glucose (KRG) phosphate buffer supplemented with Ca²⁺ (1 mM) and stored on melting ice until use. The purity of the preparation was typically around 95%. A whole-leukocyte preparation was obtained in the same way but with the Ficoll-Paque centrifugation step omitted. KRG buffer with Ca²⁺ (1 mM) was used for all experiments.

ESAT-6 and CFP-10.

Recombinant ESAT-6 and CFP-10, both purified from Escherichia coli, were purchased from BBI Solutions or Fitzgerald Laboratories. Purity and identity were confirmed by the manufacturers through Coomassie staining of protein samples on SDS-PAGE (ESAT-6, 10.2 kDa; CFP-10, 11 kDa) and reactivity with serum from tuberculosis patients on Western blots. Further, the preparations were analyzed by liquid chromatography-mass spectrometry (LC-MS/MS) (The Proteomics Core Facility at Sahlgrenska Academy, University of Gothenburg), confirming sequence identity. The endotoxin level was <1 endotoxin units (EU)/μg protein as determined by Limulus lysate assay (Clinical Microbiology, Sahlgrenska University Hospital). ESAT-6:CFP-10 complex (molecular mass, 21.2 kDa) was produced by mixing equimolar concentrations of the two proteins, as previously described (9, 19, 24).

Measurement of intracellular Ca²⁺ transients.

Ca²⁺ mobilization from intracellular stores was measured using flow cytometry (25). Neutrophils or mixed leukocytes (as indicated) were diluted in buffer supplemented with 1% fetal calf serum to 8 × 10⁶/ml and loaded with the Ca²⁺ indicator dyes Fluo-3 (Molecular Probes; fluorescence increases upon Ca²⁺ binding) and FuraRed (Molecular Probes; fluorescence decreases upon Ca²⁺ binding) at 37°C for 30 min. After two washing steps, the cells were diluted in buffer and stored on ice until use. Before each run, the cells were equilibrated at 37°C for 3 min. Fluo-3/FuraRed intensity was recorded continuously for 150 s using an Accuri C6 flow cytometer (BD Accuri). After 30 s, stimuli (M. tuberculosis products or controls) were added as indicated. Formyl-methionyl-leucyl-phenylalanine (fMLF) (Sigma) and ionomycin (Sigma) were used as positive controls. EGTA (Sigma) (2.5 mM) was used to deplete the medium of Ca²⁺. Analysis was performed using FlowJo software (v. 7.6.5; TreeStar), and results are presented as the ratio between Fluo-3 and FuraRed fluorescence intensities (normalized against the value at time = 0), reflecting the relative cytosolic Ca²⁺ concentration, over time.

Sytox green assay.

In order to assess neutrophil plasma membrane integrity after addition of CFP-10, the membrane-impermeable Sytox green DNA dye (Molecular Probes) was used. Neutrophils (5 × 10⁵ per well), diluted in buffer containing 2.5 μM Sytox green, were seeded in black 96-well plates. CFP-10 or the detergent Triton X-100 (TX100; Merck) (1%; positive control) was added to triplicate wells, and the plate was incubated at 37°C. Sytox green fluorescence was measured in a Mithras LB940 plate reader (Berthold Technologies) after 5, 30, and 120 min, and the median intensity value from each triplicate was used. Results are presented as percentages of TX100 at each time point.

LDH assay.

The effect of CFP-10 on neutrophil plasma membrane integrity was further assessed by the lactate dehydrogenase (LDH) release assay (cytotoxicity detection kit [LDH]; Roche), as described previously (26), after incubation of 5 × 10⁵ neutrophils with CFP-10 or 1% TX100 at 37°C for 2 h. Absorbance at 490 nm was measured, and results are presented as percentages of TX100 levels.

Imaging flow cytometry.

Visualization of neutrophil permeabilization by CFP-10 was performed by imaging flow cytometry (ImageStreamX MkII; Amnis). Neutrophils (1 × 10⁶) were incubated with CFP-10 at 37°C for 30 min. During the last 5 min of incubation, the membrane-impermeative DNA dye Sytox green (5 μM), which stains only permeabilized cells, and the membrane-permeative DNA dye DRAQ5 (10 μM, Abcam), which stains all cells, were included. The cells were put on ice, and 5,000 images per sample were acquired in the imaging flow cytometer. Analysis was performed in Ideas software (v. 6.0; Amnis). The proportion of Sytox green-positive cells was analyzed among the in-focus, DRAQ5-positive, single cells.

Transwell chemotaxis assay.

Neutrophil migration toward CFP-10 or fMLF (positive control) was analyzed by the use of a ChemoTx chemotaxis system (NeuroProbe) according to the manufacturer's instructions. Briefly, 30 μl chemoattractant or buffer supplemented with 0.3% bovine serum albumin was placed in a well, a filter with pore size of 3 μm was applied, and neutrophils (3 × 10⁴/sample diluted in buffer supplemented with 0.3% bovine serum albumin) were added on top of the filter. Each chemoattractant was run in triplicate and added below the filters. The plate was incubated at 37°C for 90 min, after which the number of cells in the well (both attached and in the supernatant) was assessed using microscopy and by quantification of LDH (as described above) after lysis of the cells was performed using 1% TX100. Median values were used, and results are expressed as percent cell migration, calculated as the percentages of the absorbance values obtained for a control well containing 3 × 10⁴ cells placed below the filter (100% cell control).

Measurement of NADPH-oxidase activity.

NADPH-oxidase activity was determined using isoluminol-enhanced chemiluminescence (27, 28). Chemiluminescence was measured in a six-channel Biolumat LB 9505 luminometer (Berthold Technologies) by the use of disposable polypropylene tubes with a 360-μl reaction mixture containing 2 × 10⁵ cells, 2 × 10⁻⁵ M isoluminol, and 2 U horseradish peroxidase. The tubes were equilibrated in the Biolumat luminometer for 10 min at 37°C, after which CFP-10 or fMLF (positive control) was added, and light emission, corresponding to extracellularly released O₂⁻, was recorded continuously. Disruption of the actin cytoskeleton was performed by inclusion of 100 ng/ml latrunculin A (Sigma) during equilibration.

Inhibition of G-protein-coupled receptors.

Pertussis toxin (PtX) was used to inhibit PtX-sensitive G-protein-coupled receptors (GPCRs) (29), and neutrophil activation in response to CFP-10 was measured in terms of Ca²⁺ mobilization as described above. During loading with Ca²⁺ indicator dyes, 500 ng/ml PtX (Sigma) was included in one of two vials, and after washing, PtX was again included in this vial for further incubation, along with the control vial, at 37°C. As the time required to achieve G-protein inhibition by PtX varies, inhibition was tested at 30-min intervals. This was done by stimulating PtX-treated or untreated neutrophils with fMLF (GPCR dependent) or phorbol myristate acetate (PMA; Sigma) (GPCR independent) and recording O₂⁻ production as described above using chemiluminescence. At the time point at which the response to fMLF was completely inhibited but that to PMA remained intact, the Ca²⁺ experiment was performed on the treated cells. The total time of incubation with PtX was in the range of 2.5 to 3 h.

Statistical analysis.

The numbers of independent experiments (each performed on cells from an individual donor) are stated in the figure legends, which also describe the statistical analyses performed (Graph Pad Prism v. 6.01; GraphPad Software). Statistically significant differences are indicated in the figures by a single asterisk (*; P = ≤0.05), double asterisks (**; P = ≤0.01), or triple asterisks (***; P = ≤0.001). Error bars in figures indicate the standard deviations.

RESULTS

Neutrophils recognize CFP-10, resulting in transient mobilization of intracellular Ca²⁺.

As a transient rise in intracellular Ca²⁺ is an initial signal of fundamental importance in activation of neutrophils and other leukocytes (30), we first determined the neutrophil response to the ESAT-6:CFP-10 complex using the mobilization of cytosolic Ca²⁺ as a readout. Stimulation of human neutrophils with the ESAT-6:CFP-10 complex gave rise to a Ca²⁺ signal comparable to that induced by formyl peptide receptor 1 (FPR1) agonist fMLF, used as a positive control (Fig. 1A). Surprisingly, a similar molarity of CFP-10 alone induced a Ca²⁺ signal identical to that evoked by the complex, while ESAT-6 alone induced no response (Fig. 1A). Thus, human neutrophils recognized and responded with a Ca²⁺ transient to the ESAT-6:CFP-10 complex, and the CFP-10 component was responsible for inducing this activity. The effect of CFP-10 was concentration dependent (Fig. 1B), with an activating concentration range similar to that used in previous studies investigating binding of ESAT-6 or CFP-10 to cells and activating effects of ESAT-6 (9, 19, 20). Having found that neutrophils recognize the CFP-10 component of ESAT-6:CFP-10, we performed subsequent experiments with CFP-10 alone.

FIG 1.

Cytosolic Ca²⁺ transients are induced by CFP-10 in neutrophils. Neutrophils were loaded with Fluo-3 and FuraRed, and the fluorescence intensity ratio of the two dyes, corresponding to the relative Ca²⁺ concentrations, was analyzed continuously during 150 s using flow cytometry. (A) Relative background Ca²⁺ levels were recorded for 30 s, after which stimuli were added (indicated by arrows). Neutrophils were stimulated with ESAT-6:CFP-10 complex (20 μg/ml), CFP-10 (10 μg/ml), ESAT-6 (10 μg/ml), or fMLF (10⁻⁷ M) as a positive control. The curves show the normalized ratios between Fluo-3 and FuraRed over time and are representative of the results of ≥3 independent experiments. (B) Relative peak levels of cytosolic Ca²⁺ in neutrophils stimulated with CFP-10 at the indicated concentrations or with buffer. Data are presented as peak normalized Fluo-3/FuraRed ratios and depict the means of the results of 4 to 10 experiments. A dashed line is shown at the background value. Statistical analysis was performed using ordinary one-way analysis of variance (ANOVA) and Dunnett's post hoc test, comparing each concentration to the control. AU, arbitrary units.

The Ca²⁺ transient induced by CFP-10 occurs independently of plasma membrane disturbance.

As it is well established that the ESAT-6:CFP-10 complex causes necrosis of macrophages (11–16), we next investigated the effect of CFP-10 on neutrophil viability. A plate-based assay employing the impermeable DNA dye Sytox green showed that no breach of plasma membrane integrity occurred with the CFP-10 doses and time spans used to evoke mobilization of intracellular Ca²⁺. At higher concentrations or with prolonged exposure, a permeabilizing effect of CFP-10 was noted (Fig. 2A) as seen in previous studies for pore formation by ESAT-6 in other cell and membrane types (11, 13). A membrane-disrupting effect of CFP-10 at a high concentration was confirmed by the presence of increased levels of LDH, a commonly used marker of cell necrosis, in neutrophil supernatants after prolonged incubation with CFP-10 at a high concentration (Fig. 2B).

FIG 2.

Effects of CFP-10 on plasma membrane integrity. Neutrophils were incubated at 37°C with or without CFP-10, and plasma membrane integrity was analyzed. (A) A Sytox green plate-based assay was used to assess neutrophil integrity upon the addition of CFP-10 at the indicated concentrations, buffer only, or 1% TX100 as a positive control for plasma membrane permeabilization. The diagram shows mean values of Sytox green fluorescence (expressed as percentages of TX100) after 5, 30, and 120 min of incubation from 3 to 4 independent experiments. Statistical analysis was performed using ordinary one-way ANOVA and Dunnett's post hoc test, comparing each treatment to the nontreated control. (B) The release of LDH from neutrophils incubated for 2 h with buffer or CFP-10 at 50 μg/ml was assessed, and the diagram shows the mean values of LDH release (expressed as percentages of TX100) from 5 independent experiments. Statistical analysis was performed using paired Student's t tests. (C) Neutrophils treated with or without CFP-10 at 50 μg/ml for 30 min were stained with DRAQ5 (which stains DNA in all cells) and Sytox green (which stains DNA only in permeabilized cells) and analyzed by imaging flow cytometry. The images show bright-field (BF), DRAQ5 (red), and Sytox green (green) images from 3 representative intact cells treated with buffer and from 3 permeabilized cells treated with CFP-10. The diagram shows mean percentages of permeabilized (Sytox green-positive) cells from 3 independent experiments. Statistical analysis was performed using paired Student's t tests. The scale bars represent 7 μm. (D) Fluo-3 and FuraRed fluorescence was analyzed in neutrophils during 150 s using flow cytometry in the presence or absence of the Ca²⁺ chelator EGTA (2.5 mM). Stimulus (10 μg/ml CFP-10) was added after 30 s (indicated by arrows). Normalized Fluo-3/FuraRed fluorescence ratios, representing relative Ca²⁺ levels, are shown, and traces are representative of the results of 3 independent experiments. AU, arbitrary units.

Importantly, CFP-10-induced membrane disruption did not cause cellular disintegration or rupture, as shown by incubation with CFP-10 at a high concentration and analysis by imaging flow cytometry. Upon incubation with CFP-10 for 30 min, a significant proportion of the cells were permeabilized (i.e., were positive for Sytox green), but they were not disintegrated (Fig. 2C). The results suggest pore formation rather than a detergent-like effect exerted by high concentrations of CFP-10.

Plasma membrane permeabilization causes influx of ions and other small molecules from the extracellular medium, resulting in increased cytosolic Ca²⁺ levels that were independent of receptor signaling. To rule out the possibility that the Ca²⁺ signal induced by CFP-10 was due to such unspecific leakage of Ca²⁺, the Ca²⁺ chelator EGTA was added to the cells prior to stimulation. At the concentration used (2.5 mM), EGTA chelates Ca²⁺ in the extracellular medium but does not affect intracellular Ca²⁺ levels (reference 31 and data not shown). The presence of EGTA did not affect the signal induced by CFP-10 (Fig. 2D). Thus, the cytosolic Ca²⁺ transient induced by CFP-10 was a true Ca²⁺ signal originating from an emptying of the intracellular storage organelles, a phenomenon typical for neutrophils activated by an agonist binding to a chemoattractant receptor (30, 31).

CFP-10 functions as a neutrophil chemoattractant and activates superoxide-producing NAPDH-oxidase.

Having found that CFP-10 activates neutrophils, we next investigated the functional outcomes of the interaction, focusing on neutrophil chemotaxis and production of ROS. CFP-10 significantly induced cell migration in a concentration-dependent manner, as evidenced by a plate-based transwell system where cells were allowed to migrate across a filter toward CFP-10 or fMLF as a positive control (Fig. 3A). These results indicate that CFP-10 is indeed chemotactic for human neutrophils.

FIG 3.

Neutrophil chemotaxis and NADPH-oxidase activation induced by CFP-10. (A) Neutrophils were allowed to migrate across a membrane into a well containing buffer as a negative control, fMLF (10⁻⁸ M) as a positive control, or CFP-10 at the indicated concentrations. After 90 min, chemotaxis was evaluated by lysing the cells in the well and measuring LDH in the lysates, corresponding to the number of cells present. The results are presented as mean percentages of cell migration (i.e., percentages of the cell control value, which was set at 100%). The dose-response curve shows median values from 1 representative experiment, and the bar graph shows the means of the results of 7 independent experiments. Statistical analysis was performed using paired Student's t tests. Images show wells from each treatment in one representative experiment. The scale bars represent 100 μm. (B) NADPH-oxidase activity was assessed in resting neutrophils or neutrophils pretreated with latrunculin A (100 ng/ml) for 10 min. The cells were equilibrated at 37°C in a luminometer before addition of stimuli (arrows) and recording of chemiluminescence over time (expressed in mega counts per minute [Mcpm]). The diagrams show extracellular O₂⁻ production over time in cells stimulated with 10 μg/ml CFP-10 or 10⁻⁷ M fMLF (positive control). Results from 1 experiment representative of 3 are shown.

Neutrophil activation by chemoattractants often results in activation of electron-transporting NADPH-oxidase and thus in the production and release of ROS. CFP-10 did not induce release of extracellular superoxide (O₂⁻) from unperturbed peripheral blood neutrophils (Fig. 3B). However, many agonists require priming of the cells through degranulation and increased surface expression of the corresponding receptor, or uncoupling from the actin cytoskeleton, to induce ROS production (32). Latrunculin A interferes with the polymerization of actin (32) and facilitates neutrophil secretion (33). When the cytoskeleton was disrupted through the addition of latrunculin A to the cells prior to stimulation, CFP-10 triggered substantial ROS production (Fig. 3B), and the response to fMLF (positive control) was also enhanced (Fig. 3B). Taken together, the functional outcomes of CFP-10-induced neutrophil activation include chemotaxis and ROS production, pointing toward a proinflammatory response of neutrophils to the mycobacterial protein.

CFP-10 activates cells through a pertussis toxin-sensitive G-protein-coupled receptor specific for neutrophils.

Having found indications that CFP-10 interacts with a neutrophil chemoattractant receptor, we next characterized the responsible receptor. The generation of Ca²⁺ signals in neutrophils may be achieved through activation of numerous different receptors, including members of the GPCR family of receptors which transmit their signals by activating heterotrimeric G-proteins, usually of the G_i/o family (34). These G-proteins have in common that they are sensitive to PtX, a Bordetella pertussis toxin which ADP-ribosylates the Gα subunit of the G-protein, blocking receptor interactions and thus preventing activation (35). In order to determine whether CFP-10 induces Ca²⁺ transients by binding to such a GPCR, neutrophils were treated with PtX before stimulation with CFP-10 and measurement of the Ca²⁺ response were performed. After PtX pretreatment, the CFP-10-triggered Ca²⁺ response in neutrophils was completely abolished (Fig. 4), demonstrating that CFP-10 utilizes a PtX-sensitive GPCR to induce this response. The PtX treatment also blocked the Ca²⁺ signal induced by fMLF, which acts through the PtX-sensitive GPCR FPR1, while the signal induced by the Ca²⁺ ionophore ionomycin was not affected (Fig. 4). The PtX-treated neutrophils were fully able to produce ROS in response to PMA stimulation (data not shown), affirming that the cells were viable and functional also after PtX treatment. Moreover, the NADPH-oxidase activation induced by CFP-10 in latrunculin A-treated cells was likewise sensitive to PtX (see Fig. S1 in the supplemental material) and was thus mediated through a PtX-sensitive GPCR.

FIG 4.

PtX sensitivity of the CFP-10-induced Ca²⁺ signal. Neutrophils were loaded with Fluo-3 and FuraRed and treated with PtX to inhibit G-proteins or were left untreated. The fluorescence ratio was analyzed continuously during 150 s using flow cytometry. (A) Background Fluo-3/FuraRed intensity was recorded for 30 s, after which 10 μg/ml CFP-10 (top panel), 10⁻⁸ M fMLF (middle panel), or 10⁻⁸ M ionomycin (bottom panel) was added (indicated by arrows). The diagrams show the normalized ratios between Fluo-3 and FuraRed fluorescence intensities, reflecting relative cytosolic Ca²⁺ levels over time. (B) The bar graph shows the mean peak normalized fluorescence ratios between Fluo-3 and FuraRed, reflecting relative peak cytosolic Ca²⁺ concentrations, after pretreatment and stimulation as indicated, from 4 independent experiments. A line is shown at the background value. Statistical analysis was performed using ordinary one-way ANOVA and Sidak's post hoc test, comparing the results of each stimulus with and without PtX treatment. AU, arbitrary units; Con, control.

To further characterize the unknown CFP-10 receptor, we tested several well-characterized antagonists/inhibitors for a range of neutrophil GPCRs, including FPR1, formyl peptide receptor 2 (FPR2), platelet activating factor (PAF) receptor, C5a receptor, P2Y₂ receptor, and CXCR2. However, none of the antagonists inhibited the CFP-10-induced Ca²⁺ signal (see Fig. S2 in the supplemental material), indicating that none of these GPCRs is the CFP-10 receptor. Instead, the cell specificity/selectivity of the response was investigated by stimulating a mixed leukocyte population with CFP-10. The Ca²⁺ measurement technique was combined with immunofluorescence staining of the surface marker CD45, differentially expressed on monocytes, lymphocytes, and neutrophils (36) (Fig. 5). As shown in Fig. 5, CFP-10 induced Ca²⁺ mobilization only in neutrophils, but not in lymphocytes or monocytes, demonstrating that the recognition of CFP-10 in this regard was neutrophil specific. Addition of the Ca²⁺ ionophore ionomycin to the cells gave rise to an intracellular rise in the level of Ca²⁺ in all three cell types, showing that loading with Ca²⁺ indicators was effective for all leukocyte subsets. As expected, stimulation with the FPR1 agonist fMLF yielded a Ca²⁺ signal in monocytes and neutrophils, both of which express FPR1 (37), but not in lymphocytes. The idea of neutrophil specificity of the CFP-10 response was further supported by experiments performed with the human promyelocytic leukemia (HL-60) cell line that can be in vitro differentiated into either neutrophil-like or monocyte-like cells (38). CFP-10 induced a Ca²⁺ response in neutrophil-differentiated cells but not in monocyte-differentiated HL-60 cells (see Fig. S3 in the supplemental material). Collectively, these data indicate that neutrophils are unique among leukocytes in their ability to recognize and respond to CFP-10 with a Ca²⁺ transient.

FIG 5.

Neutrophil specificity of the CFP-10 Ca²⁺ response. Mixed leukocyte populations were loaded with Fluo-3 and FuraRed and then stained using anti-CD45 antibody to allow discrimination of leukocyte subsets during Ca²⁺ measurements by flow cytometry. The ratio of the fluorescence intensities from the two Ca²⁺ indicator dyes was recorded continuously during 150 s in lymphocytes, monocytes, and neutrophils. Background Fluo-3/FuraRed intensity was recorded for 30 s, after which stimuli were added (indicated by arrows). The stimuli were 10 μg/ml CFP-10, 10⁻⁷ M fMLF, and 10⁻⁸ M ionomycin, as indicated. The dot plot shows CD45 intensity versus side scatter and the gates used for the different cell types. The diagrams show the normalized ratios between Fluo-3 and FuraRed fluorescence intensities, reflecting relative cytosolic Ca²⁺ levels over time in the indicated cell populations, and data are representative of the results of 3 independent experiments. AU, arbitrary units.

DISCUSSION

Surprisingly little is known about how M. tuberculosis interacts with host neutrophils, even though these cells are recruited in great numbers to the lung upon initial infection (4) and are present in tuberculous granulomas (39). However, a recent publication showed that an intact RD1 genomic region is required for M. tuberculosis to be able to evade neutrophil defense mechanisms and survive within neutrophils (21). We set out to investigate the direct interaction between primary human neutrophils and ESAT-6:CFP-10, a heterodimeric protein complex formed spontaneously by the cotranscribed RD1-encoded proteins ESAT-6 and CFP-10 (5–9). Several cytolytic and immune evasion functions of ESAT-6 have previously been revealed (12–16, 19), while CFP-10 has been described to act as a chaperone protein for ESAT-6, required for its secretion (10), but without known biological effects of its own (11, 19, 20). The data presented herein show that neutrophils specifically recognize CFP-10, responding by cellular activation, and thereby reveal that the CFP-10 component of the ESAT-6:CFP-10 complex possesses direct biological activity.

The ESAT-6:CFP-10 complex as well as CFP-10 on its own, but not ESAT-6 alone, induced an intracellular Ca²⁺ transient in neutrophils. This shows that CFP-10 rather than ESAT-6 was responsible for the neutrophil Ca²⁺ response to ESAT-6:CFP-10. Endotoxin contamination was evidently not the cause of this neutrophil activation, since lipopolysaccharide stimulation does not trigger Ca²⁺ signals even at very high concentrations (data not shown). ESAT-6:CFP-10 has previously been shown to bind to the surface of U937 monocyte-like cells through the long flexible arm formed by the C terminus of CFP-10, and the structure of CFP-10 has been stated to be consistent with a receptor-interacting function (9), in accordance with our findings. A few earlier studies have indicated a possible inhibitory effect of CFP-10 on dendritic cell maturation (40) and downregulation of inflammatory macrophage responses (41, 42). Several other studies, however, have found a lack of direct effect of CFP-10 on different cell types, while ESAT-6 has been shown to have several effects. These include interactions with TLR2 on macrophages, impairing signaling (19), and the induction of IL-8 production in epithelial cells through an unknown receptor (20). None of those studies included the use of neutrophils.

As a functional ESX-1 secretion system is required for inducing macrophage lysis and M. tuberculosis spread to adjacent cells (12, 13, 15, 16), we tested whether CFP-10 could also induce plasma membrane permeabilization in neutrophils. CFP-10 at high concentrations permeabilized neutrophils, but the cells appeared relatively intact by imaging flow cytometry after 30 min of incubation, pointing toward pore formation in the neutrophil plasma membrane. It has previously been shown that recombinant ESAT-6 is able to form pores in erythrocytes and macrophages, as well as to destabilize liposomes, while recombinant CFP-10 did not display this effect in these cell types (11, 13). However, neutrophils have not previously been investigated.

CFP-10 permeabilized neutrophils in a manner resembling the effect of the phenol-soluble modulin (PSMα) peptides produced by Staphylococcus aureus, substances also capable of activating neutrophils through FPR2 at lower concentrations (43). Clearly, FPR2 was not responsible for the activating effects of CFP-10 (see Fig. S2 in the supplemental material), and the cytolytic effects of CFP-10 as well as the PSMα peptides were insensitive to PtX (data not shown), indicating that the activating and cytolytic properties of these bacterial factors are mediated by different mechanisms. Further, the PSMα peptides selectively permeabilize apoptotic over viable neutrophils (43), but this was not the case for CFP-10 (data not shown).

To characterize the neutrophil response to CFP-10, the functional outcome of the interaction was investigated, and it was established that neutrophils responded to CFP-10 with both chemotaxis and NADPH-oxidase activation. We speculate that the chemotactic property of CFP-10 may play a role during early infection with M. tuberculosis—serving to attract neutrophils to the site of infection. Although several M. tuberculosis factors have been shown to activate neutrophils (44, 45), this is to our knowledge the first report of an M. tuberculosis protein that exhibits direct chemotactic effects on neutrophils. A variety of factors are known to indirectly attract neutrophils; e.g., ESAT-6 has been found to induce production of the neutrophil chemoattractant IL-8 in epithelial cells (20). Similarly to all prokaryotes, however, M. tuberculosis initiates protein synthesis with formylated methionine residues (46) and thus N-formylated peptides released from these bacteria would be suspected to attract neutrophils directly through the action of FPRs.

Chemotaxis was induced by CFP-10 in unperturbed neutrophils, while priming with latrunculin A, which disrupts filamentous actin, was required for NADPH-oxidase activation. The actin cytoskeleton is known to be central for termination of GPCR signaling, and when actin polymerization is inhibited, GPCR signaling is typically prolonged as well as enhanced (47). The results showing weak stimulatory effect of the NADPH-oxidase in resting cells exerted by CFP-10 but a potent effect on cells with a disrupted cytoskeleton are similar to those showing the neutrophil response seen in the presence of several other neutrophil chemoattractant GPCRs (32, 48). The CFP-10 receptor differs from the FPRs in that NADPH-oxidase activation through the receptor requires prior disruption of the actin cytoskeleton, an observation that is likely explained by the presence of different signaling pathways leading from receptor binding to activation of the NADPH-oxidase, in addition to differences in receptor coupling to the actin cytoskeleton (29, 32, 49).

To pinpoint the CFP-10 receptor, PtX sensitivity was studied, and the results demonstrated that PtX abolished both the CFP-10-induced Ca²⁺ response and NADPH-oxidase activation. PtX is a specific inhibitor of the G_i/o heterotrimeric G-proteins linked to the neutrophil chemoattractant receptors (35), and our results thus strongly point toward the involvement of a classical chemoattractant GPCR in mediating these responses. Several well-known chemotactic neutrophil GPCRs (FPR1, FPR2, platelet-activating factor [PAF] receptor, C5a receptor, P2Y₂, and CXCR2) were tested and can be excluded from the list of potential receptors. Further, pretreatment with ESAT-6 before stimulation with CFP-10 in the NADPH-oxidase activity assay did not affect the CFP-10 response (data not shown), indicating that ESAT-6 does not interact with the same receptor on neutrophils. The identity of the CFP-10 GPCR remains unknown. It is possible that several different receptors are engaged by the protein, mediating different responses.

Finally, experiments using mixed leukocyte preparations demonstrated that neutrophils were unique among leukocytes in their ability to recognize and respond to CFP-10 with a Ca²⁺ transient, as monocytes and lymphocytes failed to respond. This indicates that the CFP-10 receptor is either present or functional only on neutrophils and not on the other cell types. Alternatively, monocytes and/or lymphocytes do recognize CFP-10, but the interaction does not result in the release of Ca²⁺ from intracellular stores. The idea that neutrophils harbor a chemotactic GPCR that is not expressed (or is not functional) on monocytes, which are also of myeloid lineage, is intriguing, and, to our knowledge, no such GPCR has previously been described. On the other hand, GPCR formyl peptide receptor 3 (FPR3) is expressed in monocytes but not in neutrophils (50), making it conceivable that such a receptor exists. Also, experiments using monocyte- and neutrophil-differentiated HL-60 cells, where only the latter responded to CFP-10, support our findings. Our data demonstrate that CFP-10 can signal through a neutrophil-specific chemotactic GPCR of as-yet-unknown identity, and we speculate that recognition of CFP-10 through this receptor could be of benefit for the host in terms of rapid recruitment of neutrophils that can eradicate M. tuberculosis infection but also possibly for the bacterium which then gains access to potential host cells.

In conclusion, this study demonstrated a novel direct biological role for the CFP-10 component of ESAT-6:CFP-10 in specifically attracting and activating neutrophils. Thus, there is apparently more to CFP-10 than its previously described role as a chaperone protein for ESAT-6. We speculate that neutrophils may encounter ESAT-6:CFP-10 during initial M. tuberculosis infection, prior to phagocytosis or upon lysis of an M. tuberculosis-infected macrophage. This report adds to the growing body of data indicating an important role of the ESAT-6:CFP-10 complex in the M. tuberculosis-neutrophil interplay.