Involvement of Caspase-9 in the Inhibition of Necrosis of RAW 264 Cells Infected with Mycobacterium tuberculosis

Abstract

In order to know how caspases contribute to the intracellular fate of Mycobacterium tuberculosis and host cell death in the infected macrophages, we examined the effect of benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethane (z-VAD-fmk), a broad-spectrum caspase inhibitor, on the growth of M. tuberculosis H37Rv in RAW 264 cells. In the cells treated with z-VAD-fmk, activation of caspase-8, caspase-3/7, and caspase-9 was clearly suppressed, and DNA fragmentation of the infected cells was also reduced. Under this experimental condition, it was found that the treatment markedly inhibited bacterial growth inside macrophages. The infected cells appeared to undergo cell death of the necrosis type in the presence of z-VAD-fmk. We further found that z-VAD-fmk treatment resulted in the generation of intracellular reactive oxygen species (ROS) in the infected cells. By addition of a scavenger of ROS, the host cell necrosis was inhibited and the intracellular growth of H37Rv was significantly restored. Among inhibitors specific for each caspase, only the caspase-9-specific inhibitor enhanced the generation of ROS and induced necrosis of the infected cells. Furthermore, we found that severe necrosis was induced by infection with H37Rv but not H37Ra in the presence of z-VAD-fmk. Caspase-9 activation was also detected in H37Rv-infected cells, but H37Ra never induced such caspase-9 activation. These results indicated that caspase-9, which was activated by infection with virulent M. tuberculosis, contributed to the inhibition of necrosis of the infected host cells, presumably through suppression of intracellular ROS generation.

Tuberculosis caused by Mycobacterium tuberculosis is still a serious threat to human health at the global level. It has been estimated that one-third of the world's population are infected, and 8 million people develop active tuberculosis every year (15, 29). A number of studies have been carried out to identify the pathogenic determinants of M. tuberculosis, and various candidate molecules that may contribute to mycobacterial virulence have been reported (7). However, the molecular mechanisms for the virulence still remain unclear.

Macrophages play a role in the first line of host defense against bacterial infection by exerting microbicidal activity and contribute to the development of protective T cells as antigen-presenting cells through production of cytokines, including interleukin-12 (IL-12) and IL-18 (26). However, M. tuberculosis is capable of modulating such host response and survives inside macrophages (15). Therefore, some type of host response in the infected cell itself is necessary to control the replication of M. tuberculosis in the initial phase of infection. There are several reports indicating that induction of early death of infected cells is an important and alternative strategy for host defense against M. tuberculosis. For instance, it has been shown that macrophages go into apoptosis upon infection with M. tuberculosis in a caspase-dependent manner, resulting in the suppression of intracellular bacterial replication, and that arrest of macrophage apoptosis conversely enhances bacterial growth (22, 28). Furthermore, it has been reported that the apoptotic vesicles formed in the infected macrophages have an important role in transporting the mycobacterial antigen to dendritic cells and developing cellular immunity against M. tuberculosis (25). These results suggest that apoptosis of the infected cells constitutes an important part of the host resistance and affects the fate of intracellular M. tuberculosis. To date, the intracellular cascade of apoptosis has been characterized well and various caspases are known to be involved in apoptosis induction (21).

Caspases are synthesized as biologically inactive precursors and converted into active forms by sequential proteolytic cleavage. The activation process is regulated by various intracellular components and is under strict control. Upon apoptosis, which is a form of innate immunity against bacteria, however, it appears that M. tuberculosis exerts resistance by modification of the activation cascade of caspases in the cells where it resides. Sly et al. have recently reported that virulent M. tuberculosis strains cause less apoptosis than attenuated strains by induction of macrophage antiapoptotic mcl-1 gene expression (28). Balcewicz-Sablinska et al. have also shown that M. tuberculosis H37Rv inhibits apoptosis of infected macrophages by IL-10-dependent release of a soluble tumor necrosis factor (TNF) receptor that inactivates TNF-α (2). These findings suggest that though apoptosis is coupled with killing of intracellular M. tuberculosis, the bacterium possesses a virulence-associated ability to evade apoptosis.

In addition to apoptosis, it has been shown that M. tuberculosis triggers necrosis of infected macrophages. Unlike apoptosis, it appears that necrosis does not interfere with the survival of intracellular M. tuberculosis. Moreover, it is supposed that M. tuberculosis ultimately escapes macrophages by inducing necrosis, and necrotic cell death provides the nutrient source for M. tuberculosis in granuloma (30). Park et al. have shown that virulent clinical strains rapidly grow inside macrophages and induce necrosis of infected macrophages (20). Hsu et al. have demonstrated that an attenuated mutant of M. tuberculosis H37Rv failed to induce necrosis of infected macrophages (14). These results suggest that virulence of M. tuberculosis is associated with the ability to manipulate not only apoptosis but also necrosis of infected macrophages. However, little is known about the regulatory mechanism of apoptosis and necrosis or the relationship between M. tuberculosis-induced caspase activation and the fate of intracellular bacteria.

In this study, we employed various caspase inhibitors and examined their effects on the intracellular growth of a virulent H37Rv strain in macrophage-like RAW 264 cells. Unexpectedly, it was found that inhibition of caspases resulted in the necrosis of H37Rv-infected cells and our analysis revealed that the activation of caspase-9 is involved critically in the inhibition of necrosis. Furthermore, we found that H37Ra did not induce either necrosis of infected cells or activation of caspase-9. It was suggested that virulent M. tuberculosis strains avoid excessive necrosis of infected host cells by inducing caspase-9 activation.

MATERIALS AND METHODS

Reagents.

Benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethane (z-VAD-fmk; an inhibitor of various caspases) and acetyl-Tyr-Val-Ala-Asp-chloromethane (a caspase-1 inhibitor) were purchased from Peptide Institute, Inc. (Osaka, Japan). Other inhibitors, including benzyloxycarbonyl-Val-Asp(OMe)-Val-Ala-Asp(OMe)-fluoromethane (a caspase-2 inhibitor), benzyloxycarbonyl-Asp(OMe)-Gln-Met-Asp(OMe)-fluoromethane (a caspase-3 inhibitor), benzyloxycarbonyl-Ile-Glu(OMe)-Thr-Asp(OMe)-fluoromethane (a caspase-8 inhibitor), benzyloxycarbonyl-Leu-Glu(OMe)-His-Asp(OMe)-fluoromethane (a caspase-9 inhibitor), and benzyloxycarbonyl-Phe-Ala-fluoromethylketone (z-FA-fmk; an inactive caspase inhibitor analogue), were purchased from Sigma Aldrich (St. Louis, MO), Merck Biosciences, Inc. (San Diego, CA), Techne Corporation (Minneapolis, MN), R & D Systems, Inc. (Minneapolis. MN), and Calbiochem (San Diego, CA), respectively. 3(2)-t-Butyl-4-hydroxyanisole (BHA) and 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) were obtained from Wako Pure Chemical Industries (Osaka, Japan) and Molecular Probes (Eugene, OR), respectively. Rabbit anti-mouse caspase-9 antibody was obtained from Cell Signaling Technology, Inc. (Danvers, MA).

Bacteria.

The M. tuberculosis H37Rv and H37Ra strains maintained in our laboratory were grown at 37°C to mid-log phase in Middlebrook 7H9 broth (Becton Dickinson Microbiology Systems, Sparks, MD) supplemented with 0.5% albumin, 0.2% dextrose, 3 μg/ml catalase, and 0.2% glycerol. H37Rv was harvested and stirred vigorously with glass beads to disperse the bacterial clumps and stood for 30 min. An upper part of the suspension without visible clumps was collected and stored at −80°C in aliquots. After being thawed, the bacterial suspension was centrifuged at 150 × g for 3 min to remove clumps, and only the upper part of the suspension was used for the experiments to ensure an even infection of each cell. Viable bacteria were enumerated by plating the diluted suspension on Middlebrook 7H10 agar plates containing 50 μg/ml oleic acid, 0.5% albumin, 0.2% dextrose, 4 μg/ml catalase, and 0.85 mg/ml sodium chloride and counting the number of colonies 3 weeks after incubation at 37°C.

Measurement of intracellular bacterial growth.

RAW 264 cells were seeded in 24-well microplates at 1.0 × 10⁵ cells/well and incubated for 12 h at 37°C in 5% CO₂ in a culture medium consisting of RPMI 1640 medium supplemented with 10% fetal bovine serum and 5 μg/ml of gentamicin. Cells were washed and infected with 5 × 10⁵ CFU of H37Rv for 4 h. After three washes with the culture medium for removal of extracellular bacteria, the cells were cultured for 7 days in the presence or absence of various caspase inhibitors and/or BHA. Cells were lysed in 0.05% Triton X-100 solution, and the number of viable bacteria in each well was determined by plating the lysate on Middlebrook 7H10 agar plates. In one experiment, thioglycolate-induced peritoneal macrophages (1.0 × 10⁵ cells) were infected with H37Rv and the intracellular bacterial number was determined 7 days later.

Detection of DNA fragmentation.

Two days after infection at a multiplicity of infection (MOI) of 5, 5 × 10⁶ cells were lysed in a lysis buffer consisting of 10 mM Tris-HCl (pH 7.6), 0.15 M NaCl, 5 mM MgCl₂, and 0.5% Triton X-100. Intact nuclei were collected by centrifugation at 1,000 × g for 5 min, suspended in 10 mM Tris-HCl (pH 7.6) buffer containing 0.4 M NaCl, 1 mM EDTA, and 1% Triton X-100, and centrifuged at 12,000 × g for 15 min to segregate the nucleoplasm from high-molecular-weight chromatin. The semipurified nucleoplasm was consecutively incubated at 37°C with 20 μg/ml of RNase for 1 h and 100 μg/ml of proteinase K for 2 h. DNA was extracted with the phenol-chloroform method and electrophoresed on a 1.4% agarose gel. After being stained with ethidium bromide, DNA was visualized on a UV transilluminator. Alternatively, oligonucleosomes were quantified by using a Cell Death Detection ELISA^PLUS kit (Roche Diagnostics, Penzberg, Germany) according to the manufacturer's protocol. The degree of DNA fragmentation was expressed as an arbitrary unit calculated by the following formula: arbitrary unit = (A₄₀₅ of experimental group − A₄₀₅ of negative control [medium only])/(A₄₀₅ of untreated cells − A₄₀₅ of negative control).

Flow cytometric analysis.

RAW 264 cells were collected 2 and 4 days after infection and washed with phosphate-buffered saline (PBS) containing 0.2% albumin. Cells were incubated with 0.2 mM propidium iodide (PI; Molecular Probes, Eugene, OR) for 10 min on ice in the dark, washed, and fixed with 1% paraformaldehyde in PBS. Fluorescence intensity was analyzed by FACSCalibur (BD Biosciences, San Jose, CA). In order to detect intracellular reactive oxygen species (ROS), RAW 264 cells infected with H37Rv 2 days before were incubated with 5 μM DCFH-DA for 15 min at 37°C. DCFH-DA diffused into cells and was hydrolyzed to DCFH (2′, 7′-dichlorohydrofluorescein). Cells were detached from culture plates, and the fluorescence intensity of DCFH, which was converted into oxidized form by intracellular ROS, was analyzed by FACSCalibur according to a method described previously (3).

Detection of LDH.

RAW 264 cells and peritoneal macrophages were infected with H37Rv or H37Ra, and the culture supernatants were collected 2 and 4 days later. The amount of lactate dehydrogenase (LDH) released from the infected cells was measured using an LDH cytotoxicity detection kit (TaKaRa BIO Inc., Shiga, Japan). The percentage of LDH release was calculated according to the following formula: percent release = 100 × (experimental LDH release − spontaneous LDH release)/(maximal LDH release − spontaneous LDH release). A value for maximal LDH release was obtained from the supernatant of cells treated with 1% Triton X-100.

Transmission electron microscopy.

RAW 264 cells were infected with H37Rv for 3 days. The cells were washed twice with PBS and once with 0.1 M cacodylic acid buffer and fixed with 2.5% glutaraldehyde in 0.1 M cacodylic acid buffer. After fixation, the cells were treated with 2% osmium tetroxide in 0.1 M cacodylic acid buffer, dehydrated by treatment with graded ethanol solutions, and embedded in Quetol-812 resin mixture-embedding media. The ultrathin sections were stained with uranyl acetate and lead citrate and examined with a JEOL model JEM-1200EX electron microscope. The percentage of cells undergoing apoptosis or necrosis was estimated by investigating the morphologies of 100 cells in each experimental group.

Measurement of caspase activities.

RAW 264 cells were lysed 1 and 2 days after infection, and the caspase-8, caspase-3 and/or -7, and caspase-9 activities in the cleared lysate were measured by using Caspase-Glo 8, Caspase-Glo 3/7, and Caspase-Glo 9 assays (Promega Corporation, Madison, WI) according to the manufacturer's protocols.

Statistical analysis.

Student's t test was used to determine the statistical significance of the values obtained, and P values of <0.05 were considered statistically significant.

RESULTS

Effect of z-VAD-fmk on intracellular growth of H37Rv.

In order to determine the effect of inhibition of caspase activities on the intracellular growth of H37Rv in RAW 264 cells, we first infected the cells with H37Rv at an MOI of 5 in the presence or absence of z-VAD-fmk and monitored the number of intracellular bacteria. H37Rv replicated slowly in RAW 264 cells in the absence of z-VAD-fmk (Fig. 1). A similar pattern of growth was observed in z-VAD-fmk-treated macrophages for 2 days after infection. However, the growth rate was suppressed afterwards. To rule out the possibility that the growth inhibition is due to the direct action of z-VAD-fmk on H37Rv, we incubated H37Rv (5 × 10⁵ CFU) in Middlebrook 7H9 broth including albumin, dextrose, and catalase for 7 days in the presence or absence of z-VAD-fmk and recovered 1.52 × 10⁷ CFU or 1.56 × 10⁷ CFU of H37Rv from the culture, respectively. A similar result was observed when H37Rv was cultured in the cell culture medium (data not shown). The results showed that z-VAD-fmk by itself did not influence bacterial growth and raised the possibility that some initiator and/or effector caspases were involved in facilitating the intracellular survival of H37Rv. Several studies clearly showed that caspase-8 and caspase-9 (initiator caspases) and caspase-3 (an effector caspase) are activated after M. tuberculosis infection and cause cell death in infected macrophages (22). Thus, we speculated that these caspases might have some activity influencing the fate of intracellular bacteria. To make this point clear, we examined whether these caspases were activated after infection with H37Rv and whether z-VAD-fmk inhibited the activation. As shown in Fig. 2, a significant level of caspase-8 activation was observed 1 day after infection and the activity decreased back to the control level by 2 days. The activities of caspase-3/7 and caspase-9 were increased on day 2 about four- and twofold, respectively. These caspase activities were at the control level on day 5, indicating that H37Rv induced activation of caspases to some extent. On the other hand, H37Rv-induced activations of these caspases were mostly inhibited in the presence of z-VAD-fmk and were not affected by z-FA-fmk (an inactive caspase inhibitor analogue). Since growth inhibition of the intracellular bacteria was detected in z-VAD-fmk-treated cells later than 2 days after infection, the results suggest that caspases may play a role in the intracellular survival of virulent M. tuberculosis.

FIG. 1.

Kinetics of intracellular growth of H37Rv in RAW 264 cells treated with or without z-VAD-fmk. RAW 264 cells were infected with H37Rv at an MOI of 5 and cultured in the presence or absence of 40 μM z-VAD-fmk. Cells were lysed at the indicated days, and the number of viable bacteria was determined by a CFU assay. Data represent the means ± standard deviations for triplicate assays and are representative of three independent experiments. *, P < 0.05.

FIG. 2.

Caspase activities after infection with H37Rv. RAW 264 cells were infected with H37Rv, and caspase activities were measured at 1, 2, and 5 days after infection by a Caspase-Glo assay. VAD, z-VAD-fmk; FA, z-FA-fmk; RLU, relative light units. Data represent the means ± standard deviations for triplicate assays and are representative of three independent experiments. *, P < 0.05.

z-VAD-fmk treatment causes necrosis of infected RAW 264 cells.

To find the reason z-VAD-fmk treatment inhibited bacterial growth, on day 3 of infection, infected cells in the presence or absence of z-VAD-fmk were examined under an electron microscope. We did not detect any morphological change between normal cells (Fig. 3A) and cells treated only with z-VAD-fmk for 3 days (Fig. 3B). However, we found that infection of RAW 264 cells with H37Rv influenced cell morphology. Among the infected cells, 42% of the cells maintained their cellular structures (Fig. 3C and D) and 26% showed apoptotic structural changes (Fig. 3E). The remaining 32% of the cells displayed morphological changes characteristic of necrosis (Fig. 3F). The treatment with z-VAD-fmk provoked much severe damage in the infected cells. As shown in Fig. 3G and H, the caspase inhibitor caused necrotic morphological changes in as many as 95% of the infected cells. These results raised the possibility that some caspases contribute to inhibition of necrosis of infected cells in conventional infection in vitro. To confirm that the morphological changes resulted from apoptosis or necrosis of the infected cells, we investigated a generation of oligonucleosomes and assessed the membrane integrity of the infected cells by measuring the population of PI-stained cells and LDH release into the culture medium. As shown in Fig. 4A, an oligonucleosomal DNA ladder was observed in RAW 264 cells infected with H37Rv but not in uninfected cells. Treatment with z-VAD-fmk clearly inhibited the cleavage of DNA. To analyze the DNA fragmentation quantitatively, the cell lysates were subjected to a sandwich enzyme-linked immunosorbent assay specific for oligonucleosomes. The number of oligonucleosomes in H37Rv-infected cells was about 5 times as high as that in the uninfected cells, but z-VAD-fmk significantly inhibited the generation of oligonucleosomes (Fig. 4B). These results clearly indicated that z-VAD-fmk inhibited the apoptotic process induced by infection with H37Rv. On the other hand, though the infected cells were not stained with PI for 2 days after infection, irrespective of z-VAD-fmk treatment, the fluorescence intensities in both groups increased at day 4 postinfection (Fig. 4C and D). In the absence of z-VAD-fmk treatment, cells were stained intermediately with PI and could be divided into two populations based on fluorescence intensity. Since the number of cells expressing the brighter fluorescence reached 32.2% of the total number of cells and 32% of the infected cells exhibited necrotic morphologies (Fig. 3), it seemed that the cells stained highly with PI represented the cells undergoing necrosis. On the other hand, most of the z-VAD-fmk-treated cells were stained strongly with PI 4 days after infection, while z-FA-fmk treatment did not increase the number of damaged cells in the population (Fig. 4E). PI-stained cells were never observed in noninfected cells treated with z-VAD-fmk alone (Fig. 4F). In addition, z-VAD-fmk treatment caused a lower but a significant level of LDH release from the infected cells 2 days after infection, and the release was dramatically enhanced on day 4. Again, there was no LDH release at all from the cells treated with z-VAD-fmk alone even after 4 days in the absence of M. tuberculosis infection (Fig. 4G). Furthermore, z-FA-fmk treatment did not induce LDH release from the infected cells. These results strongly suggested that H37Rv induced the activation of caspases, resulting in not only the induction of apoptosis but also the inhibition of necrosis of infected cells.

FIG. 3.

Effect of z-VAD-fmk on morphologies of infected cells. RAW 264 cells infected with H37Rv were cultured for 3 days in the absence (C to F) or presence (G, H) of z-VAD-fmk. Cells were fixed, and morphologies were observed under an electron microscope. (E) A representative apoptotic cell. (F) A cell showing necrotic damage. (C and D) Infected cells which maintained normal cell morphologies. (A and B) Nontreated cells (A) and cells treated only with z-VAD-fmk (B). Micrographs were taken at magnifications of ×5,000 (A to F and H) and ×3,000 (G).

FIG. 4.

Inhibition of apoptosis and induction of necrosis by z-VAD-fmk treatment. RAW 264 cells were infected with H37Rv and cultured for 2 days. Cells were lysed, and the DNA ladder and oligonucleosomes were detected by agarose gel electrophoresis (A) and quantified by an enzyme-linked immunosorbent assay (B), respectively. Two days (C) and 4 days (D, E) after infection, the cells were stained with PI and fluorescence intensity was measured. VAD, z-VAD-fmk; FA, z-FA-fmk. (F) Fluorescence intensity of the cells treated only with z-VAD-fmk for 4 days. (G) Culture supernatants were collected 2 and 4 days after infection, and LDH activity was assayed. Data represent the means ± standard deviations for triplicate assays and are representative of three independent experiments. *, P < 0.05.

Involvement of ROS accumulation in z-VAD-fmk-induced necrosis of infected cells.

Vercammen et al. have shown that L929 fibrosarcoma cells treated with z-VAD-fmk rapidly died from necrosis after treatment with TNF-α. Necrotic cell death was induced also by overexpression of cytokine response modifier A (CrmA), a serpin-like caspase inhibitor (12, 32). In these reports, they indicated that the necrosis was provoked by ROS that was generated by inhibition of caspases. Since H37Rv infection induced severe necrosis of infected cells when the cells were treated with a caspase inhibitor, we examined whether z-VAD-fmk treatment triggers ROS generation in H37Rv-infected cells by using DCFH-DA, a fluorescent detector of ROS. RAW 264 cells were infected with H37Rv for 2 days in the presence or absence of z-VAD-fmk and treated with DCFH-DA for 15 min. The fluorescence of DCFH emitted in cytoplasm after oxidization by ROS was measured by FACSCalibur. The fluorescence intensity of the infected cells was increased significantly by treatment with z-VAD-fmk (Fig. 5A). However, enhancement of fluorescence was diminished by addition of BHA, a scavenger of ROS generated intracellularly, indicating that z-VAD-fmk treatment induced the generation of ROS in the infected cells (Fig. 5B). We further found that BHA treatment suppressed the z-VAD-fmk-induced necrosis of H37Rv-infected cells, because both the fluorescence intensity of the PI-stained cells and the LDH release from the cells were decreased markedly by BHA (Fig. 5C and D). In addition, though treatment with z-VAD-fmk inhibited the intracellular growth of H37Rv, the inhibitory activity was cancelled appreciably by treatment with BHA (Fig. 5E). Furthermore, our data showed that the low level of intracellular ROS that was generated by H37Rv alone did not affect the bacterial growth, because BHA treatment did not influence the intracellular bacterial number. These results indicated that inhibition of caspase activities by z-VAD-fmk induced the high level of ROS generation in the cytoplasm of H37Rv-infected cells. The intracellular ROS appeared to contribute to the induction of necrosis and the arrest of intracellular growth of H37Rv.

FIG. 5.

Involvement of intracellular ROS accumulation in induction of necrosis and inhibition of intracellular bacterial growth. (A) RAW 264 cells were infected with H37Rv and cultured for 2 days in the presence or absence of z-VAD-fmk. Cells were treated with 5 μM DCFH-DA, and the fluorescence intensity of the oxidized DCFH was measured. (B) RAW 264 cells were infected and cultured for 2 days with or without z-VAD-fmk and an antioxidant reagent, BHA (25 μM). The cells were treated with DCFH-DA, and fluorescence intensity was measured. (C) Two days after infection, the cells were collected and cell permeability to PI was measured. (D) RAW 264 cells were infected and cultured for 4 days. The culture supernatant was collected, and the LDH activity was measured. (E) RAW 264 cells were infected and cultured for 7 days with or without z-VAD-fmk and BHA. The cells were lysed, and the number of intracellular bacteria was determined. Data represent the means ± standard deviations for triplicate assays and are representative of three independent experiments. *, P < 0.05.

Critical involvement of caspase-9 in the inhibition of necrosis.

Because z-VAD-fmk is a broad-spectrum caspase inhibitor capable of inhibiting the activity of caspases in general (23, 27), we next tried to identify the particular caspase involved in the inhibition of necrosis of infected cells. To address this point, we employed specific inhibitors for caspase-1, caspase-2, caspase-3, caspase-8, and caspase-9 (10, 17, 31) instead of z-VAD-fmk and determined the effect of each inhibitor on necrosis of H37Rv-infected cells by measuring membrane permeability. Although inhibitors of caspase-1, caspase-2, caspase-3, and caspase-8 did not change the PI-staining pattern of the infected cells (Fig. 6A to D), only the caspase-9 inhibitor (z-LEHD-fmk) enhanced the fluorescence intensity (Fig. 6E). The inhibitor itself did not affect the membrane permeability (Fig. 6F). Furthermore, we found that z-LEHD-fmk enhanced the fluorescence intensity of oxidized DCFH in the infected cells (Fig. 6G) and inhibited the intracellular growth of H37Rv (Fig. 6I). Activation of caspase-9 in H37Rv-infected cells was significantly abolished in the presence of z-LEHD-fmk (Fig. 6H). These results suggested that caspase-9 which was activated by H37Rv infection contributed to the inhibition of necrosis through regulation of intracellular ROS generation.

FIG. 6.

Effects of inhibitors (120 μM) for caspase-1, -2, -3, -8, and -9 on induction of necrosis and intracellular bacterial growth. RAW 264 cells were infected with H37Rv and incubated with or without inhibitors specific for caspase-1 (A), caspase-2 (B), caspase-3 (C), caspase-8 (D), or caspase-9 (E). After 3 days of cultivation, membrane permeability was assessed by PI staining. (F) The caspase-9 inhibitor does not affect membrane permeability. (G) RAW 264 cells were infected and cultured for 2 days with or without z-LEHD-fmk, and the cells were stained with DCFH-DA and measured for fluorescence derived from oxidized DCFH. (H) RAW 264 cells were infected and cultured for 2 days with or without z-LEHD-fmk. Caspase-9 activity was measured by a Caspase-Glo assay. (I) Seven days after infection and cultivation with z-LEHD-fmk, the number of intracellular bacteria was determined by a CFU assay. Data represent the means ± standard deviations for triplicate assays and are representative of three independent experiments. *, P < 0.05. RLU, relative light units.

Detection of caspase-9-dependent inhibition of bacterial growth and necrosis in peritoneal macrophages.

To analyze whether caspase-9-dependent inhibition of necrosis is a general intracellular event after H37Rv infection, we infected thioglycolate-induced peritoneal macrophages with H37Rv in the presence or absence of z-VAD-fmk or z-LEHD-fmk and measured LDH release. Similar to H37Rv infection in RAW 264 cells, H37Rv induced only a low level of LDH release from the infected macrophages and z-FA-fmk did not augment the response (Fig. 7A). However, treatment with z-VAD-fmk and z-LEHD-fmk caused a high level of LDH release from the infected macrophages compared to treatment with caspase inhibitor alone or H37Rv infection without treatment with these inhibitors (Fig. 7A). In addition, we found that treatment with z-VAD-fmk or z-LEHD-fmk significantly reduced the intracellular number of bacteria (Fig. 7B). These data are consistent with those obtained from RAW 264 cells, suggesting that caspase-9 generally contributes to inhibition of necrosis in macrophages after M. tuberculosis infection.

FIG. 7.

Effects of caspase inhibitors on necrosis of H37Rv-infected macrophages and intracellular growth of bacteria. Peritoneal exudate macrophages were infected with H37Rv at an MOI of 5 in the presence or absence of caspase inhibitors. (A) Culture supernatants were collected 1 day after infection, and LDH activity was assayed. (B) Cells were infected with H37Rv and incubated for 7 days in the presence or absence of caspase inhibitors. The number of intracellular bacteria was determined by a CFU assay. Data represent the means ± standard deviations for triplicate assays and are representative of three independent experiments. *, P < 0.05.

Difference between the abilities of H37Rv and H37Ra to induce necrosis and apoptosis and activate caspase-9.

We next investigated the abilities of H37Rv and H37Ra in induction of apoptosis and necrosis and activation of caspase in RAW 264 cells to determine whether H37Rv-induced responses are associated with the virulence of M. tuberculosis. Two days after infection, H37Rv induced LDH release from the infected cells while such a level of LDH release was not detected in the supernatant of cells infected with H37Ra (Fig. 8A). In contrast, a significant number of oligonucleosomes was detected in the lysate of H37Ra-infected cells, which was higher than that in the H37Rv-infected cells (Fig. 8B). In addition, we found that H37Ra induced a higher level of caspase-3/7 and caspase-8 activation than H37Rv. These results indicated that H37Rv but not H37Ra caused necrosis of RAW 264 cells and that H37Ra had a higher ability in induction of apoptosis than H37Rv. In this experimental model, as described formerly, H37Rv induced a high level of LDH release from the cells cultured in the presence of z-VAD-fmk, while H37Ra caused only a moderate level of LDH release (Fig. 8A). In contrast to the strong activities of caspase-9 and the active caspase-9 fragment in the lysate of H37Rv-infected cells (Fig. 8C and D), H37Ra could not induce such levels of caspase-9 activation. The amount of the active form of caspase-9 detected in H37Ra infection was smaller than that induced by H37Rv infection (Fig. 8C and D). It may be suggested that virulent M. tuberculosis has the ability to cause necrosis of the infected macrophages but, at the same time, is capable of activating caspase-9 in order to evade immediate necrotic cell death after infection in host cells where M. tuberculosis must reside for a longer period.

FIG. 8.

Differences between the abilities of the M. tuberculosis H37Rv and H37Ra strains to induce apoptosis, necrosis, and caspase activation in RAW 264 cells. RAW 264 cells were infected with H37Rv or H37Ra at an MOI of 5 in the presence or absence of z-VAD-fmk. (A) The culture supernatant was collected 4 days later, and LDH activity was measured. Two days after cultivation, cells were lysed and the DNA ladder was detected by agarose gel electrophoresis (B) and caspase activities were measured by a Caspase-Glo assay (C). (D) The cell lysate was applied on a sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel, and the amount of active form of caspase-9 was analyzed by Western blotting with caspase-9-specific antibody. The relative intensities of the bands are indicated in the lower graph. Data represent the means ± standard deviations for triplicate assays and are representative of three independent experiments. *, P < 0.05.

DISCUSSION

In the present study, we showed that H37Rv induced the activation of various caspases and that some of the infected cells underwent apoptosis 2 days after infection. Our initial presumption was, therefore, that the addition of the broad-spectrum caspase inhibitor z-VAD-fmk might simply block the apoptosis of RAW 264 cells infected with H37Rv and facilitate the bacterial growth inside. Indeed, generation of oligonucleosomes, which is a representative parameter for apoptosis, was significantly inhibited by the addition of z-VAD-fmk. To our surprise, however, the treatment instead caused necrosis in a very high proportion of the infected cells. These results strongly suggested that some caspases contribute to the inhibition of necrosis of RAW 264 cells induced by H37Rv infection. By using a panel of inhibitors specific for each caspase, we were able to find that caspase-9 is responsible for such an effect through inhibition of intracellular ROS generation. The less virulent strain H37Ra hardly induced caspase-9 activation, and necrosis of infected cells could not be observed even in the presence of z-VAD-fmk. It was suggested that caspase-9-dependent inhibition of necrosis is related to the virulence of M. tuberculosis.

Based on our preliminary study, we selected concentrations of caspase inhibitors suitable for suppression of caspase activities. Schaible et al. used the caspase inhibitor z-VAD-fmk at the same concentration for their investigation (25). Although the appropriate concentration might be high, it might not be high, and it probably differs on the basis of experimental conditions. In the presence of a broad-spectrum caspase inhibitor or caspase-9-specific inhibitor, the growth of H37Rv was limited in RAW 264 cells that underwent necrosis. The current consensus is that apoptosis of macrophages results in the limitation of intracellular survival of M. tuberculosis but that necrosis does not affect the intracellular bacteria and helps M. tuberculosis in dissemination to other macrophages. Recent evidence further demonstrates that virulent M. tuberculosis possesses some inhibitory mechanisms of apoptosis that can be easily induced by less virulent strains (6, 15). In the present experiments, using RAW 264 cells, we observed that H37Ra induced a higher level of DNA fragmentation than H37Rv. In contrast, H37Rv infection caused a significant level of LDH release whereas H37Ra hardly induced LDH release during the initial period of infection. These results are consistent with observations published elsewhere (6, 15). In addition, we found that intracellular growth was inhibited in RAW 264 cells in which severe apoptosis was induced by treatment with actinomycin D (data not shown). As shown here, z-VAD-fmk and z-LEHD-fmk treatment caused severe necrosis of infected cells and the magnitude of necrosis was markedly different from that induced by H37Rv infection alone. It appears that induction of an excessive level of necrosis or apoptosis may eliminate the favorable niche for bacterial growth, resulting in the inhibition of bacterial multiplication in host cells. Our finding is not against the consensus, and we believe that this study could give insight into the role of caspase-9 in the fate of intracellular M. tuberculosis. Although the present data did not reveal whether necrosis of the infected macrophages affected bacterial replication in vivo, it has been reported that uncontrolled mycobacterial growth was observed in necrotic regions in the lungs of sst1^s mice and TNF-α or gamma interferon knockout mice (4, 8, 11, 19, 24). Because activation of the necrosis pathway may allow acceleration of bacterial growth and exacerbation of mycobacterial infection, the caspase-9-dependent necrosis inhibition that was presented in this study may be of additional importance in host defense against infection with virulent M. tuberculosis.

It has been shown that several kinds of caspase species contribute to inhibition of necrosis (18). Vercammen et al. have reported that TNF-α-stimulated L929 cells undergo necrosis when the cells are treated with a caspase inhibitor. They suggested that caspase-1 and caspase-3 might play a role in the inhibition of both ROS generation and necrosis (12, 31). On the other hand, other reports have shown that interaction of FasL and Fas induces necrosis if caspase-8 is inhibited (13, 17, 31). It is reported that caspase-8 inhibits necrosis by inhibition of binding of receptor-interacting protein to the death domain of the Fas receptor (13). In the present study, using M. tuberculosis infection in vitro, we found that caspase-9, but not caspase-1, -3, or -8, exerted a critical role in the inhibition of necrosis. Since there were differences in the requirements of particular caspases to inhibit necrosis of infected cells, it appeared that distinctive signal pathways were activated. One possible interpretation for the caspase-9-dependent inhibition of necrosis is that caspase-9 contributes to stabilization of the mitochondrial membrane and inhibition of ROS production from mitochondria. It has been shown that an excessive generation of ROS was induced from mitochondria in L929 cells after stimulation with TNF-α in the presence of an inhibitor for caspase-1 or caspase-3 (12, 31). Matsumura et al. have shown that a reduction of mitochondrial transmembrane potential (Δψm) was observed in JmF cells treated with FasL and z-VAD-fmk (17) and that pyrrolidine dithiocarbamate, a metallo chelator and antioxidant, efficiently inhibited FasL-induced necrosis. Our preliminary study also showed that z-VAD-fmk treatment caused a reduction of Δψm in H37Rv-infected cells (data not shown). Since the intracellular concentration of ROS was increased when cells were infected with H37Rv in the presence of the caspase-9 inhibitor, caspase-9 might contribute to the inhibition of mitochondrial membrane damage. Further studies are needed to determine the precise mechanism.

It has been reported that various bacterial components are involved in apoptosis induction in cells infected with M. tuberculosis. There are several reports showing that 19-kDa lipoprotein and lipomannan derived from M. tuberculosis induced the apoptosis of macrophages or neutrophils (1, 5, 6, 9, 16). It was also reported that TNF-α was produced after infection with M. tuberculosis and caused apoptosis of macrophages (2, 22). On the other hand, other reports demonstrated that M. tuberculosis possesses an activity inhibiting apoptosis induction. Sly et al. showed that H37Rv had a weaker activity in induction of apoptosis than the attenuated H37Ra strain, and this was due to up-regulation of antiapoptotic gene expression in H37Rv-infected cells (28). Because the intracellular growth of H37Ra in RAW 264 cells was limited compared to that of H37Rv (data not shown), it is possible that activation of the inhibitory process facilitates the intracellular replication of H37Rv. On the other hand, H37Rv caused necrosis when infected cells were treated with the caspase-9 inhibitor. However, H37Ra hardly induced necrosis of cells treated with the inhibitor. Furthermore, we found that caspase-9 was not activated by infection with the attenuated H37Ra strain. Hsu et al. have shown that a mutant strain of H37Rv which is deficient for the RD1 (region of difference 1) region is attenuated for virulence and necrosis-inducing abilities (14). In addition, Park et al. have shown that virulent clinical isolates of mycobacteria strongly induced necrosis of infected macrophages (20). Taken together, these results and our findings suggest that necrosis-inducing activity is associated with the virulence of M. tuberculosis and that caspase-9 activation is probably linked with some mycobacterial virulence determinant.

In conclusion, our present study clearly demonstrated that caspase-9 has a pivotal role in regulation of necrosis induced by infection with H37Rv. We are now trying to address how caspase-9 is activated and how the caspase inhibits necrosis of infected cells. In addition, because necrosis induction appears to be associated with the virulence of mycobacteria, further analysis on the bacterial factor responsible for caspase-9 induction may provide some novel insight for further understanding of host-M. tuberculosis interaction.