Survival of Mycobacterium tuberculosis in Host Macrophages Involves Resistance to Apoptosis Dependent upon Induction of Antiapoptotic Bcl-2 Family Member Mcl-1¹

Mycobacterium tuberculosis, the causative agent of pulmonary tuberculosis, infects one-third of the world’s population (1). It accounts for more deaths each year than any other single infectious bacteria (2). Its ability to survive and replicate in the host macrophage is critical to its pathogenesis, emphasizing a need for a clearer understanding of its interactions with the host macrophage. In vitro experimental models have demonstrated that M. tuberculosis infection causes apoptosis of host macrophages. This has been demonstrated by the detection of annexin V binding to surface-exposed phosphatidylserine early after infection with M. tuberculosis (3). In vitro infection of human macrophages with M. tuberculosis strain H37Ra or strain H37Rv caused TNF-α-dependent apoptosis demonstrated by a genomic DNA ladder, nuclear fragmentation, and condensation as well as by TUNEL labeling (4). Keane et al. (5) demonstrated further that avirulent M. tuberculosis strains and Mycobacterium bovis bacillus Calmette-Guérin (BCG)⁴ cause more apoptosis than virulent strains, implicating a role for apoptosis in host defense against M. tuberculosis.

In addition, there is evidence that apoptosis occurs in vivo during the course of pulmonary tuberculosis. Placido et al. (6) reported that bronchoalveolar lavage from patients with reactive or disseminated tuberculosis demonstrated a 3-fold increase in apoptotic cells compared with control patients as determined by propidium iodide staining, TUNEL labeling, and tissue transglutaminase expression. Furthermore, morphological changes consistent with apoptosis have been noted in epithelioid cells from granuloma of tuberculous patients (7).

Apoptosis during M. tuberculosis infection has been shown to be an innate mechanism of host defense. Apoptosis, but not necrosis, was shown to reduce M. tuberculosis CFUs recovered during in vitro infection of monocyte-derived macrophages (MDMs) (8). Induction of apoptosis in Mycobacterium avium-infected macrophages has been demonstrated to be the cause of intracellular growth restriction induced by picolinic acid (9). Fas ligand (FasL), TNF-α, and ATP-induced apoptosis all caused a reduction in M. tuberculosis strains H37Ra and H37Rv recovered from infected macrophages in vitro (8, 10).

However, it has been demonstrated that Fas-FasL interactions and ATP-induced apoptosis are not responsible for the intracellular death of M. tuberculosis. CD4⁻CD8⁻ cytotoxic T cells mediated Fas-FasL interactions but did not cause the death of intracellular mycobacteria (11). Rather CD8⁺ T cells can cause the death of intracellular M. tuberculosis and other pathogens by a direct granule-dependent mechanism involving the release of the effector protein granuloysin (12). The effect of ATP-induced killing of the intracellular M. tuberculosis is not due to apoptosis but rather has been attributed to enhanced phagosome-lysosome fusion and acidification of the Mycobacterium-containing phagosome (13, 14, 15).

Proapoptotic and antiapoptotic members of the Bcl-2 family have been shown to be differentially regulated by mycobacterial infection. Bcl-2 family members are homologues of the prototypical antiapoptotic protein Bcl-2 (16). The antiapoptotic family member bfl-1, which encodes the protein A1, is up-regulated in macrophages in response to infection with M. bovis BCG (17). The antiapoptotic gene Bcl-x_L has also been shown to be up-regulated during M. tuberculosis infection with a coordinate down-regulation of Bcl-2 (18). In neutrophils, M. tuberculosis infection results in an increase in proapoptotic family member Bax and down-regulation of antiapoptotic family member Bcl-x_L, and these events were shown to be dependent on the generation of reactive oxygen species (19). In an animal model of primary murine tuberculosis, Bcl-2 was up-regulated while Bax was down-regulated, resulting in a net decrease in apoptosis in the host (20).

Mcl-1 is an antiapoptotic member of the Bcl-2 family. Prosurvival members of the Bcl-2 family may prevent apoptosis by maintaining the integrity of the mitochondrial membrane, thereby preventing release of cytochrome c, activation of caspases. and DNA degradative enzymes (21, 22). The Bcl-2 homologue, Mcl-1, is present in cells of the hemopoietic lineage (14, 15, 16, 23, 24, 25). In polymorphonuclear cells, antisense oligonucleotides to mcl-1 have been used to reduce Mcl-1 protein expression, resulting in increased apoptosis in response to aging and hypoxia (26).

To study the role of Mcl-1 in M. tuberculosis-infected macrophages, we examined steady-state levels of mcl-1 mRNA and Mcl-1 protein expression in M. tuberculosis-infected THP-1 cells and the association of Mcl-1 levels with bacterial survival.

RPMI 1640 and HBSS were obtained from StemCell Technologies (Vancouver, BC, Canada). FCS was purchased from HyClone Laboratories (Logan, UT). Middlebrook 7H9, 7H10, and Middlebrook OADC enrichment were obtained from Difco (Detroit, MI). The Kinyoun staining kit was purchased from BBL Microbiology Systems (Cockeysville, MD). Rabbit polyclonal anti-human Mcl-1 Ab was obtained from Upstate Biotechnology (Lake Placid, NY). Goat polyclonal anti-actin Ab was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Secondary goat anti-rabbit Ab conjugated to HRP was obtained from Bio-Rad (Hercules, CA) and secondary anti-goat Ab conjugated to HRP was purchased from Santa Cruz Biotechnology. TRIzol and Superscript Reverse Transcriptase II were obtained from Life Technologies (Grand Island, NY). The Primer Express program and SYBR green master mix were purchased from Applied Biosystems (Foster City, CA). The pGEMT-Easy vector was obtained from Promega (Madison, WI). Annexin V-PE was purchased from BD PharMingen (San Diego, CA). Actinomycin D was obtained from Calbiochem (La Jolla, CA). LipofectAMINE was purchased from Life Technologies. Phosphorothioate-modified oligonucleotides (s-oligos) were synthesized and purified by Life Technologies.

MDMs were isolated as described previously (27). Mononuclear cells were allowed to adhere to cell culture flasks for 45 min at 37°C, 5% CO₂, in a humidified atmosphere. Nonadherent cells were removed with three washes with HBSS and medium was replenished. Adherent cells were used the following day for infections.

The human promonocytic cell line THP-1 was obtained from the American Type Culture Collection (TIB-202; Manassas, VA). M. tuberculosis strains H37Rv and H37Ra were used as virulent and attenuated type strains, respectively. The human myeloid cell line THP-1 was used as a model for human macrophage cells in infection studies because of its similarity to human MDMs and its availability for use (28, 29). THP-1 cells were cultured in RPMI 1640 supplemented with 10% FCS and 2 mM glutamine and were maintained between 2 and 10 × 10⁵ cells/ml. M. tuberculosis was grown to late log phase in Middlebrook 7H9 with OADC. Aliquots were frozen at −70°C. Representative samples were thawed and CFUs per ml were enumerated by plating on Middlebrook 7H10 with OADC. Heat-killed bacteria were prepared by heating aliquots at 80°C for 2 h, resulting in a decrease in CFUs recovered of greater than five logs. Cells were infected as described previously (29). Briefly, THP-1 cells were seeded in six-well flat-bottom tissue culture plates and allowed to adhere and differentiate at 37°C in a humidified, 5% CO₂ atmosphere for 18 h in the presence of 20 ng/ml PMA. Cells were washed and medium was replaced to remove PMA 4 h before the addition of bacteria. Where indicated, bacteria were opsonin coated by incubating with 50% fresh serum in RPMI 1640 for 30 min before addition to culture. Bacteria, opsonin coated or not, were added to THP-1 cells at infection ratios of 5:1, 20:1, 50:1, or 200:1, resulting in infection rates as reported in Results. Latex beads (LB) were used at a ratio of 50:1, resulting in a rate of one to seven beads per cell in ∼90% of cells. Nonopsonin-coated bacteria were added to MDMs at a ratio of 5:1, resulting in an infection rate of one bacilli per macrophage. Infection was evaluated by CFU assays. After 4 h of coincubation of bacteria or LB with cells, cells were washed three times with HBSS to remove noningested particles and the medium was replenished.

SDS-PAGE was performed according to the method of Laemmli (30). Western-blotted membranes were developed by ECL as previously described (31).

Total RNA was isolated from THP-1 cells using TRIzol according to the manufacturer’s instructions. cDNA was prepared by reverse transcription using Superscript Reverse Transcriptase II according to the manufacturer’s instructions. Q-PCR primers for mcl-1 were designed using the Primer Express program and sequences were as follows: forward, 5′-TGGGTTTGTGGAGTTCTTCCA-3′; and reverse, 5′-ACACCTGCAAAAGCCAGCA-3′. Primers were synthesized by Applied Biosystems. The mcl-1 Q-PCR product was cloned into pGEMT-Easy vector according to the manufacturer’s instructions to compare unknown samples to those with known copy numbers of the mcl-1 PCR product. Q-PCR was conducted on cDNA samples and dilutions of plasmid containing the mcl-1 PCR product in triplicate using the SYBR green master mix. Negative controls included RNA with no reverse transcriptase for DNA contamination and no template for environmental contamination.

s-Oligos were prepared to the sense and antisense strands of mcl-1 and incorporated into cells as described previously (29). Briefly, s-oligos to human mcl-1 were synthesized and HPLC purified by Life Technologies. s-Oligos were phosphorothioate-modified to prevent intracellular degradation and purified to remove incomplete synthesis products. The 19-mer antisense sequence was 5′-GGGGCTTCCATCTCCTCAA-3′ as described previously (26) and the sense sequence was 5′-TTGAGGAGATGGAAGCCCC-3′. To incorporate s-oligos into THP-1 cells, 5 × 10⁶ were resuspended in 500 μl of RPMI 1640 containing 2.5% LipofectAMINE/5 μM s-oligos and incubated on a rotary shaker for 4 h at 37°C before differentiation with PMA for THP-1 cells or before allowing monocytes to adhere for MDMs.

Cells were scraped, spun down, and washed twice with 1 ml of cold PBS/10⁶ cells. Cells (1 × 10⁶) were resuspended in annexin V-PE binding buffer (10 mM HEPES (pH 7.4), 140 mM NaCl, and 2.5 mM CaCl₂) at 10⁶ cells/ml. An aliquot (100 μl) was removed and 5 μl of annexin V-PE was added. The mixture was vortexed gently and incubated for 15 min at room temperature in the dark. The cells were washed once with binding buffer and fixed in 30% formaldehyde in methanol. The volume was increased to 500 μl with binding buffer for analysis by flow cytometry.

TUNEL assays were performed using the Fluorescein-FragEL DNA fragmentation kit (Oncogene Research Products, San Diego, CA) according to the manufacturer’s instructions. Treated cells were cytospun onto slides, stained, and examined by fluorescent microscopy. Cells were considered apoptotic if they were TUNEL positive (green fluorescent nuclear staining). A positive control included incubation of THP-1 cells with actinomycin D (50 ng/ml for 19 h).

Enumeration of CFUs was performed as described previously (29, 32). Bacilli were plated immediately after 4 h of coincubation with cells and washing (time 0) and at 4 and 7 days after infection. Bacilli were released from cells in cold PBS/0.1% Triton X-100, serially diluted in Middlebrook 7H9 with OADC, and 20 μl of three dilutions were plated on Middlebrook 7H10 with OADC in triplicate. CFUs were counted after 14 days of incubation at 37°C and plates were maintained for 21 days to ensure that no additional CFUs appeared.

Data presented are expressed as means ± SD. Statistical analyses for Q-PCR, annexin V-PE binding, and TUNEL assays were performed by an unpaired Student’s t test. Comparisons for CFUs were done by ANOVA for each time point. Differences were considered significant at a level of p < 0.05.

Mcl-1 protein levels were determined at several M. tuberculosis infection ratios and at different times after infection by Western blotting. Multiplicity of infections (MOIs) for nonopsonized and opsonized M. tuberculosis strain H37Rv were titrated to determine the optimal MOI for inducing Mcl-1. Infection with nonopsonized bacteria at initial ratios of 5:1, 20:1, 50:1, or 200:1 resulted in MOIs of 0.04, 0.44, 0.81, or 1.80 bacteria per macrophage, respectively, as determined by recovery of CFUs immediately after infection and washing (time 0). Infection at the same ratios with opsonized bacteria resulted in MOIs of 0.59, 1.20, 4.21, and 8.56, respectively. Control cells constitutively expressed low levels of Mcl-1 and enhanced expression of Mcl-1 was observed in macrophages infected with either nonopsonized or opsonized M. tuberculosis (Fig. 1,A). With nonopsonin-coated bacteria, induction of Mcl-1 protein occurred with a MOI of as low as 0.44 and was maximal at 0.81 (Fig. 1,A, lanes 2–5). With opsonized bacteria, induction of Mcl-1 was detected at a MOI of 0.59 and was maximal at 1.20 (Fig. 1 A, lanes 6–9).

Mcl-1 induction by H37Rv infection occurred early after infection. Nonopsonin-coated H37Rv was added to differentiated THP-1 cells at a 50:1 ratio and THP-1 cell lysates were prepared at different times after infection to determine the optimal time for Mcl-1 induction. The amount of Mcl-1 protein present in THP-1 cells was increased as early as 4 h after infection with a maximum at 24 h and that was maintained through 48 h after infection (Fig. 1 B).

Expression of mcl-1 mRNA was evaluated in THP-1 cells that were infected with either H37Rv or H37Ra as well as in cells exposed to either HK H37Rv or LB. In comparison to control cells, there was a 9.6-fold increase in mcl-1 mRNA in H37Rv-infected macrophages (Fig. 2,A). In contrast, no significant increases in mcl-1 gene expression were observed for cells exposed to any H37Ra bacteria, heat-killed M. tuberculosis strain H37Rv (HK) bacteria, or LB (Fig. 2,A). In THP-1 cells, densitometry of ECL-developed bands revealed a 4.8-fold increase of Mcl-1 protein, as determined by Western blotting of whole cell lysates, in H37Rv-infected cells vs uninfected control cells, whereas steady-state levels of Mcl-1 protein were not increased by exposure to either H37Ra, HK bacteria, or LB (Fig. 2,B). Similarly, in MDMs, there was a 2.5-fold increase of Mcl-1 protein measured in cells infected with H37Rv vs uninfected control cells, whereas there was no difference in the amount of Mcl-1 protein detected after exposure of cells to either H37Ra, HK bacteria, or LB (Fig. 2 C).

To determine whether Mcl-1 expression could be reduced in THP-1 cells and MDMs, phosphorothioate modified antisense and sense oligonucleotides (s-oligos) were synthesized based upon sequences in human mcl-1. Incorporation of s-oligos into cells was verified by measuring the fluorescence of cells treated with FITC-labeled s-oligos compared with control cells by flow cytometry. FITC-labeled s-oligos were incorporated into 80–90% of cells in two independent experiments (data not shown). Treatment with antisense, but not sense, s-oligos decreased the amount of Mcl-1 protein in whole cell lysates of noninfected THP-1 cells by >90% and in noninfected MDMs by >96% (Fig. 3,A). Mcl-1 protein was similarly reduced by 90% in whole cell lysates of THP-1 cells and MDMs infected with H37Ra and by 86% or 84%, respectively, in whole cell lysates of THP-1 cells or MDMs infected with H37Rv (Fig. 3 B).

The contribution of Mcl-1 to THP-1 cell and MDM apoptosis during M. tuberculosis infection was examined by annexin V-PE binding to phosphatidylserine on the cell surface. Phosphatidylserine surface expression, an early marker of cell apoptosis, was assessed by annexin V-PE binding and flow cytometric analysis at 1 and 4 days after infection. One day after infection, compared with controls, the frequencies of annexin V-PE-positive cells were increased in response to infection with H37Ra but not H37Rv (Fig. 4, A and C). This effect was not influenced by treatment with mcl-1 antisense. One day after infection with H37Rv, the percentages of annexin V-PE-positive THP-1 cells were similar in untreated and sense s-oligo-treated cells at 12 ± 1% and 15 ± 1%, respectively. However, the percentage of H37Rv-infected cells that were positive for annexin V-PE binding increased dramatically to 29 ± 4% in THP-1 cells that were treated with antisense s-oligos to mcl-1 (Fig. 4,A, left graph). Pretreatment of cells with antisense s-oligos to mcl-1 brought about a similar increase in annexin V-PE binding positive cells at 4 days postinfection with H37Rv (Fig. 4,A, right graph). M. tuberculosis strain H37Rv-infected cells treated with sense s-oligos were 28 ± 2% annexin V-PE binding positive, whereas antisense s-oligo treatment increased this to 53 ± 6% positive cells (Fig. 4,A, right graph). Whereas annexin V binding was again higher in H37Ra-infected cells, this time at 4 days after infection (Fig. 4 A, right graph), as had been observed at 1 day postinfection, this response was not affected by antisense mcl-1 treatment.

Qualitatively similar results were observed with annexin V-PE binding to untreated and sense- and antisense-treated MDMs that were noninfected or infected with M. tuberculosis strains H37Ra and H37Rv 1 and 4 days after infection (Fig. 4,C, left and right graphs, respectively), although infection caused a larger amount of annexin V-PE binding in these cells. Infection with strain H37Ra caused significantly more annexin V-PE binding than did infection with strain H37Rv. Furthermore, antisense oligonucleotides to mcl-1 caused an increase in annexin V-PE binding positivity in H37Rv-infected cells, but not in control cells or H37Ra-infected cells at 1 and 4 days after infection (Fig. 4 C).

To ensure that the cause of the annexin V-PE binding to cells was apoptosis, we evaluated the number of treated THP-1 cells displaying nuclear fragmentation by TUNEL assay at 1 and 4 days after treatment and infection (Fig. 4,B). One day after infection with H37Rv, the percentages of TUNEL-positive THP-1 cells were similar in untreated and sense s-oligo-treated cells at 7 ± 1% and 8 ± 1%, respectively. However, the percentage of H37Rv-infected cells that were TUNEL positive increased dramatically to 21 ± 2% in THP-1 cells that were treated with antisense s-oligos to mcl-1 (Fig. 4,B, left graph). Pretreatment of cells with antisense s-oligos to mcl-1 caused a similar increase in TUNEL-positive cells 4 days after infection with H37Rv (Fig. 4,B, right graph). M. tuberculosis strain H37Rv-infected cells treated with sense s-oligos were 17 ± 2% annexin V-PE binding positive, whereas antisense s-oligo treatment increased this to 36 ± 3% positive cells (Fig. 4,B, right graph). Although TUNEL positivity was again higher in H37Ra-infected cells at 4 days after infection (Fig. 4 B, right graph), as had been observed at 1 day postinfection, this response was not affected by antisense mcl-1 treatment.

Mcl-1 induction in H37Rv-infected cells did not protect host cells from apoptosis induced by FasL or by ATP (data not shown). THP-1 cells were noninfected or infected with H37Rv or H37Ra followed by treatment with either 20 U/ml FasL or 20 mM ATP to induce apoptosis. Apoptosis was then measured by annexin V-PE binding to the surface of the cell membrane. FasL treatment resulted in induction of apoptosis in ∼72% of cells after 24 h and this was not significantly different in noninfected cells, H37Rv-infected cells, or H37Ra-infected cells. ATP resulted in annexin V-PE binding in >84% of cells after 24 h and this also was not significantly different in noninfected, H37Rv-infected, or H37Ra-infected cells.

To investigate whether M. tuberculosis strain H37Rv-induced Mcl-1 expression correlated with the intracellular survival of bacteria, growth of strains H37Ra and H37Rv was evaluated in THP-1 cells (Fig. 5,A) and MDMs (Fig. 5,B) that were left untreated or treated with sense or antisense s-oligos to mcl-1. Intracellular growth of H37Ra was similar whether cells were left untreated or treated with sense or antisense s-oligos to mcl-1. Growth of strain H37Rv was similar in untreated cells or cells treated with sense s-oligos to mcl-1. However, there were dramatic decreases in CFUs recovered from antisense s-oligo-treated cells at both 4 and 7 days after infection with H37Rv compared with cells treated with sense s-oligos (Fig. 5). Recovery of CFUs from antisense-treated THP-1 cells was 42 ± 3% of CFUs recovered from sense s-oligo-treated, H37Rv-infected cells at 4 days after infection, and this was reduced further to only 21 ± 2% at 7 days after infection (Fig. 5,A). Recovery of CFUs from antisense-treated MDMs was 56 ± 4% of CFUs recovered from sense s-oligo-treated, H37Rv-infected cells at 4 days after infection, and this was reduced further to only 19 ± 2% by 7 days after infection (Fig. 5 B).

This study examined changes in the expression of Bcl-2 family member Mcl-1 during infection with M. tuberculosis and the role of these changes to host cell apoptosis and to intracellular survival of bacteria. Intracellular infection with various species of mycobacteria has been reported to cause host cell apoptosis (33). Furthermore, infection with relatively avirulent species and strains such as M. tuberculosis H37Ra, M. bovis BCG, and Mycobacterium kansasii have been observed to induce more apoptosis than their more virulent counterparts including M. tuberculosis H37Rv, M. bovis, and a clinical M. tuberculosis isolate (5). This implies a role for apoptosis in host defense against the infection. Indeed, in infection models, apoptosis of the host macrophage has been shown to reduce the viability of intracellular M. tuberculosis (8, 9, 10).

The results presented in Fig. 1,A demonstrate that induction of Mcl-1 by infection with virulent M. tuberculosis strain H37Rv occurred whether or not bacteria were serum opsonized (Fig. 1 A). This suggests that Mcl-1 induction is independent of the route of entry of the bacilli. It has been reported that at least seven different receptors including CR1, CR3, CR4, mannose receptor, scavenger receptors, CD14, and sp-A are involved in the entry of M. tuberculosis into macrophages (34). Different receptor pathways may be favored by serum opsonization, but Mcl-1 induction occurs in the presence or absence of serum opsonins.

It has been observed that low MOIs for M. tuberculosis cause less apoptosis than high MOIs and in contrast that production of TNF-α was inversely proportional to MOI (35). Whereas this may represent a mechanism by which M. tuberculosis reduces host cell apoptosis and replicates within the host macrophage, we observed Mcl-1 induction to occur over a wide range of MOIs. Mcl-1 induction was evident at a MOI as low as 0.44 and increased further up to a MOI of 8.6, suggesting that induction of this antiapoptotic mechanism does not require low MOIs (Fig. 1,A). Optimal induction of Mcl-1 occurred at 24 h after infection (Fig. 1 B), suggesting that this mechanism is not triggered by phagocytosis per se, but rather by something elaborated by internalized bacteria.

In this study, H37Rv infection induced mcl-1 gene expression (Fig. 2,A) and Mcl-1 protein expression (Fig. 2, B and C), an antiapoptotic gene product, but H37Ra does not. This is consistent with a model in which host cell apoptosis provides an antibacterial defense mechanism that is subverted by virulent strain H37Rv (4, 5, 36). Mcl-1 induction did not occur after infection with H37Ra or phagocytosis of HK or LB, verifying that the response is not triggered by phagocytosis (Fig. 2), but rather infection with viable, virulent M. tuberculosis is required for induction. Taken together with the finding that maximal induction of Mcl-1 occurred between 24 and 48 h after infection, these results suggest that the induction of Mcl-1 is dependent upon a factor elaborated by viable and virulent M. tuberculosis within the phagocyte.

Antisense oligonucleotides to mcl-1 reduced the expression of Mcl-1 protein in control cells (Fig. 3,A) and in cells infected with either M. tuberculosis strain H37Ra or strain H37Rv (Fig. 3,B). Mcl-1 expression in antisense s-oligo-treated, H37Rv-infected cells was lower than basal expression in untreated, uninduced control cells (Fig. 3,B, lane 5 vs lane 1). The magnitude of the reduction in Mcl-1 expression was similar to that observed previously using the same antisense oligonucleotide sequence (26) where mcl-1 antisense treatment resulted in an increase in apoptosis in polymorphonuclear cells induced by hypoxia while having no effect on control cells (26). Treatment with antisense s-oligos to mcl-1 also resulted in an increase in apoptotic host cells in response to infection with H37Rv (Fig. 4). This suggested that Mcl-1 induction by virulent bacteria is an antiapoptotic mechanism. In contrast, for the attenuated strain H37Ra, which did not induce Mcl-1 expression, antisense treatment did not influence the level of apoptosis observed (Fig. 4).

Suppression of apoptosis by strain H37Rv may provide a means for increased intracellular bacterial proliferation. Inhibition of mcl-1 expression by treatment with antisense s-oligos resulted in a reduced ability of H37Rv to grow within differentiated THP-1 cells and MDMs, respectively (Fig. 5). The induction of apoptosis in antisense-treated, H37Rv-infected THP-1 cells (Fig. 4, A and B) correlated with reduced recovery of CFUs from infected cells (Fig. 5,A). Similarly, the annexin V-PE binding of antisense s-oligo-treated, H37Rv-infected MDMs (Fig. 4,C) correlated with reduced recovery of CFUs from infected MDMs (Fig. 5 B). This finding is consistent with a growing appreciation that apoptosis of infected macrophages leads to decreased survival of M. tuberculosis. For example, Molloy et al. (8) first described the antimycobacterial effect of host apoptosis in ATP-treated macrophages compared with the lack of antimycobacterial activity of necrotic macrophages (hydrogen peroxide treated). In addition, FasL-induced apoptosis had an antimycobacterial effect in vitro against M. tuberculosis strains H37Ra and H37Rv as did TNF-α-induced apoptosis (10). Recently, the antimycobacterial activity of picolinic acid against M. avium was attributed to induction of host cell apoptosis (9).

Although the increase in apoptosis in antisense oligonucleotide, H37Rv-infected THP-1 cells and MDMs correlated with a decrease in intracellular proliferation of this strain, the amount of cellular apoptosis did not correlate with intracellular growth throughout this study. For example, H37Ra infection was observed consistently to cause more host cell apoptosis than strain H37Rv (Fig. 4), yet its growth in untreated cells or sense oligonucleotide-treated cells was comparable to that of strain H37Rv. Hence, it may be that there is a threshold amount of apoptosis required to negatively effect intracellular growth of H37Rv. Alternatively, there may be mechanisms in addition to apoptosis that are negatively regulated by Mcl-1 which contribute to the reduced recovery of H37Rv when they are released from the inhibitory control of Mcl-1.

Attenuation of host cell apoptosis via induction of the antiapoptotic mechanisms by intracellular pathogens may provide a means to evade the host response, allowing increased bacterial proliferation (37). One example of a mycobacterial antiapoptotic mechanism is reduced Fas expression during M. tuberculosis infection, which may provide a means for M. tuberculosis to limit FasL-mediated apoptosis (10), promoting intracellular survival. A second example may relate to inhibition of TNF-α-induced apoptosis by virulent M. tuberculosis. Despite the fact that infection of macrophages with strains H37Rv or H37Ra leads to similar levels of TNF-α production, M. tuberculosis strain H37Ra causes more apoptosis than does strain H37Rv. This paradox may be explained by the finding that infection with M. tuberculosis strain H37Rv brings about greater expression of soluble TNFR 2 (sTNFR2), thereby blocking the activity of TNF-α (38). It has been proposed, therefore, that expression of sTNFR2 provides strain H37Rv with a means to decrease host cell apoptosis, thereby increasing its intracellular survival. Increased expression of sTNFR2 by virulent strain H37Rv, but not attenuated strain H37Ra is consistent with the observation that avirulent strains of mycobacteria cause more apoptosis in host macrophages than virulent strains. Furthermore, the association of more host cell apoptosis with avirulent strains suggests that macrophage apoptosis is a host defense mechanism used to prevent proliferation of internalized bacilli (4, 5, 36). In this study, infection with M. tuberculosis strain H37Ra promoted significantly more macrophage apoptosis than was found during infection with strain H37Rv or in uninfected control cells 1 and 4 days after infection (Fig. 4). This observation correlates with the strain’s inability to induce Mcl-1 expression and provides an additional mechanism whereby virulent strains of mycobacteria may cause less apoptosis than avirulent strains.

Mcl-1 induction may be a direct effect of signaling from M. tuberculosis. Alternatively, it may be an indirect effect of M. tuberculosis infection-induced cytokine production or a combination of both. It has recently been shown that TNF-α signaling in human prostate cancer cells leads to antiapoptotic bcl-2 gene expression and bcl-2 protein expression via NF-κB (39). At present, although there has not been any direct evidence presented to link induction of other bcl-2 family members to TNF-α signaling or NF-κB activation, it is conceivable that other family members may be similarly regulated. Induction of another antiapoptotic Bcl-2 family member, A1 protein encoded by the gene bfl-1, has recently been reported during infection of murine macrophages with live or dead M. bovis BCG (18). A1 induction was associated with the protection of host macrophages from NO-induced apoptosis. Mcl-1 induction did not protect H37Rv-infected macrophages from either FasL or ATP-induced apoptosis. The role of A1 in intracellular survival of BCG was not assessed (18).

Virulent mycobacteria use multiple, distinct strategies to prevent host cell apoptosis. Infection with virulent strains has been shown to prevent apoptosis by minimizing external signals that induce host cell apoptosis. Two such strategies have been demonstrated thus far including the decrease in Fas expression, which reduces FasL-induced apoptosis (10), and the increase in sTNFR expression, which competes for binding of TNF-α with the membrane receptor, thereby reducing TNF-α-mediated host cell apoptosis (38). Infection with virulent M. tuberculosis also increases expression of antiapoptotic proteins and this effect may be a direct effect of M. tuberculosis or may be an indirect effect of M. tuberculosis infection-induced cytokine production. Antiapoptotic genes induced after infection with virulent M. tuberculosis include bcl-x_L, bfl-1, and mcl-1, as demonstrated here (17, 18).

In this study, Mcl-1 was induced by infection of THP-1 cells with viable and virulent M. tuberculosis, but not in response to phagocytosis or ingestion of heat-killed or attenuated M. tuberculosis. Induction of this antiapoptotic protein limited the extent of apoptosis in virulent M. tuberculosis-infected macrophages as inhibition of Mcl-1 expression resulted in increased apoptosis of infected cells. This mechanism appears to contribute to the differential amounts of apoptosis observed when comparing virulent and avirulent mycobacterial strains and species. Finally, the induction of antiapoptotic Mcl-1 is important for the intracellular survival and/or proliferation of virulent M. tuberculosis as reduction of Mcl-1 expression by antisense oligonucleotides resulted in decreased CFUs recovered from H37Rv-infected macrophages. An important objective at this stage is to determine the nature of the factor produced during infection with viable and virulent strain H37Rv that leads to Mcl-1 induction. The evidence suggests that this factor must be either actively released or synthesized intracellularly since Mcl-1 induction was not observed in response to internalization of either dead bacilli or attenuated strain H37Ra. In summary, these findings present a novel mechanism by which virulent M. tuberculosis evades apoptosis, an innate, protective host cell response to infection, thereby promoting its intracellular survival and proliferation.

We thank Drs. Vince Duronio and Kathryn Schubert for their valuable advice and technical assistance.