Mycobacterium tuberculosis Lipoprotein MPT83 Induces Apoptosis of Infected Macrophages by Activating the TLR2/p38/COX-2 Signaling Pathway

With an estimated 10.4 million new cases and 1.4 million deaths in 2015, tuberculosis (TB) caused by the bacterial pathogen Mycobacterium tuberculosis remains a major global health challenge (1). Inhaled aerosolized droplets containing M. tuberculosis were first phagocytosed by alveolar macrophages, which comprise the first line of defense against the invading pathogen and could also serve as a sanctuary for M. tuberculosis if the immune responses mounted fail to contain the bacteria. As a highly adapted human pathogen, M. tuberculosis has successfully evolved numerous strategies to circumvent host immune defenses (2). Virulent M. tuberculosis evades elimination in macrophages by blocking phagosome maturation, manipulating host cell death programs, modulating the inflammatory responses, and suppressing MHC class I– and class II–mediated Ag presentation (3–5).

It has been well documented that subversion of the cell death pattern of infected macrophages plays a crucial role in the pathogenesis of TB. Abundant necrosis-characterized cell death (plasma membrane lysis and escape of the pathogens) was observed in virulent M. tuberculosis–infected macrophages, affording a protective milieu for the spreading of bacteria. In contrast, apoptosis-characterized cell death (plasma membrane intact and promotion of pathogen killing) was induced by avirulent strains of M. tuberculosis, leading to killing of intracellular bacilli and activation of adaptive immunity by cross-presentation (6, 7). Accumulating evidence indicates a close relationship between lipid mediators and M. tuberculosis–driven macrophage death. Infection of macrophages with avirulent M. tuberculosis induces PGE₂ production to prevent mitochondrial damage and repair the plasma membrane disruptions caused by M. tuberculosis, and this leads to apoptosis (8, 9). However, virulent M. tuberculosis–infected macrophages produce predominant LXA₄, which reduces PGE₂ biosynthesis through blocking cyclooxygenase-2 (COX-2) expression (8). Thus, M. tuberculosis interferes with host eicosanoid biosynthetic pathways to promote necrosis or avoid apoptosis (9).

Increasing evidence supports that M. tuberculosis–secreted proteins and lipoproteins interfere with macrophage apoptosis. Through a gain-of-function screening in Mycobacterium smegmatis, NuoG (Rv3151) was identified as a key M. tuberculosis gene involved in the inhibition of macrophage apoptosis via neutralizing host NOX2-derived reactive oxygen species (10, 11). Additionally, mutation of the secA2 (Rv1821) gene in M. tuberculosis H37Rv strain enhanced the apoptosis of infected macrophages by impairing secretion of bacterial superoxide dismutase (12). Moreover, M. tuberculosis–bearing plcA and plcB (Rv2351c and Rv2350c) genes led to alveolar macrophage necrosis by subversion of PGE₂ production (13). Besides, M. tuberculosis also used CpnT (Rv3903c) protein to induce macrophage necrosis by hydrolyzing NAD (14, 15). Although M. tuberculosis induces significant apoptosis, few M. tuberculosis genes involving in promoting apoptosis have been identified. ESAT6 (Rv3875) was reported to induce THP-1 apoptosis through upregulating the expression of caspase-1, -3, -5, -7, and -8 genes (16). LpqH (Rv3763) and PPE32 (Rv1808) were found to trigger apoptosis in both caspase-dependent and caspase-independent manners (17, 18).

MPT83 (Rv2873), an M. tuberculosis surface-expressing lipoglycoprotein, facilitates both innate and adaptive immune response to M. tuberculosis infection. Following recognition by TLR2, MPT83 induces the production of cytokines, including TNF-α, IL-6, and IL-12 p40 as well as IFN-γ–induced MHC class II presentation in macrophages. As a result of increased IL-12 production, activated macrophages presented MPT83 peptide to CD4⁺ T cells after immunization with rMPT83 (19, 20). Preimmunization with MPT83 DNA or protein vaccine generated a strong IFN-γ⁺ T cell responses and CD8⁺ T cell response to the specific epitope MPT83 (127–135) (PTNAAFDKL), which conferred protection in mice from M. tuberculosis infection manifesting with lower bacterial burden in the lung and spleen tissues (21). However, little is known about the mechanism underlying the protective effect of MPT83-relevant subunit vaccine. In this study, we report that recognition of MPT83 by TLR2 mediated MAPK p38 activation, followed by COX-2 production, which in turn led to enhanced macrophage apoptosis. Furthermore, expression of MPT83 in M. smegmatis improved host bacterial clearance and enhanced mouse survival.

Tlr2^−/− and tlr4^−/− mice on a C57BL/6 background were obtained from The Jackson Laboratory. Mice were bred in specific pathogen-free conditions at the Laboratory Animal Center of Tongji University. Female mice at the age of 6–8 wk old were used in the experiments. All animal experiments were reviewed and approved by the Animal Experiment Administration Committee of Tongji University School of Medicine, China. For experiments, mice were randomly chosen from different genotype groups (n > 30) housed in individual cage.

COX-2 inhibitor [N-(2-cyclohexyloxy-4-nitrophenyl)methane sulphonamide (NS398)], p38 inhibitor (SB203580), and NF-κB inhibitor (PDTC) were purchased from Sigma-Aldrich. Abs against phospho-p65 (3033), phospho-p38 (9215), and GAPDH (2118) were obtained from Cell Signaling Technology. COX-2 Ab (ab62331) was purchased from Abcam. All of the Abs were used at a dilution of 1:1000 for Western blot.

All M. smegmatis strains used in this study were derived from the laboratory strain mc2155. Coding sequences for MPT83 (663 bp) were amplified by PCR from M. tuberculosis H37Rv genomic DNA using the primers MPT83 forward (5′-AGCCAAGACAATTGCGGATCCATGATCAACGTTCAGGCCAAACCGG-3′) and reverse (5′-TCGAATTCTGCAGCTGGATCCTTACTGTGCCGGGGGCATCAGCACC-3′). The cDNA of MPT83 was cloned into the mycobacterial−Escherichia coli shuttle vector pMV261 and then genotyped by sequencing. M. smegmatis was transformed with either empty pMV261 or pMV261-MPT83 by electroporation and then selected on Middlebrook 7H10 agar containing 10% oleic acid–albumin–dextrose–catalase (OADC) and 50 μg/ml kanamycin. The resulting recombinant transformants consisting of either pMV261 or pMV261-MPT83 were defined as MS_Vec or MS_MPT83, respectively. Strains of MS_Vec and MS_MPT83 were cultured at 37°C in 7H9 liquid medium (Becton Dickinson) containing 0.5% (v/v) glycerol, 0.05% (v/v) Tween 80, 10% (v/v) OADC (Becton Dickinson), and 50 μg/ml kanamycin.

MPT83 gene was cloned pET28a with His tagged at the C terminus. E. coli BL21 bacteria carrying recombinant plasmid was induced with isopropyl β-d-thiogalactopyranoside (IPTG), and His-tagged rMPT83 was purified with Ni-NTA columns (Invitrogen). The LPS was removed by a ToxinEraser endotoxin removal kit (GenScript), and the residual LPS was detected by an endotoxin detection kit (Houshiji) and the concentration of LPS left in the purified protein is <0.05 endotoxin unit/mg. The purity of the rMPT83 protein was analyzed by SDS-PAGE, followed by Coomassie blue staining. The concentration of MPT83 protein was detected with a BCA protein assay kit (Thermo Fisher).

Mouse bone marrow–derived macrophages (BMDMs) were differentiated in vitro from isolated bone marrow cells. Briefly, bone marrow cells collected from mouse femurs and tibias were incubated for 7 d in DMEM containing 25% heat-inactivated FBS, 1% penicillin, 1% streptomycin, and 25% L929 cell–conditioned medium.

Immortalized BMDMs (iBMDMs) were gifted by Dr. F. Shao (National Institute of Biological Sciences, Beijing, China). Lentivirus-mediated knockdown of specific genes was performed to generate COX-2 knockdown iBMDMs. Briefly, pLKO.1-puro vector containing scrambled sequence and sequence targeting mouse COX-2 (5′-CCGGCCGTACACATCATTTGAAGAACTCGAGTTCTTCAAATGATGTGTACGGTTTTTG-3′) was transfected into HEK293T cells with packing plasmids pSPAX2 and pMD2.G at a ratio of 4:3:1. Lentiviruses were collected 48 h later for the transfection of iBMDMs, followed by selection with puromycin (Santa Cruz Biotechnology) at the concentration of 2 μg/ml. The knockdown efficiency of COX-2 was detected by Western blot.

Apoptosis and necrosis were detected according to the manufacturer’s instructions (MultiSciences Biotech). Briefly, cultured cells (∼2 × 10⁵ cells) treated as indicated were washed twice with PBS and incubated in 200 μl of binding buffer containing 5 μl of Annexin V^FITC and 10 μl of propidium iodide (PI) in the dark for 5 min at room temperature. Then, the stained samples were analyzed on the flow cytometer (BD FACSVerse). Total cell death was detected through lactate dehydrogenase (LDH) measurement by a CytoTox 96 nonradioactive cytotoxicity assay (Promega) following the manufacturer’s instructions.

BMDMs were either infected with M. smegmatis or stimulated with rMPT83 protein for indicated times. Subsequently, cells were lysed with 1× SDS loading buffer on ice, followed by boiling at 95°C for 10 min. Cell lysates and prestained protein markers were subjected to SDS-PAGE and transferred to nitrocellulose membrane. ECL reagents (Thermo Scientific) were applied for immunoblot analysis.

Total RNA was isolated using TRIzol reagent (Life Technologies) according to the manufacturer’s instructions. After DNase treatment, cDNA was synthesized using a PrimeScript RT reagent kit (TaKaRa Bio, Kyoto, Japan). Quantitative real-time PCR analysis was performed using an Applied Biosystems 7300 real-time PCR system and SYBR Green real-time PCR master mixes (Toyobo, Osaka, Japan) for cox2 and gapdh expression, respectively. The relative expression level of cox2 was normalized to that of internal control gapdh using the 2^−ΔΔCt cycle threshold method. Primer sequences were listed as follows: cox-2 forward, 5′-ACTGGCTTGTATAGACTACTGGT-3′, reverse, 5′-TTCAACACACTCTATCACTGGC-3′; and gapdh forward, 5′-AGAAGCGTTTGCGGTACTCAT-3′, reverse, 5′-GGGGTCGTTGATGGCAACA-3′. All of the quantitative PCR experiments were conducted in duplicates in each experiment, and experiments were replicated at least three times.

M. smegmatis cultures for infection were grown from low-passage freezer stocks to midlog phase, washed, and resuspended in PBS. To obtain single-cell suspensions, bacteria were sonicated before infection. Mice were infected with M. smegmatis through the tail vein either at 5 × 10⁷ CFU in 200 μl for mouse survival experiment or 5 × 10⁶ CFU in 200 μl for tissue CFU detection.

The left part of the lung tissues from infected mice was homogenized in 1 ml of PBS. Homogenates in 10-fold serial dilutions were plated on Middlebrook 7H10 agar supplemented with 10% OADC enrichment (both from Becton Dickinson) and incubated at 37°C. Colonies were counted after 3 d.

Densitometry quantification was performed with ImageJ software (ImageJ 1.51e). Data are expressed as the mean ± SD unless otherwise indicated. GraphPad Prism5 (GraphPad Software) was used for statistical analysis. Statistical significance between groups was determined by a two-tailed unpaired t test. Bacterial titers were analyzed using the Mann–Whitney U test. For the mouse survival study, Kaplan–Meier survival curves were generated and analyzed by a Gehan–Breslow–Wilcoxon test. A p value <0.05 represented a statistically significant difference.

M. tuberculosis infection leads to the apoptosis or necrosis of macrophages, which finally decides the outcome of the infection (4). Previous studies have indicated that M. tuberculosis MPT83 as an important immunomodulator in the process of M. tuberculosis infection (19, 20). To further characterize the modulatory function of MPT83, we systematically studied the effect of MPT83 on mycobacteria-driven apoptosis in macrophages by constructing recombinant M. smegmatis strain overexpressing MPT83 (MS_MPT83) and purification of rMPT83 protein. A PI and annexin V labeling system was used to distinguish apoptosis and necrosis. Annexin V⁺PI⁻ indicates early apoptosis and annexin V⁺PI⁺ indicates late apoptosis, whereas annexin V⁻PI⁺ indicates necrosis. First, mouse BMDMs isolated from wild-type (WT) C57BL/6 mice and PMA-differentiated human monocytic THP-1 cells were infected with M. smegmatis transfected by the shuttle vector pMV261 (MS_Vec) or MS_MPT83 for 24 h and the apoptotic cells were detected by FACS assay. In both mouse and human macrophages, MS_MPT83 induced significantly more apoptosis than did MS_Vec (Fig. 1A, 1B). Furthermore, rMPT83 induced the apoptosis of mouse BMDMs in a dose-dependent manner (Fig. 1C, 1D, Supplemental Fig. 1). The direct effect of rMPT83 on the induction of apoptosis indicated that a cell surface molecule on macrophages could be involved in the process. Taken together, our data indicate that MPT83 enhances macrophage apoptosis in both mouse and human macrophages.

COX-2, an inducible PGH synthase, is required for the synthesis of prostanoid, including PGD₂, PGE₂, PGF_2α, PGI₂, and thromboxane (22). It has been well documented that PGE₂ is critical for the repair of membrane microdisruptions caused by M. tuberculosis infection, which finally decides the fate of infected cells (9). The large amounts of PGE₂ production in response to avirulent M. tuberculosis infection supports macrophage apoptosis and containment of the mycobacteria. In contrast, virulent M. tuberculosis develops a strategy to inhibit the production of COX-2 and PGE₂, which shifts the infected macrophages toward necrosis and favors the spreading of M. tuberculosis (8, 9). Because MPT83 markedly enhanced macrophage apoptosis, we then interrogated the correlation between COX-2 expression and MPT83-induced apoptosis. First, the kinetics of induction of COX-2 transcripts and protein in mouse BMDMs infected with MS_Vec or MS_MPT83 were evaluated by quantitative real-time PCR and Western blot, respectively. Overexpression of MPT83 in M. smegmatis markedly upregulated the expressions of COX-2 at both the mRNA and protein levels in macrophages (Fig. 2A, 2B). Consistently, rMPT83 induced COX-2 expression in mouse BMDMs in a time- and dose-dependent manner (Fig. 2C–E).

Next, to determine whether COX-2 plays an essential role in the MPT83-induced apoptosis process, we stimulated mouse BMDMs with rMPT83 in the absence or presence of NS398, which has been reported to selectively inhibit COX-2 activity, and assessed cell death by FACS or LDH assay. NS398 treatment resulted in a marginal increase of apoptotic BMDMs due to the cytotoxicity of the compound, whereas NS398 dramatically inhibited the apoptosis of macrophages treated with rMPT83 (Supplemental Fig. 2A, 2B). Consistently, inhibition of COX-2 by NS398 markedly reduced rMPT83-induced death of BMDMs as measured by LDH (Supplemental Fig. 2C). Furthermore, blockade of COX-2 with NS398 eliminated the enhanced cell death that resulted from overexpression of MPT83 in M. smegmatis observed in DMSO-treated mouse BMDMs (Supplemental Fig. 2D). Furthermore, we generated COX-2 knockdown iBMDMs by a lentivirus transfection system (Fig. 2F). Consistent with the inhibition of COX-2, COX-2 silencing resulted in reduced apoptosis and death of macrophages in response to rMPT83 treatment (Fig. 2G–I). These results suggest that MPT83 induces macrophage apoptosis mainly by regulating COX-2 expression or activity.

It has been reported that engagement of TLR2 by MPT83 protein upregulated proinflammatory cytokine expression in macrophages (20). We then examined whether TLR2 is required for MPT83-induced COX-2 expression. Thus, we detected the abundance of COX-2 transcripts and protein in WT and tlr2^−/− BMDMs infected with either MS_Vec or MS_MPT83 by quantitative real-time PCR and Western blot, respectively. We observed that TLR2 deficiency led to significant reduction of COX-2 at both the transcriptional and protein levels in BMDMs infected with either MS_Vec or MS_MPT83 (Fig. 3A, 3B). Additionally, the MS_MPT83-induced augmentation of COX-2 expression in WT BMDMs was eliminated in TLR2-deficient cells. Consistently, rMPT83-induced COX-2 protein expression was significantly reduced in TLR2-deficient BMDMs (Fig. 3C).

We next determine the possible involvement of TLR2 in MPT83-induced apoptotic process. As expected, the deficiency of TLR2 but not TLR4 significantly blocked MPT83-induced augmentation of cytotoxicity in infected BMDMs (Fig. 3D). Consistently, although MS_MPT83 induces more apoptosis in WT BMDMs in comparison with MS_Vec, as demonstrated by FACS analysis, the increment was not observed in infected tlr2^−/− BMDMs (Fig. 3E, 3F). Moreover, the deficiency of TLR2 dramatically reduced rMPT83-driven apoptosis in infected mouse BMDMs as detected by FACS assay (Fig. 3G), indicating an essential role of TLR2 in MPT83-induced apoptosis. Taken together, these results indicate that MPT83 might trigger TLR2-mediated COX-2 expression to induce macrophage apoptosis.

To further understand how TLR2 regulates MPT83-induced COX-2 expression, we next assessed the effect of MPT83 on TLR2 downstream signaling activation including MAPKs and NF-κB. In line with a previous report (20), MS_MPT83 infection induced significantly more phosphorylation of p38 and p65 compared with MS_Vec (Fig. 4A). Consistently, stimulation of mouse BMDMs with rMPT83 at different concentrations robustly induced the phosphorylation of p38 and p65 (Fig. 4B). To determine whether the activation of p38 and NF-κB pathways are involved in the regulation of COX-2 expression, we treated BMDMs with p38 inhibitor (SB203580) or NF-κB inhibitor (PDTC) for 1 h before stimulating with rMPT83. Inhibition of p38 by SB203580 dramatically reduced MPT83-induced COX-2 expression at both the transcriptional and translational levels, whereas no effect was observed with PDTC treatment (Fig. 4C, 4D). Collectively, these data suggested that MPT83 might induce COX-2 expression to promote macrophage apoptosis through the TLR2/p38 signaling pathway.

Apoptosis of infected macrophages has been reported to lead to intracellular bacteria clearance (23), and it acts as a reservoir of Ag that facilitates initiation and shaping of acquired T cell immunity (6). Therefore, we assumed that promotion of apoptosis by MPT83 may protect mice from Mycobacterium infection. To test this hypothesis, we employed an acute mycobacterial infection model by i.v. injection of M. smegmatis in mice. By infecting WT and tlr2^−/− C57BL/6 mice with MS_Vec or MS_MPT83, we observed that MS_MPT83 significantly reduced the bacterial load in the lung of infected WT mice but not tlr2^−/− mice (Fig. 5A). Consistently, MS_MPT83 infection significantly improved mouse survival in comparison with MS_Vec (Fig. 5B). Thus, MPT83-induced apoptosis may contribute to the protection against mice from mycobacteria infection.

Numerous studies have shown that M. tuberculosis induces macrophage apoptosis and necrosis, which exert distinct functions in host defense mycobacterial infection (9, 23–25). However, many questions remain to be addressed, including the identification of mycobacterial factors responsible for evoking apoptotic responses, clarification of the mechanisms underlying the apoptotic response to M. tuberculosis infection, and definition of the exact effect of apoptotic cell death on the outcome of infection. In this study, we identified MPT83 as an M. tuberculosis component that induces macrophage apoptotic cell death through the TLR2/p38/COX-2 signaling pathway (Supplemental Fig. 3). It is well accepted that apoptosis is an effective host strategy defense to mycobacterial infection through evoking innate immunity and subsequently boosting M. tuberculosis–specific T cell immune response (9, 23, 24). As a proof of concept, overexpressing MPT83 in M. smegmatis showed reduced virulence in a mouse infection model manifesting with reduced bacterial load in the tissues and enhanced mouse survival. Moreover, recent studies have revealed that induction of apoptosis improved the protective effect of vaccines. C57BL/6 vaccination of mice with apoptotic vesicles prepared from mycobacteria-infected macrophages induced a CD8 T cell response and protection against M. tuberculosis infection (25). Furthermore, deletion of secA2 (Rv1821) in an attenuated M. tuberculosis strain, which induced more apoptosis, strongly improved its protective potential when used as a live vaccine (26). More recently, deletion of nuoG, an apoptosis inhibitory gene, also conferred a more protective effect against M. tuberculosis infection (27). Therefore, the deletion of antiapoptotic or introduction of proapoptotic M. tuberculosis genes may be an effective strategy for improving the efficacy of live vaccines. Indeed, the protective effects of MPT83 as protein or a DNA subunit vaccine against M. tuberculosis infection have been reported. Immunization of C57BL/6 mice with MPT83 before aerosol challenging with M. tuberculosis led to a high frequency of Ag-specific IFN-γ T cells, as well as significant reduction of both pulmonary and splenic bacterial loads (21, 28). Our revelation of MPT83 as a novel M. tuberculosis proapoptotic gene provides an explanation for its protective effect, which could also serve as a potential candidate for the development of anti-TB vaccines in the future.

Lipid mediators have been demonstrated to be critical factors involved in modulating the death pattern of M. tuberculosis–infected macrophages. PGE₂ is important for the repair of plasma membrane damage caused by M. tuberculosis infection through promoting translocation of lysosome and Golgi apparatus–derived vesicles, which in turn shift necrosis to apoptosis (9). COX-2, an enzyme responsible for the production of intermediate PGH₂ from arachidonic acid, is critical for the generation of PGE₂ (22). H37Ra-infected macrophages generate PGE₂ through TLR2/p38/COX-2/PGE₂ with an EP4-dependent positive feedback loop (29), whereas H37Rv subverts PGE₂ production through inhibiting COX-2 expression (8). To understand the mechanism of MPT83-induced apoptosis, we detected expression of COX-2 in macrophages stimulated with rMPT83 or infected with M. smegmatis overexpressing MPT83. We observed that MPT83 dramatically increased COX-2 production in mRNA and protein levels indispensable to the TLR2/p38 signaling pathway. Accordingly, both TLR2 defeciencey and COX-2 silencing markedly reduced the MPT83-induced apoptotic response, indicating that a TLR2/p38/COX-2 signaling axis is involved in the regulation of MPT83-induced apoptosis. Unexpectedly, the inhibition or knockdown of COX-2 did not skew M. smegmatis–infected macrophages from apoptosis to necrosis, which may be explained by the fact that M. smegmatis mainly cause apoptosis of macrophages. In support of this idea, inhibition of COX-2 dramatically inhibited cell apoptosis induced by hypoxia/reoxygenation (30), although cells treated with the COX-2 inhibitor NS398 exhibited a dose-dependent increase of apoptosis (31, 32). Therefore, the effect of COX-2 on cell death may be context-dependent. In this study, all the experiments were conducted with purified rMPT83 protein or recombinant M. smegmatis overexpressing MPT83, and the function of MPT83 in the pathogenesis of TB warrants further study by deletion of the gene in M. tuberculosis.

In conclusion, in this study we identified MPT83 as a novel M. tuberculosis gene capable of inducing apoptosis. Moreover, the TLR2/p38/COX-2 axis is responsible for the enhanced apoptosis resulting from MPT83. Our study provides a novel proapoptotic gene of M. tuberculosis, hopefully serving as a candidate gene for future TB vaccine development.

We thank members of B.G.'s laboratory, the Shanghai Key Laboratory of Tuberculosis, and the Clinical and Translational Research Center for helpful discussions.

The authors have no financial conflicts of interest.