M. leprae inhibits apoptosis in THP-1 cells by downregulation of Bad and Bak and upregulation of Mcl-1 gene expression.

BACKGROUND
Virulent Mycobacterium leprae interfere with host defense mechanisms such as cytokine activation and apoptosis. The mitochondrial pathway of apoptosis is regulated by the Bcl-2 family of proteins. Expression of Fas ligand and apoptotic proteins is found in leprosy lesions and M. leprae has been shown to activate pro-apoptotic Bcl-2 genes, Bak and Bax. However, the mechanism by which M. leprae modulates apoptosis is as yet unclear. We investigated expression of apoptotic genes in THP-1 monocytes in response to infection by M. leprae and non-pathogenic M. bovis BCG.


RESULTS
M. leprae did not induce apoptosis in THP-1 cells, while BCG induced a significant loss of cell viability by 18 h post-infection at both (multiplicity of infection) MOI-10 and 20, with an increase by 48 h. BCG-induced cell death was accompanied by characteristic apoptotic DNA laddering in cells. Non-viable BCG had a limited effect on host cell death suggesting that BCG-induced apoptosis was a function of mycobacterial viability. M. leprae also activated lower levels of TNF-alpha secretion and TNF-alpha mRNA expression than BCG. Mycobacterium-induced activation of apoptotic gene expression was determined over a time course of infection. M. leprae reduced Bad and Bak mRNA expression by 18 h post-stimulation, with a further decrease at 48 h. Outcome of cell viability is determined by the ratio between pro- and anti-apoptotic proteins present in the cell. M. leprae infection resulted in downregulation of gene expression ratios, Bad/Bcl-2 mRNA by 39% and Bak/Bcl-2 mRNA by 23%. In contrast, live BCG increased Bad/Bcl-2 mRNA (29 %) but had a negligible effect on Bak/Bcl-2 mRNA. Heat killed BCG induced only a negligible (1-4 %) change in mRNA expression of either Bak/Bcl-2 or Bad/Bcl-2. Additionally, M. leprae upregulated the expression of anti-apoptotic gene Mcl-1 while, BCG downregulated Mcl-1 mRNA.


CONCLUSION
This study proposes an association between mycobacterium-induced apoptosis in THP-1 cells and the regulation of Bcl-2 family of proteins. M. leprae restricts apoptosis in THP-1 cells by downregulation of Bad and Bak and upregulation of Mcl-1 mRNA expression.


Background
Virulent mycobacteria such as Mycobacterium leprae and M. tuberculosis manipulate host cells to persist within them. Apoptosis, or programmed cell death is essential for the homeostatic regulation of cells, restriction of intracellular pathogens and also for stimulation of the host adap-tive immune response. Virulent mycobacteria modulate host cell apoptosis to create a protected niche within cells [1][2][3], and downregulate protective host cytokines such as, TNFα and IFNγ [4,5] to reduce effector T cell and macrophage responses to the pathogen. Virulent M. tuberculosis and M. bovis induce lower levels of apoptosis than avirulent strains [6]. Virulent mycobacteria such as M. leprae induce reduced activation of pro-inflammatory cytokine such as TNFα, as compared with the non-pathogenic M. bovis BCG strain [7]. The apoptotic response to mycobacteria is found to be dependent on cytokine activation [8] whereby, TNFα has been shown to activate mycobacterium-induced apoptosis [9], while IL-10 downregulates apoptosis in macrophages [10].
M. leprae causes a disease spectrum ranging from disseminated multibacillary (MB) leprosy to restricted paucibacillary (PB) disease. T cell and macrophage responses are less affected in PB leprosy than in MB, where effector T cell and macrophage responses are severely compromised. Multibacillary leprosy granulomas show increased expression of Fas ligand which may protect these sites from attack by Fas-expression cytotoxic T lymphocytes [11]. Production of pro-inflammatory cytokines such as TNFα is found to be greater in PB disease as compared with MB disease [12]. Although apoptotic proteins are activated in leprosy lesions from all patients [13], reports indicate that apoptosis is greater in PB disease, correlating with increased TNFα secretion. A role for TNFα is further corroborated as M. leprae induced apoptosis is reduced by treatment with TNFα inhibitor pentoxyfylline [14] or anti-TNFα antibodies [15].
Regulation of programmed cell death occurs via either the intrinsic pathway, involving mitochondrial release of cytochrome C and activation of caspase 9, or the extrinsic pathway, involving the stimulation of death receptors expressed on the cell surface and activation of caspase 8. The Bcl 2-family of proteins comprise of anti-apoptotic members; Bcl-2, Bcl-x L and Mcl-1, and pro-apoptotic members; Bad, Bid, Bax and Bak. Pro-apoptotic Bcl-2 proteins located in the cytosol act as sensors of cellular stress and damage [16]. Upon activation, they relocate to the mitochondrial surface where anti-apoptotic proteins are located. Activation of Bax or Bad and blockade of antiapoptotic Bcl-2 proteins are pivotal steps in the mitochondrial pathway of apoptosis. Bad positively regulates cell apoptosis by forming heterodimers with Bcl-x L and Bcl-2 and reversing their death repressor activity. Pro-apoptotic Bak is sequestered by Mcl-1 and Bcl-x L ; but not Bcl-2 until displaced by Bid, Bad or Bak [17]. When there is an excess of pro apoptotic proteins, cells are more sensitive to apoptosis.
Mycobacteria interfere with host apoptosis by modifying expression of Bcl-2 proteins. Virulent M. tuberculosis has been shown to repress apoptosis by upregulation of Bcl-2 [18] and Mcl-1 [19], and by deactivation of Bad [20] and Bax [2]. Previously, M. leprae has been shown to induce apoptosis in monocytes via induction of Bak and Bax genes [14]. However, the mechanism by which M. leprae influences host cell apoptosis is as yet unclear.
The human acute monocytic leukemia cell line THP-1 has been established as a model to study mycobacterial growth and persistence [21] in addition to mycobacterium-induced apoptosis [22][23][24]. Here we have investigated the mechanism by which virulent M. leprae regulates apoptosis in THP-1 human monocytes as compared with the non-pathogenic M. bovis BCG vaccine strain. The pattern of genomic DNA fragmentation in mycobacterium infected cells was determined in order to investigate the cause of loss in cell viability observed at 18 h postinfection. BCG-infection of cells resulted in DNA laddering characteristic of apoptotic cells, with an increase in apoptosis with dose from MOI-10 to MOI-20 (Fig. 2B). In contrast, apoptotic laddering was not observed in cells infected with M. leprae (data not shown).

Mycobacterium leprae induces less cell death as
We employed gamma-irradiated M. leprae in our assays as this was the only strain available to us and we have previously used it to model M. leprae infections in cells from both healthy controls and leprosy patients [25,26]. The M. bovis BCG strain served as a comparison of a non-pathogenic species, while heat-killed BCG was employed as a control for BCG viability in the assays. Heat-killed BCG at the lower infection dose of MOI-10 did not cause any significant loss in THP-1 viability but did so at MOI-20 (P < 0.05) at 18 h. However, cell death was not induced by heat-killed BCG at 48 h post-infection (data not shown).

M. leprae is a poorer inducer of TNFα than M. bovis BCG
A role for TNFα is indicated in mycobacterium-induced apoptosis. M. leprae-induced apoptosis in primary macrophages has been shown to be dependent on TNFα activation [14], while BCG-induced apoptosis in THP-1 cells is dependent on an endogenous TNF response [24]. Virulent mycobacteria are poorer inducers of TNFα than avirulent strains [27,28]. M. leprae has previously been shown to induce only negligible TNFα secretion in THP-1 cells [29]. We investigated M. leprae induced TNFα secretion and mRNA expression in THP-1 cells to determine whether the low level of apoptosis observed could be associated with a lack of TNFα activation in the assay. M. leprae and BCG -induced cytokine secretion was measured at 6, 18 and 48 h post-stimulation. BCG infection of cells induced TNFα secretion within 6 h of stimulation (mean values: BCG10, 40 pg/ml; BCG20, 80 pg/ml) with peak secretion at 18 h (mean values: BCG10, 238 pg/ml; BCG20, 299 pg/ml) and a decrease at 48 h (mean values: BCG10, 165 pg/ml; BCG20, 204 pg/ml). M. leprae did not induce any detectable TNFα secretion in THP-1 cells at either MOI 10 or 20 at 6 h, with negligible secretion (ML10, 0 pg/ml; ML20, 4 pg/ml) at the 18 h peak interval. However, M. leprae induced TNFα mRNA expression upon infection of cells but at a lower level than as compared with BCG (Fig 3).

Bak and Bad gene expression is down-regulated in M. leprae infected THP-1 cells
M. leprae infection of peripheral blood monocytes has been shown to result in the activation of the pro-apoptotic protein Bad and induction of Bak mRNA [14]. In this study we did not observe apoptosis in THP-1 cells infected with M. leprae, while cells infected with BCG showed increased apoptosis. To investigate the mechanism responsible for this difference we measured Mycobacterium-induced mRNA expression of pro-apoptotic (Bad and Bak) and anti-apoptotic (Bcl-2) genes at 6, 18 and 48 h post infection. Gene expression was compared relative to that of the β-actin housekeeping gene in each case. Outcome of cellular viability/apoptosis is determined not only by the expression of specific pro-and anti-apoptotic genes, but is dependent on the ratio between pro and anti apoptotic genes. Therefore, to further understand Mycobacterium-induced change in host apoptotic gene expres- has been shown to be optimal for real-time PCR quantification of gene expression in human cells [30]. At the ear-  (Fig. 6).

Discussion
Our study shows that M. leprae inhibits apoptosis in the THP-1 monocytic cell line by downregulation of proapoptotic genes Bak and Bad and concomitant upregulation of anti-apoptotic Mcl-1 mRNA. In contrast, the increased apoptosis evidenced by BCG infection of THP-1 cells is coordinated by an increase in Bad mRNA expression relative to Bcl-2 and a downregulation of Mcl-1 mRNA.
Our data correlates with previous reports that virulent mycobacteria induce lower levels of cell death than avirulent mycobacteria [6]. Cell death induced by BCG was dose dependent as shown previously [24]. BCG-induced apoptosis (as shown by DNA laddering of cells) increased with time from 18 to 48 h post-infection. Loss in cell viability was observed in response to heat-killed BCG at 18 h but was no longer apparent by 48 h. The earlier response to non-viable BCG observed at 18 h may be the result of non-specific binding and entry of the mycobacterium into the host cell which may trigger host defenses. Initial cell attachment and uptake of BCG into host macrophages has shown to be common amongst live and dead mycobacterium, with subsequent differences occurring later due to differential trafficking of live and dead organisms within host the endosomal/lysosomal pathway [31]. As BCG-induced apoptosis is most probably dependent on bacterial viability it is therefore only sustained in the case of live BCG.
M. leprae-induced less TNFα secretion and gene expression than BCG did. A role for TNFα has been defined previously in BCG-induced apoptosis in THP-1 cells [24]. In addition, M. leprae-induced apoptosis in primary monocytes has been shown to be TNFα dependent [14], and inhibition of TNFα via treatment with anti-TNFα inhibitors or pentoxyfylline is found to reduce host inflammation and cell death in reactional leprosy [32]. Therefore, in the case of M. leprae infection, the reduced levels of TNFα Mycobacterium-induced TNFα, Bcl-2, Bad and Bak gene expression Figure 3 Mycobacterium-induced TNFα, Bcl-2, Bad and Bak gene expression. THP-1 cells were infected with M. leprae and BCG as described previously and total cellular RNA was extracted, reverse transcribed and PCR was carried out for β-actin, TNFα, Bcl-2, Bad and Bak genes. Gene expression was quantified using scanning densitometry. Panels illustrate gene expression in unstimulated cells 'sp', M. leprae 'ML10', 'ML20', and BCG 'BCG10' and 'BCG20' infected cells at 18 h post-stimulation in a representative experiment. Graphs illustrate expression of TNFα, Bcl-2, Bad and Bak mRNA relative to β-actin in each case. We found M. leprae to downregulate expression of proapoptotic genes Bak and Bad relative to both β-actin and anti-apoptotic Bcl-2. This correlates with previous studies where reduced Bax and increased Bcl-2 mRNA expression has been shown to be responsible for slowly progressive murine tuberculosis [33]. The role of Bad gene expression in determining outcome of cellular viability is further indicated where cell wall lipoarabinomannan (LAM) from M. tuberculosis-induces survival in monocytes by phosphorylation of the protein Bad, hence preventing it from binding Bcl-2 [22]. However, our work differs from that by Hernandez, et al. who demonstrate that infection of primary monocytes with irradiated M. leprae resulted in apoptosis due to activation of apoptotic genes, Bax and Bak [14]. This apparent discrepancy can be attributed to differential outcome of mycobacterial infection dependent on host cell type, as evidenced by the variation in apoptosis induction by M. tuberculosis observed in alveolar epithelial cells and U937 macrophages [20]. In addition, Hernandez et al [14] showed that M. leprae increases apoptosis in monocytes-derived macrophages of leprosy patients. However, macrophages from leprosy patients are more prone to apoptotic death than those from healthy individuals [15]. Therefore, the spontaneous activation of apoptotic genes in leprosy patients may make the cells more prone to extracellular triggers such as infection with mycobacteria.
We demonstrate a role for Mcl-1 in M. leprae mediated inhibition of apoptosis in the THP-1 cells. The anti-apoptotic Bcl-2 family member gene Mcl-1 has been shown to play a role in cell survival of human polymorphonuclear cells [34] and neutrophils [35]. Virulent M. tuberculosis H37Rv has been shown to persist in THP-1 cells by upregulation of the anti-apoptotic Bcl-2 family member gene Mcl-1 [19]. Pro-apoptotic Bak is sequestered by Mcl-1 [17] therefore upregulation of Mcl-1 expression would result in a coordinate decrease in available Bak leading to further downregulation of apoptosis.
We found BCG-induced apoptosis to be accompanied with increased expression of Bad/Bcl-2 mRNA. Previously, BCG-induced apoptosis has been shown to be a consequence of downregulated Bcl-2 expression [1] and also by activation of caspases [24]. We show for the first time in our study that BCG induces downregulation of anti-apoptotic Mcl-1 mRNA. Overall, our results indicate that the differential effect of virulent M. leprae and non-pathogenic BCG on host cellular survival is determined by the combined gene expression of pro-and anti-apoptotic genes Bad, Bak, Bcl-2 and Mcl-1.  D. E. F.

Materials and methods
Mycobacterial strains used M. bovis BCG (Montreal vaccine strain) was used as described previously [7]. Gamma-irradiated M. leprae (prepared from armadillo liver tissue) was provided by Colorado State University, USA, by NIH contract (NO1-AI-55262, 'Leprosy research support'). BCG was were cultured in 7H9 Middlebrook medium supplemented with 0.02 % glycerol, 10 % ADC Middlebrook enrichment and 0.5 % Tween-80 (DIFCO Laboratories, Detroit, MI, USA). BCG was heat killed by treatment at 80°C for 30 min. Mycobacterium sp. was stored in single use aliquots at -70°C and used as described previously [25].

ELISA for TNFα
ELISA reagents for TNFα were from R&D Systems (USA). Assays were carried out according to the manufacturer's recommendation and as reported previously. [7] The lower limit of detection was 3.9 pg/ml.

Fluorescence viability staining
For fluorescence assays, cells were seeded on glass coverslips within 24 well plates as described above. Staining dye cocktail contained acridine orange (100 µg/ml) and ethidium bromide (100 µg/ml) in PBS [36]. One ml of dye cocktail was added to each coverslip, incubated in the dark for 2 min and subsequently visualized by microscopy.
Apoptotic DNA laddering THP-1 cells (2 × 10 5 /sample) were lysed, and the genomic DNA extracted according to the protocol for TACS DNA Laddering Kit (R&D Systems, USA). Samples were analyzed on a 1.5 % agarose gel with ethidium bromide staining.

Gene expression quantification for βactin, Bcl-2, Bak and Bad
Total cellular RNA was harvested from adherent cells using Trizol reagent (GIBCO-BRL, USA). cDNA was prepared by reverse transcription using MuLV reverse transcriptase and PCR for the β-actin gene was carried out using custom designed primers as described previously [25]. PCR primers for Bcl-2, Bak and Bad were [14,20]