Introduction

Mycoplasmas, the smallest self-replicating wall-less bacteria, exist as parasites in humans, and were shown to interact with various cells of the immune system, either by activation or suppression, resulting in alteration of cell functions.1 Mycoplasma fermentans (M. fermentans), a human pathogen, was shown to induce polyclonal activation of the immune cells, cytokine production, increased major histocompatibility complex (MHC) class II expression, increased cytotoxicity of T cells and expression of oncogenes.2, 3, 4, 5, 6 In the last decade, there have been increasing numbers of publications implicating M. fermentans in the development of rheumatoid arthritis (RA),7, 8, 9 a disease characterized by impaired apoptosis.8, 10, 11, 12

Apoptosis is induced in two different pathways: the intrinsic and extrinsic pathways. tumor necrosis factor α (TNFα), an inducer of the extrinsic pathway, crosslinks the TNF receptor (TNFR), which results in the activation of caspases, a class of cysteine proteases with specific functions in the execution of apoptosis.13 As a result of caspase activation, certain cellular substrates are cleaved and the cells undergo apoptosis.14

Several studies were published regarding mycoplasmas and apoptosis, but the results were contradictory. Mycoplasma proteins were shown to induce apoptosis in several cells,5, 15 and it was shown that in Mycoplasma-contaminated cell cultures the internucleosomal DNA degradation is due to mycoplasmal endonuclease activity.16 Another study showed that mycoplasmal infections do not induce apoptosis, and even prevent it.17 In the present study, the indirect effect of mycoplasmal infection on apoptosis was investigated, by examining the effect of M. fermentans infection on apoptosis induced by TNFα, in the human myelomonocytic U937 cell line, as an experimental model.

Results

Infection of U937 cells with M. fermentans did not induce apoptosis in the cells

To investigate whether M. fermentans infection of U937 cells induces direct apoptosis of the cells, the viability of the cells was examined at 24 and 48 h post infection. The FACSCalibur flow cytometer (FACS) analysis of Annexin-V-fluorescein isothiocyanate (FITC)- and propidium iodide (PI)-stained cells (Figure 1a) and cell cycle (Figure 2) demonstrated no difference in the percent of apoptotic cells between noninfected and infected cells. Thus, under the conditions used, infection of U937 cells with M. fermentans did not induce apoptosis.

Figure 1
figure 1

(a) Cell viability in M. fermentans-infected and noninfected U937 cells determined by Annexin-V-FITC–PI staining as analyzed by FACS. U937 cells were infected for 24 h (n=9) or 48 h (n=3) prior to cell staining. The mean percentage of live cells±standard deviation were as follows: 91±4% (at 24 h) and 92±1% (at 48 h), for noninfected cells (black); 92±3% (at 24 h) and 91±4% (at 48 h), for control culture treated with SP4 (M. fermentans medium) (white); and 88±6% (at 24 h) and 89±4% (at 48 h), for cells infected with M. fermentans (1000 CFU/cell) (gray). (b) Number of U937 cells in M. fermentans-infected and noninfected cultures, determined by trypan blue exclusion dye. U937 cells were infected for 24 h prior to cell staining (at a ratio of 1000 CFU/cells). The mean percentage of cell number × 105/ml±standard deviation was as follows: 5.2±1.7 in noninfected cells; 4.5±1 in SP4 control; 4.2±1 in M. fermentans-infected cells (1000 CFU/cell). *P<0.05, **P – not significant

Full size image
Figure 2
figure 2

Cell cycle in M. fermentans-infected and noninfected U937 cells, determined by PI staining as analyzed by FACS. U937 cells were infected with M. fermentans (1000 CFU/cell) for 24 h prior to cell staining. The data are presented as the mean percentage±standard deviation of cells at different phases of the cell cycle, from three independent experiments. M. fermentans caused a G1 arrest

Full size image

Infection of U937 cells with M. fermentans causes a cell cycle arrest

To investigate whether M. fermentans infection of U937 cells induces proliferation of the cells, we determined, at 24 h post infection, the number and viability of the cells by trypan blue exclusion dye, and analyzed the cell cycle. As seen in Figure 1a, there was no significant difference in percentage of viable cells between the infected and noninfected cultures. By cell cycle analysis, a small but statistically significant decrease in the percentage of cells in S phase was observed (P<0.05 for both controls), as well as a corresponding increase in the percentage of cells in G1 phase (P<0.05 for both controls) (Figure 2).

Confocal microscopy of U937 cells infected with M. fermentans

To determine the localization of M. fermentans with regard to U937 cells, we examined the cells by confocal microscopy. As seen in Figure 3, at 12 and 24 h post infection, most of M. fermentans were localized at the surface of U937 cells. Some M. fermentans were observed inside the cells.

Figure 3
figure 3

Confocal microscopy of M. fermentans-infected and noninfected U937 cells, at various infection times. Slides were prepared as described in Materials and methods. (a) Morphology of infected and noninfected cells. (b) The location of M. fermentans as detected with anti-M. fermentans antibodies

Full size image

TNFα-induced apoptosis is reduced in U937 cells infected with M. fermentans

Prior to investigating TNFα-induced apoptosis, the levels of TNFα secreted by U937 upon infection with M. fermentans were measured in the supernatants of the cultures, 24 h post infection. As shown in Figure 4, M. fermentans induced some secretion of TNFα, in a dose–response manner depending on the infection load (P<0.05 for an infection ratio of 1000/1), but the amount secreted even at the highest colony-forming units (CFU)/cell ratio (the ratio we used in all experiments) is negligible (20 pg/ml) as compared to the amount we used to induce apoptosis (20 ng/ml). Indeed, as shown in Figure 1a, the infected cells, after 24 h, are not undergoing apoptosis.

Figure 4
figure 4

TNFα secretion by M. fermentans-infected and noninfected U937 cells, determined by ELISA. U937 cells were infected with M. fermentans for 24 h, and TNFα in the supernatants was measured. Mean percentage±standard deviation of TNFα levels in the different cell cultures, from four separate experiments, are depicted

Full size image

The effect of M. fermentans on TNFα-induced apoptosis (20 ng/ml) in U937 cells was determined 24 h post infection. The rationale for choosing 24 h after infection emerged from the results obtained both in the confocal microscopy and in electron microscopy (data not shown). The percentage of apoptotic cells was examined by two techniques: (1) Acridine orange (AO)–ethidium bromide (EB) staining, which distinguishes between apoptotic and necrotic cells by the morphological changes in the nucleus (lack of DNA condensation in necrotic cells as opposed to apoptotic cells),18 was used to determine whether TNFα addition in our system induced apoptosis or necrosis. Firstly, TNFα addition induced apoptosis and not necrosis. Secondly, the apoptosis was reduced by ∼60% in U937 cells infected with M. fermentans (11.9±4.7%), in comparison to noninfected cells (27.5±7.7%) and SP4-treated cells (27.2±3.9%); P<0.05 for both controls (Table 1); (2) A typical experiment of Annexin-V-FITC–PI staining is shown in Figure 5a. By this technique, necrotic and late apoptotic cells cannot be distinguished from each other, but, relying on the AO–EB technique, we concluded that the double-positive cells (for Annexin-V-FITC and PI) represent late apoptotic cells and not necrotic ones. The results, obtained in nine independent experiments, showed approximately 60% reduction in the percentage of apoptotic U937 cells infected with M. fermentans at a ratio of 1000 CFU/cell (6.5±2.8%), in comparison to noninfected cells (19.8±4.3%) and SP4 control (19.5±4.1%); P<0.01 for both controls (Figure 5b). The specificity of the inhibitory effect of M. fermentans is demonstrated by the magnitude of inhibition, which was dependent on the ratio of CFU/cell (Figure 5b).

Table 1 Cell death induced by TNFα in M. fermentans-infected and noninfected U937 cells, as determined by fluorescence microscopy
Full size table
Figure 5
figure 5

Measurement of TNFα-induced apoptosis by Annexin-V-FITC–PI staining in M. fermentans-infected and noninfected U937 cells. (a) A representative experiment: Cells were infected for 24 h (1000 CFU/cell), followed by addition of TNFα (20 ng/ml) for 4 h, stained with Annexin-V-FITC-PI and analyzed by FACS. Dot plots of forward and side scattered (with the gate used) are depicted on the upper-left corner of each treatment. Double negative staining represents live cells (LL), positive staining for Annexin-V-FITC and negative staining for PI represents early apoptotic stage (LR), and double-positive staining represents late apoptotic stage (UR). (b) Mean percentages of apoptotic cells±standard deviation were: 19.8±4.3% in noninfected cells, 19.5±4.1 in SP4 control, and 13.4±4.7, 11.2±2 and 6.5±2.8% in M. fermentans-infected cells (50, 200 and 1000 CFU/cell, respectively) (P<0.05, P<0.01 and P<0.01 in M. fermentans-infected cells (50, 200 and 1000 CFU/cell, respectively) for both controls as determined by ANOVA single factor)

Full size image

Loss of mitochondrial inner transmembrane potential induced by TNFα is reduced in U937 cells infected with M. fermentans

In many apoptosis scenarios, including TNFα-mediated apoptosis, the mitochondrial inner transmembrane potential (ΔΨm) collapses.19, 20 To investigate whether the antiapoptotic effect of M. fermentans in TNFα-induced apoptosis is upstream or downstream of the mitochondria, we measured the loss in ΔΨm, induced by TNFα (20 ng/ml), in infected and noninfected cells. At 24 h post infection, the cultures were stimulated with TNFα (20 ng/ml) for 2 h, and each culture was stained with 3,3′-dihexyloxacarbocyanine iodide (DiOC6 (3)) and analyzed by FACS (a typical experiment is shown in Figure 6a).

Figure 6
figure 6

Determination of decrease in mitochondrial inner transmembrane potential induced by TNFα in M. fermentans-infected and noninfected U937 cells, by FACS of DiOC6(3) fluorescence. (a) A representative experiment: Cells were infected for 24 h, followed by addition of TNFα (20 ng/ml) for 2 h, stained with DiOC6(3) and analyzed by FACS. The percentage of cells with higher DiOC6(3) staining (higher ΔΨm) is on the right of each histogram, while the percentage of cells with lower DiOC6(3) staining (lower ΔΨm) is on the left. The line in the control histogram represents the positive control (mCiCCP). (b) Mean percentages of low DiOC6(3) fluorescence dye±standard deviation (n=3) were: 23.7±7.2% in noninfected cells, 22.6±2.2% in SP4 control and 5.6±1.6% in M. fermentans-infected cells (1000 CFU/cell) (P⩽0.01 for both controls, as determined by ANOVA single factor). Overall, a fourfold reduction of ΔΨm loss was measured in M. fermentans-infected U937 cells

Full size image

In three independent experiments we found that, in U937 cells infected with M. fermentans, there was a fourfold reduction in ΔΨm loss, 5.6±1.6% cells with low DiOC6(3) fluorescence dye, compared to noninfected cells (23.7±7.2%) and SP4-treated cells (22.6±2.2%); P≤0.01 for both controls (Figure 6b).

Protease activity of caspase-8 induced by TNFα is reduced in U937 cells infected with M. fermentans

Caspase-8 is at the apex of the caspase pathway and links death domain protein signaling (on TNF receptor) and caspase activation.14 To investigate whether the reduction of TNFα-induced apoptosis in infected cells is upstream or downstream of caspase-8, we measured the protease activity of caspase-8 induced by TNFα, in infected and noninfected cells. At 24 h post infection, the cultures were stimulated with TNFα (20 ng/ml) for 4 h and cytosolic extracts were prepared. The assay for caspase-8 activity was performed as described in Materials and Methods.

A 60% reduction in protease activity of caspase-8 was measured in M. fermentans-infected U937 cells, compared to noninfected cells; P<0.05 (Figure 7).

Figure 7
figure 7

Protease activity of caspase-8 induced by TNFα in M. fermentans-infected and noninfected U937 cells. Cells were infected for 24 h, followed by addition of TNFα (20 ng/ml) for 4 h, and cytosolic extracts were prepared. Mean percentages of caspase-8 activity±standard deviation (n=3) were: 1.87±0.3-fold in TNFα-treated compared to untreated in noninfected cells, and 1.17±0.29-fold in TNFα-treated compared to untreated in M. fermentans-infected cells (1000 CFU/cell) (P<0.05 as determined by ANOVA single factor). The decrease in protease activity of caspase-8 caused by M. fermentans was ∼60%

Full size image

TNFα-induced apoptosis is reduced in U937 cells treated with non-live M. fermentans

To clarify whether the inhibitory effect of M. fermentans on TNFα-induced apoptosis was exerted by live or non-live mycoplasma, cells were stimulated with sonicated M. fermentans prior to induction of apoptosis by TNFα. The percentage of apoptotic cells was examined by Annexin-V-FITC–PI staining. The results, obtained in four independent experiments, showed an ∼60% reduction of apoptotic cells in U937 cells stimulated with 40 μg/ml of sonicated M. fermentans (9.6±3.8%), in comparison to noninfected cells (23.7±4.7%) and phenylmethylsulfonyl fluoride (PMSF) control cells (26.3±3.8%); P<0.01 for both controls (Figure 8). The inhibitory effect of sonicated M. fermentans on TNFα-induced apoptosis was dose dependent (Figure 8).

Figure 8
figure 8

Measurement of TNFα-induced apoptosis by Annexin-V-FITC–PI staining in U937 cells treated with non-live sonicated M. fermentans. Cells were treated with non-live sonicated M. fermentans for 24 h, followed by addition of TNFα (20 ng/ml) for 4 h, stained with Annexin-V-FITC/PI and analyzed by FACS. Mean percentages of apoptotic cells±standard deviation were: 23.7±4.7% in noninfected cells, 26.3±3.8 in PMSF control, and 20.8±4.2%, 15.3±2 and 9.6±3.8% in cells treated with the sonicated M. fermentans (2, 8 and 40 μg/ml respectively) (*P<0.05, **P<0.01 and ***P<0.01 for both controls as determined by ANOVA single factor)

Full size image

Discussion

In this article, we report that infection of human myelomonocytic U937 cell line with M. fermentans does not cause a direct cell death but, rather, inhibits TNFα-induced apoptosis. The block of apoptosis lies upstream of the mitochondria and upstream of caspase-8. It seems that the inhibitory effect of M. fermentans resides in constitutive components, since both live and non-live M. fermentans exert a similar effect.

M. fermentans has been shown to be both an extracellular and intracellular bacteria.7, 21 We found, by confocal microscopy, that most M. fermentans cells were localized on the surface of U937 cells and that some of them became intracellular during the 8–24 h post infection (Figure 3). These findings suggest that M. fermentans might affect the cell either from the outer side of the membrane, or by stimulating the cell from its location within the cell.

The effect of mycoplasmas on apoptosis was previously examined in several studies.5, 15, 16, 17 Most of the studies were engaged with the direct effect of Mycoplasma or mycoplasmal components on the apoptotic processes in infected cells. M. fermentans proteins were shown to induce apoptosis in various cell lines.5, 15 In contrast, it was shown that mycoplasmal infections (under similar conditions, for example, multiplicity of infection (MOI)) do not induce apoptosis, and even prevent it.17 In the present study, we found that infection of U937 cells with M. fermentans for 24 or 48 h did not induce apoptosis (Figures 1a and 2), and induced an incomplete arrest in the cell cycle at G1 (Figure 2), as was previously reported.6 The contradiction between the studies might be explained by the fact that non-live mycoplasmal derivatives were used when apoptosis was demonstrated,5, 15 whereas, when live mycoplasmas were employed, apoptosis was not induced.17 In another study, it was reported that Mycoplasma contamination of astrocytes (using similar MOI) could cause apoptosis, as a result of choline depletion in the medium, due to the mycoplasmal nutritional requirements rather than a direct effect of Mycoplasma on the cells.22 Indeed, Rawadi et al5 reported that the proapoptotic effect resides exclusively in the nonlipid protein fraction of M. fermentans and not in the lipid fraction; this might explain why live Mycoplasma (enveloped only by a lipid membrane) do not cause apoptosis or cell death. In our system, U937 cells were infected with live M. fermentans. Although we were not able to grow M. fermentans from the culture 24 h post infection (data not shown), it is likely that U937 cells came into contact with M. fermentans membrane lipoproteins. This might explain why, in our system, there was no apoptosis during the 24 or 48 h post infection.

To our knowledge, this is the first report presenting the ability of Mycoplasma to inhibit TNFα-induced apoptosis. Only a few papers have been published regarding inhibition of TNFα-induced apoptosis by Chlamydia pneumonia23, 24 and Hepatitis C Virus.25 These support the hypothesis26 that parasite bacteria might use a mechanism to prevent host cell apoptosis, in order to promote their survival and replication.

In TNFα-mediated apoptosis, the ΔΨm collapses, resulting in cytochrome c release.19, 20 A fourfold reduction in ΔΨm loss was observed in infected cells as compared to noninfected cells (Figure 6). Therefore, we concluded that the effect of M. fermentans is upstream of the mitochondria. This result is similar to a recent observation regarding the effect of Chlamydia pneumoniae infection,23, 24 which was reported to render epithelial and monocyte cell lines resistant to TNFα-induced apoptosis, via blockage of mitochondrial cytochrome c release.

It is well known that the Bcl-2 protein family regulates the ΔΨm loss in apoptosis,27, 28 but, as yet, we have no evidence that these proteins are upregulated or downregulated in our infected cells (data not shown).

Binding of TNFα to the TNFR1 causes fas-associated death domain (FADD) to bind the receptor. Procaspase-8 binds the receptor-bound FADD, leading to its proteolytic activation. This, in turn, can lead to an apoptotic process via two branches, either directly by activation of caspase-3 (extrinsic), or by release of cytochrome c from the mitochondria (intrinsic).13 By comparing caspase-8 activity (induced by TNFα) in infected and noninfected cells, we found that there was a 60% reduction of protease activity of caspase-8 in infected cells, indicating that the inhibitory effect of M. fermentans on TNFα-induced apoptosis is upstream of caspase-8 (Figure 7).

Our finding that M. fermentans inhibit TNFα-induced apoptosis might be explained by the effect of M. fermentans on the transcription factor, nuclear factor κB (NF-κB).NF-κB activation is known to inhibit TNFα-induced apoptosis via suppression of caspase-8 activation.29 It was previously reported that M. fermentans cause NF-κB activation in human monocytic cell line and murine macrophages.5, 17, 30 In a preliminary study, we found that, in our system, NF-κB was translocated to the nucleus in M. fermentans-infected cells (data not shown).

Alternatively, since most of M. fermentans were associated with the U937 cell membrane (Figure 3), the notion that the antiapoptotic effect of M. fermentans is exerted via modulation of TNF receptors has surfaced. This is currently under investigation in our laboratory, and preliminary results indicate that there is no significant difference in TNFR1 expression between infected and noninfected cells (data not shown).

The inhibition of TNFα-induced apoptosis was also evident when cells were stimulated by non-live M. fermentans, similar to the effect of infection with live M. fermentans (Figure 8). Although we infected cells with live M. fermentans, we cannot conclude whether the inhibition of TNFα-induced apoptosis was due to live or non-live Mycoplasma, since live M. fermentans could not be cultivated after 24 h. However, in preliminary experiments, when live M. fermentans were recovered from the culture, 2 h post infection, the antiapoptotic effect was already demonstrated (data not shown). These results suggest that both live and non-live M. fermentans exerted inhibition of the TNFα-induced apoptosis.

In summary, the findings that M. fermentans causes a significant inhibition of TNFα-induced apoptosis in U937 cells strengthen the hypothesis that M. fermentans might play a pathogenic role in the development of diseases characterized by impaired apoptosis. These findings also imply that the presence of mycoplasmas in various cell lines might be affecting the results of apoptosis research.31

Materials and Methods

Reagents and antibodies

Recombinant TNFα and TNFα determination kit (enzyme-linked immunosorbent assay (ELISA)) were purchased from R&D Systems, Inc. (Minneapolis, MN, USA). Annexin-V-FITC and PI were purchased from Bender MedSystems (Vienna, Austria). The 3,3′-dihexyloxacarbocyanine iodide (DiOC6(3)) and goat-anti-mouse conjugated to Alexa-Fluor 488 were purchased from Molecular Probes, Inc. (Junction City, OR, USA). Carbonyl cyanide m-chlorophenylhydrazone (mCiCCP), EB, AO and PMSF were purchased from Sigma Chemical Co. (St. Louis, MO, USA). The caspase-8 Assay Kit was obtained from Calbiochem (Nottingham, UK). Cell culture media and supplements were purchased from Biological Industries Ltd (Beit Haemek, Israel).

Mycoplasma culture and mycoplasmal non-live preparation

M. fermentans K7 (originally obtained from JG Tully (National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD, USA)) was cultured in SP4 broth containing 8.5% fetal calf serum (FCS).32, 33 Aliquots (1 ml) (containing 4 × 108 CFU) were frozen at −70°C. For each experiment, one aliquot was thawed and cultivated in SP4 broth first, at a dilution of 1/10 for 24 h (37°C), and then stepped up at a 1/50 dilution for another 24 h, when log phase was observed. M. fermentans was quantified by counting of CFU on SP4 agar plates.32, 33

Non-live Mycoplasmas were prepared as follows: M. fermentans was cultivated as above and the 500 ml culture was pelleted (13 000 × g, 30 min at 4°C), and washed twice with phosphate-buffered saline (PBS). The pellet was then resuspended in 2 ml of PBS. CFU/ml was determined and then stored at −70°C for further experiments. Frozen pellets of M. fermentans were thawed and sonicated at 4°C for 4 × 30 s at 80% power (50% duty cycle; Heat Systems Ultrasonics, Inc.) in the presence of the protease inhibitor PMSF (10−4 M). These conditions of sonication resulted in a non-live mycoplasmal preparation (no growth was observed in repeated culturing procedures). Protein concentration of sonicated M. fermentans was determined by using the Bio-Rad protein assay kit (Richmond, CA, USA). An amount of 1 × 108 CFU corresponds to 10 μg of mycoplasmal proteins.

Cell culture

The human myelomonocytic U937 cell line was cultivated at 37°C and 5% CO2, in RPMI 1640 culture medium containing 10% FCS, 1% HEPES, 1% penicillin–streptomycin and 1% glutamine. The cell line was tested every 4 weeks by a polymerase chain reaction (PCR)-based detection assay for M. fermentans contamination.9

Stimulation of U937 cells with live M. fermentans or sonicated M. fermentans

Prior to infection with live M. fermentans, Mycoplasmas (at a log phase) were washed once in PBS (13 000 × g, 30 min at 4°C), and resuspended in RPMI 1640 medium containing 10% FCS, 1% HEPES, 1% penicillin–streptomycin and 1% glutamine. Then, M. fermentans was added to U937 cells, pre-washed once with PBS (500 × g, 10 min at 4°C), and resuspended in a new culture medium at a final concentration of 2 × 105 cells/ml. The ratio of Mycoplasmas/cell to be used was pre-determined by a series of dose–response experiments, at various ratios (50/1, 200/1, 1000/1). A ratio of 1000 CFU/cell was chosen since, under these conditions, the effect of M. fermentans on cells was more pronounced. The same procedure, with the same volume as in the M. fermentans culture, was performed with SP4 broth to be used as on the SP4 control. For experiments with non-live M. fermentans, the sonicated Mycoplasma (prepared as described above) was added to U937 (2 × 105 cells/ml) cells at concentrations of 2, 8 and 40 μg/ml. PMSF control is a noninfected cell culture with the addition of PMSF, at the same concentration as in the sonicated M. fermentans.

Cell cycle analysis

At 24 h post infection, U937 cultures were washed and fixed with 1 ml of 70% cold ethanol. Cells were kept in the fixative for at least 24 h at −20°C. Followed by centrifugation and removal of the ethanol, pellets were resuspended in 900 μl of a solution containing 0.1% Triton X-100 and RNase (10 μg/ml) in PBS. The cells were kept for 40 min at room temperature, and stained with 100 μl of PBS containing PI (150 μg/ml). After incubation on ice for 10 min, the cells were analyzed by a FACS (Becton-Dickinson, Mansfield, MA, USA). The population of cells in each cell cycle phase was determined by using ModFit LT software (Becton Dickinson Immunocytometry Systems).

Confocal microscopy

U937 cells were infected with M. fermentans, as described above, for different time periods (0, 1, 2, 4, 12 and 24 h). Each culture was fixed with 3.7% formaldehyde (in Eppendorf tubes) for 30 min at room temperature and washed twice with PBS. Cells were then permeabilized by 0.2% Triton X-100 for 5 min at room temperature, washed twice with PBS, and blocked with 1% bovine serum albumin (BSA). Cells were then incubated with anti-M. fermentans hyperimmune mouse serum, prepared as described elsewhere,34 for 40 min at room temperature. Then, the cells were washed three times with PBS, incubated with goat-anti-mouse conjugated Alexa-Fluor 488 for 40 min at room temperature and washed three times with PBS. The labeled cells were mounted on slides and examined by confocal laser-scanning microscopy (LSM410, Carl Zeiss, Jena, Germany).

Measurement of TNFα secreted from U937

At 24 h post infection of U937 cells with M. fermentans, as described above, the supernatants were collected by centrifugation (500 × g for 10 min) and stored at −20°C. TNFα ELISA was performed according to the manufacturer's instructions.

Induction of apoptosis

Infected U937 cells (24 h post infection) were counted, examined for viability by the trypan blue exclusion dye, centrifuged at 500 × g for 10 min, and transferred to a new culture medium at a final concentration of 4 × 105 cells/ml. The cells were divided in 24-well sterile plates (Corning, Corning, NY), and apoptosis was induced by TNFα at a concentration of 20 ng/ml.

Measurement of apoptosis

Two different techniques were used: (1) AO–EB staining, an exclusion dye method, enables the differentiation between live, early-apoptotic, late-apoptotic and necrotic cells.18 At 8 h post induction of apoptosis by TNFα, infected and noninfected cells (4 × 105) were collected, centrifuged at 500 × g for 10 min, and resuspended in 100 μl PBS. Samples of 25 μl from each culture were stained with AO–EB (final concentrations 1 μg/ml for each AO and EB), and observed under a fluorescence microscope. At least 200 cells were randomly counted in each sample (in duplicates), and the percentage of apoptotic cells was calculated; (2) Annexin-V-FITC–PI staining,35 which detects the exposure of phosphatidylserine (PS) to the external leaflet of the plasma membrane in early apoptosis, enables to differentiate between live, early apoptotic and dead cells (it does not differentiate between late apoptotic and necrotic cells).36 At 4 h post induction of apoptosis by TNFα, infected and noninfected cells (2 × 105) were washed with PBS, centrifuged and resuspended in 400 μl of a binding buffer (10 mM HEPES/NaOH, 140 mM NaCl, 2.5 mM CaCl2, pH 7.4). Annexin-V-FITC (5 μl; 10 μg/ml) was added to a sample of 195 μl of cell suspension, mixed, incubated for 10 min at room temperature in the dark, washed with PBS and resuspended in 190 μl of binding buffer containing 10 μl of PI (1 μg/ml). The double-stained cells were analyzed by the FACS, within 10 min (10 000 cells/sample).

In all measurements of apoptosis, we determined the percent of cells that underwent apoptosis by TNFα, by subtracting the percent of spontaneous apoptosis (unstimulated cells) from the total apoptosis (stimulated cells).

Measurement of mitochondrial inner transmembrane potential (ΔΨm)

At 2 h post induction of apoptosis by TNFα, a total of 8 × 105 infected and noninfected cells were washed with PBS and resuspended in 990 μl of PBS containing 10 μl of DiOC6(3) (final concentration 40 nM). Cells were mixed, incubated for 30 min at 37°C in the dark, and kept on ice until analyzed by the FACS within 10 min (10 000 cells per sample). Cells with low DiOC6(3) fluorescence represent cells with permeabilized mitochondria. A positive control, comprising 8 × 105 cells with 5 mM of mCiCCP, a mitochondrial uncoupler and a known reducer of ΔΨm,37 was incubated for 30 min at 37°C prior to the addition of DiOC6(3).

Measurement of protease activity of caspase-8

At 4 h post induction of apoptosis by TNFα, a total of 1.6 × 106 infected and noninfected cells were washed with PBS, resuspended in 80 μl of ice-cold cell lysis buffer (50 mM HEPES, 1 mM DTT, 0.1 mM EDTA, 0.1% CHAPS, 0.1% Tween20, pH 7.4), and incubated for 5 min on ice. The resulting cell lysates were centrifuged at 13 000 × g for 10 min at 4°C and the supernatant (cytosolic fraction) was kept at −70°C until analyzed. The test was performed according to the manufacturer's protocol (Calbiochem). Briefly, a sample of 15 μl of each cell lysate was added to a well of a microtiter plate (Corning), containing 75 μl of assay buffer (100 mM NaCl, 50 mM HEPES, 10 mM DTT, 1 mM EDTA, 0.1% CHAPS, 10% glycerol, pH 7.4). The positive control in this assay was a recombinant human caspase-8. To validate the caspase-8 activity in the lysate, a negative control for each sample was analyzed. This control was an identical lysate sample to which 20 μl of Granzyme B Inhibitor II (Ac-IETD-CHO), final concentration 50 ng/ml, was added. The plate was incubated at 37°C for 10 min. Then, the reaction was initiated by adding 10 μl of Granzyme B Substrate I, Colorimetric (Ac-IETD-pNA), final concentration 200 mM, and the change in absorbance was monitored at 405 nm for 2 h, at 10-min intervals.

Protein concentration of all lysates was determined by using the Bio-Rad protein assay kit (Richmond, CA, USA), and protease activity of caspase-8 was calculated as a specific activity (pmol/min/μg protein).

Statistical analysis

The statistical significance of differences between groups was performed by analysis of variance ANOVA: single factor, given in P-value. P-values of <0.05 were considered statistically significant; P-values of <0.01 were considered highly significant.