INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Mycobacterium tuberculosis, the causative agent of tuberculosis, is responsible for 1.5 million deaths per year, although only 10% of immunocompetent people infected with M. tuberculosis develop active tuberculosis (13). Most people contain, although probably do not eliminate, the initial infection as a result of cell-mediated immunity. It has been established that T cells provide protection against M. tuberculosis (reviewed in reference 9), and there are studies supporting roles for both CD4 and CD8 T cells (4, 6, 16, 27, 40, 51). The organism resides and multiplies within macrophages (M), which can be activated by gamma interferon (IFN-) and other signals to control the infection.

The interactions between dendritic cells (DCs) and pathogens are of prime importance in establishing an appropriate immune response. It has been demonstrated that DCs internalize various microbes (14, 20, 23-25, 32, 36, 37, 39). We reported previously that human peripheral blood-derived DCs phagocytose M. tuberculosis and subsequently display a phenotype consistent with that of a mature DC (24). Recently, it has been reported that a murine DC line also can be infected with M. tuberculosis, which results in a mature phenotype with secretion of inflammatory cytokines (50). The initiation of a protective immune response depends on the interaction of antigen-presenting cells (APCs) and naive T cells, which occurs in lymphoid organs, including lymph nodes. DCs are considered to be the most potent APCs and play a crucial role in the initiation of an adaptive immune response. Following phagocytosis of an antigen (such as a bacterium), mannose and Fc receptors on DCs are downregulated, while adhesion, antigen-presenting, and costimulatory molecules for T cells are upregulated, resulting in a mature DC (42, 54). The murine model of tuberculosis has provided considerable insight into the immune responses to M. tuberculosis. In this study we extend the findings from our human DC studies to the mouse system, indicating that the responses of DCs and M to M. tuberculosis infection differ. In addition, we present data showing that the fate of the organism within these two cell types is different. This may have implications for the immune response against this microbe, as well as for the persistence of bacilli in the host.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Mice. Adult female 8- to 10-week-old C57BL/6 mice (Jackson Laboratories, Bar Harbor, Maine) were used. Nitric oxide synthase (NOS2)^/ mice on a C57BL/6 background were generated as described by MacMicking et al. (30) and kindly provided by Timothy Billiar (University of Pittsburgh School of Medicine). All mice were maintained in a specific-pathogen-free biosafety level 3 facility.

Culture and purification of DCs and M. DCs and M were generated from the bone marrow cells of C57BL/6 mice. Briefly, cells were extracted from the femur and tibia bones of mice in Dulbecco modified Eagle medium (DMEM). For the M cultures, the cells were washed twice in DMEM, and 2.5 × 10⁶ cells were plated in LabTek PS petri dishes (Fisher Scientific, Pittsburgh, Pa.) in 25 ml of DMEM supplemented with 10% certified fetal bovine serum, 1 mM sodium pyruvate, 2 mM L-glutamine (Life Technologies, Grand Island, N.Y.), and 33% supernatant from L-cell fibroblasts cultured for 5 to 6 days. All reagents were lipopolysaccharide (LPS)-free, and no antibiotics were used. The medium was changed on day 3. On day 5, adherent cells were washed twice with ice-cold phosphate-buffered saline (PBS) (Life Technologies), incubated for 20 min on ice, and harvested by using cell scrapers (Becton Dickinson Labware, Lincoln Park, N.J.). The cell concentration was adjusted to 1.0 × 10⁶ cells/ml, and cells were placed in Teflon jars (1 ml) (Savillex, Minnetonka, Minn.) or aliquoted into a 96-well plate (200 µl/well) for infection.

Bacteria. M. tuberculosis strain Erdman (obtained from the Trudeau Institute, Saranac Lake, N.Y.) was passed through mice, grown once in liquid (7H9 Middlebrook; Difco) medium, and frozen in aliquots at 80°C. Aliquots were used to start cultures at a concentration of 2.5 × 10⁶ /ml in 7H9 medium; bacteria were grown in 5% CO₂ at 37°C. Cultures were used at day 6 or 7 to infect cells. The bacteria were washed and resuspended in DC or M medium, sonicated 10 s in a cup-horn sonicator, and then added to the cell cultures after estimation of bacterial numbers based on previous experience. Enumeration of viable bacteria to confirm the multiplicity of infection (MOI) was done by plating for viable CFU on 7H10 Middlebrook medium and incubated for 18 days at 37°C with 5% humified CO₂.

Infection of DCs and M. After culture for 5 days, DCs (at 10⁶/ml in DC medium plus rm-GM-CSF but without IL-4) were infected in 24- or 96-well plates with M. tuberculosis at an estimated MOI of 3 to 5. After 12 h, unincorporated bacteria were removed by pelleting the DCs at low speed (<1,000 rpm) and reculturing them with fresh media. In some experiments, M were cultured and infected in 96-well plates; monolayers were then washed to remove extracellular bacteria, and fresh medium was added. In experiments involving quantitative cultures of intracellular bacterial growth, stasis, or killing, the MOI was reduced to 1, and extracellular bacteria were removed after 4 h. To estimate the percentage of infected cells for each experiment, DCs and M were either air dried on poly-L-lysine-coated slides or grown in parallel in glass culture well slides (Nalgene) and fixed in 1% paraformaldehyde at each time point. Slides were stained by the Kinyoun method for acid-fast bacteria. For phenotypic assays, DCs and M were cultured for an additional 48 h.

Transmission electron microscopy. Uninfected and M. tuberculosis-infected DCs were cultured as described above for 72 h postinfection. Cells (4 × 10⁵) were gently pelleted at 2,000 rpm for 2 min in microcentrifuge tubes, washed twice in PBS, and then fixed in 1.5% paraformaldehyde and 1.0% glutaraldehyde in PBS. Cells were postfixed in 1% osmium tetroxide in PBS, dehydrated through a graded series of alcohols, and embedded in Epon (Energy Beam Sciences, Agawam, Mass.). Thin (60-nm) sections were cut using a Reichert Ultracut S (Leica, Deerborn, Mich.), mounted on 200-mesh copper grids, counterstained with 2% aqueous uranyl acetate for 7 min and 1% aqueous lead nitrate for 2 min, and observed using a JEOL 1210 transmission electron microscope (Peabody, Mass.).

Flow cytometry analysis of cell surface markers. DCs and M were obtained and infected as described above. Approximately (2 to 5) × 10⁵ were aliquoted into tubes and stained for surface markers using antibodies against major histocompatibility complex (MHC) class I (phycoerythrin [PE]-conjugated anti-mouse H2D^b clone KH95 with control BALB/c immunoglobulin G2b [IgG2b]), MHC class II [fluorescein isothiocyanate (FITC) anti-mouse IA^b clone AF6-120.1 with control mouse IgG2a ()], intercellular adhesion molecule-1 (ICAM-1) (FITC anti-mouse CD54 clone 3E2 with control hamster IgG), B7.1 (FITC-conjugated anti-mouse CD80 clone 16-10A1 with control hamster IgG), B7.2 (FITC-conjugated anti-mouse CD86 clone GL-1 with control rat IgG2a), and CD14 (PE-conjugated anti-mouse clone rm-C5-3 with control PE-conjugated rat IgG1) in PBS containing 20% mouse serum, 0.1% bovine serum albumin, and 0.1% sodium azide (fluorescence-activated cell sorter [FACS] buffer) for 30 min at 4°C. All antibodies were used at 0.2 µg/10⁶ cells and were obtained from Pharmingen (San Diego, Calif.). Cells were fixed in 2% paraformaldehyde for 4 to 15 h and analyzed by flow cytometry (FACSCaliber) using Lysis II (acquisition) and CellQuest (analysis) software (Becton Dickinson Immunocytometry Systems, San Jose, Calif.).

Phagocytosis assay. Infected or control bone marrow-derived DCs and M were harvested from culture and resuspended at 4 × 10⁵/ml in DMEM and kept on ice. Then, 8 µl of FITC-latex conjugated beads 3 µm in diameter (Polysciences, Warrington, Pa.) was added to the cells and mixed well. The cells were incubated with the beads for 2 h at 37 or 4°C. The cells were washed five times after incubation with ice-cold FACS medium and then fixed for 1 h with 1% formaldehyde before analysis by flow cytometry (FACScan; Becton Dickinson).

Cytokine production. Supernatants from control and M. tuberculosis-infected DCs and M cultures were harvested postinfection, filtered (0.2-µm-pore-size filters), and stored at 80°C. Enzyme-linked immunosorbent assay (ELISA) antibody pairs were used to detect IL-10 (JES5-SXC1 and JES5-2A5), IL-12p70 heterodimer (C15.6 and C17.15) (Biosource International, Camarillo, Calif.), and tumor necrosis factor alpha (TNF-; MP6-XT22 and MP6-XT3) (Pharmingen) in the supernatants. Recombinant cytokines (Pharmingen and The Genetics Institute, Cambridge, Mass.) were used to generate standard curves. The ELISAs were performed according to Pharmingen's protocol.

RPA. Determination of the levels of mRNA for the genes of interest at various time intervals after the infection was performed using a multiprobe RNase assay system (Pharmingen). Total RNA was extracted from DCs and M uninfected or infected with M. tuberculosis as detailed above, using Trizol reagent (Life Technology). The extracted RNA was subjected to an RNase protection assay (RPA) according to the manufacturer's instructions. Protected [³²P]UTP-labeled probes were resolved on a 6% polyacrylamide gel and analyzed by autoradiography. Cytokine analysis was performed using custom-made probe sets specific for NOS2, IL-12p40, IL-1, IL-1, IFN-, and IL-10 (mCK3) and for IL-4, IL-2p40, TNF-, IFN-, IL-1, and IL-1 (mCK2B). The expression of specific genes was quantified densitometrically (ImageQuant; Molecular Dynamics, Sunnyvale, Calif.) relative to the abundance of the housekeeping genes for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or L32.

T-cell proliferation and cytokine production assays. DCs and M uninfected or infected (MOI = 4) as described above for 24 h and then plated in triplicate in 96-well U-bottom plates (Corning Incorporated) at various concentrations to achieve 1:20 to 1:100 APC/T-cell ratio in RPMI 1640 medium (Life Technologies) containing 10% certified fetal bovine serum, 1 mM sodium pyruvate, 2 mM L-glutamine, 25 mM HEPES (Life Technologies), and 50 µM 2-mercaptoethanol (Sigma). As a source of sensitized T cells, spleens were obtained from C57BL/6 mice infected for 4 to 5 weeks, and single-cell suspensions were obtained by crushing the spleens in cell strainers (Becton Dickinson Labware). Red blood cells were lysed with NH₄Cl-Tris solution, and cells were washed twice. M were depleted by adherence on plastic petri dishes for 2 h at 37°C. In some experiments, B cells were depleted by adherence on anti-IgG-anti-IgM antibody-coated plates (Zymed Laboratories, Inc., San Francisco, Calif.). Lymphocyte-enriched splenocytes were added at (2 to 4) × 10⁵ cells/well and cultured with different APCs for 3 days. The proliferation of T cells in medium alone served as a baseline. As a positive control, cells were stimulated with concanavalin A (ConA; Boehringer Mannheim Corp., Indianapolis, Ind.) at 5 µg/ml. Cells were pulsed for the final 12 to 18 h of culture with 1 µCi of [³H]thymidine (Amersham Life Sciences, Inc., Arlington Heights, Ill.) per well, and the incorporation of radioactivity was measured by counting cell lysates on filters in a scintillation counter. The stimulation index (SI) was determined as follows: SI = (cpm of T cells plus infected APCs)/(cpm of T cells plus uninfected APCs). Culture supernatants were harvested after 3 days of culture, and IFN- production was measured by sandwich ELISA using the antibodies R4-A62 and XMG1.2 (biotinylated) (Pharmingen) according to the manufacturer's protocol. Recombinant murine IFN- used to generate a standard curve was a gift from Genentech (San Francisco, Calif.).

Antimycobacterial activity of M and DCs. The antimycobacterial activity of DCs and M was assessed by metabolic labeling of intracellular M. tuberculosis with [³H]uracil as previously described (8, 43). DCs and M (2 × 10⁵ cells/well, duplicate or triplicate wells for each condition) were primed with IFN- (100 U/ml; Genentech, Inc., San Francisco, Calif.) for 12 to 24 h, and then LPS (1 µg/ml) (Sigma, St. Louis, Mo.) was added. Activated and resting cells were infected with M. tuberculosis (sonicated in a cup-horn sonicator for 20 s to reduce clumping) at an MOI of 3 to 5. Aminoguanidine (AG), an NOS2 enzyme inhibitor, was added to some conditions of the IFN--treated and untreated cells 4 h prior to LPS addition and infection to inhibit NOS2 activity. At 24 h postinfection, the cells were pulsed with 2.5 µCi of [³H]uracil, which is incorporated predominantly by the bacteria and not by M or DCs. The supernatants were removed 8 to 16 h later, the cell pellets were lysed with 1% saponin and trichloroacetic acid (TCA) precipitated onto GF/C glass fiber filters (Fisher Scientific, Pittsburgh, Pa.), and radioactive incorporation was measured by using a -scintillation counter. The percent inhibition was calculated as follows: 100  [(cpm of activated cells/cpm of resting cells) × 100].

Determination of nitrite accumulation. Nitrite (NO₂) accumulation in the supernatant of cultured cells was measured as an indicator of NO production by a Griess assay, with a sodium nitrite standard, as previously described (21). Supernatants from 2 × 10⁵ cells (100 µl) of each condition were assayed in duplicate or triplicate, and the absorbancy was measured at 570 nm using an Emax precision microplate reader (Molecular Dynamics).

Statistics. For statistical analysis of samples, paired and unpaired Student t tests were used (Instat, v.2.03; GraphPad Software, San Diego, Calif., and Stat View, Abacus Concepts, Berkeley, Calif.). P values of <0.05 were considered significant.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Infection of murine DCs with M. tuberculosis. Immature DCs are highly phagocytic and readily take up various microbes (46, 48). We generated immature murine DCs by culture of bone marrow cells for 5 days in LPS-free media supplemented with 5% mouse serum, rm-GM-CSF, and rm-IL-4. These DCs have a very immature phenotype, as judged by low cell surface expression of MHC class II, B7.1, and B7.2, as well as moderate expression of ICAM (Fig. 1), morphology, and a high phagocytic ability, as determined by the uptake of FITC-labeled beads (Fig. 2). M were also derived from bone marrow precursors. The DCs and M were infected with M. tuberculosis at an estimated MOI of 3 to 5 for 12 h, which resulted in 50 to 70% of the cells infected, as judged by staining for acid-fast bacilli. The use of a higher MOI (5 to 10) resulted in a higher percentage of infected cells or more bacteria per cell but was associated with a loss of cell viability. To confirm that the mycobacteria were internalized, infected DCs were examined using transmission electron microscopy (Fig. 3). M. tuberculosis bacilli were observed within vacuoles of the DCs, and multiple bacilli were often present within the cell. Bacteria free in the cytoplasm were not observed. After infection with M. tuberculosis, both DCs and M undergo a change in cell density and size (Fig. 1B).

View larger version (36K): [in this window] [in a new window]   FIG. 1.   Cell surface phenotype of DCs infected with M. tuberculosis. Purified bone marrow-derived DCs and M were infected with M. tuberculosis or uninfected and harvested 48 h postinfection. (A) The cells were stained with antibodies as follows: MHC class II (IAb)-FITC conjugated and an isotype control mouse IgG2a; ICAM (CD54)-FITC and B7.1 (CD80)-FITC and an isotype control hamster IgG; B7.2 (CD86)-FITC and an isotype control rat IgG2a; and CD-14 (PE conjugated) and an isotype control rat IgG1(). The cells were then fixed and analyzed by FACS analysis. Top row, M; bottom row, DCs; dotted line, IgG control; thin gray line, uninfected cells; thick black line, M. tuberculosis-infected cells. (B) The forward versus side-scatter plots show that both the DCs and the M undergo morphologic changes after infection. A representative experiment of 10 independent experiments is shown.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Phagocytic capability is decreased in M. tuberculosis-infected DCs but not macrophages. Purified DCs and M were cultured without (top panel) or with M. tuberculosis (bottom panel). After 72 h, cells were incubated with FITC-labeled latex beads (3 µm) for 2 h at 4 or 37°C as indicated and examined by flow cytometry. Dead cells and free beads were excluded from the gate, and 104 cells were collected within the gate for each sample. The MFI is indicated in each histogram. The bar above the histograph represents the gate. Percent gated: DCs, uninfected, 20% at 4°C and 76% at 37°C; DCs plus M. tuberculosis, 5% at 4°C and 37% at 37°C; M, uninfected, 33% at 4°C and 93% at 37°C; M plus M. tuberculosis, 14% at 4°C and 83% at 37°C. The experiment was repeated once.

View larger version (113K): [in this window] [in a new window]   FIG. 3.   Immature murine DCs internalize live M. tuberculosis. DCs infected with M. tuberculosis were fixed, embedded, sectioned, and examined by transmission electron microscopy. The arrows indicate M. tuberculosis. The micrograph shown is representative of 25 individual cells observed. Bar = 1 µm.

Murine DCs and M respond differently to M. tuberculosis infection. We previously reported that M. tuberculosis infection of human peripheral blood mononuclear cell (PBMC)-derived DCs resulted in the phenotypic maturation of the cells. To confirm a similar effect on murine DCs, we examined cell surface marker expression following M. tuberculosis infection of murine bone marrow-derived DCs and M by flow cytometry. The DCs generated from bone marrow were cultured in mouse serum and have a very immature phenotype. DCs infected with M. tuberculosis showed a consistent upregulation in expression of the cell surface molecules ICAM, B7.2, and B7.1 compared to uninfected cells (Fig. 1A, Table 1), suggesting a shift to a more mature phenotype. Although the M express MHC classes I and II, B7.1, B7.2, and ICAM-1, these were not generally upregulated following infection with M. tuberculosis. The M express high levels of CD14, whereas the DCs express low-level CD14 which is slightly upregulated upon infection, as described by others (31) (Fig. 1A, Table 1). The morphology of both populations of cells is modified postinfection (Fig. 1B).

Inflammatory cytokines induced by M. tuberculosis infection of DCs. M infected with M. tuberculosis secrete inflammatory cytokines, which can influence cytokine production by T cells (3). The production of cytokines by DCs during the priming of T cells in the lymph node is likely to affect the initiation of an immune response against an infection. TNF-, IL-12, and IL-10 secretion and gene expression as measured by ELISA and RPA, respectively, were assessed in M. tuberculosis-infected DC and M cultures. TNF- was produced by M. tuberculosis-infected DCs and M by 4 h postinfection, with decreased production at later time points (Fig. 4A). Infected DCs consistently produced more TNF- early postinfection compared with infected M, but TNF- production by infected M increased over time. This pattern was also observed at the RNA level (Fig. 4B). Uninfected DCs and M produced negligible amounts of TNF- (Fig. 4, time zero). IL-12 protein was detectable by 12 h in M. tuberculosis-infected DCs and by as early as 4 h in infected M and increased over the course of infection (Fig. 4A). Infected M produced two- to fivefold more IL-12 than infected DCs. Uninfected DCs and M produce negligible amounts of IL-12 (Fig. 4, time zero). IL-12 gene expression was barely detectable in M. tuberculosis-infected DCs and M until 24 h; no gene expression was observed in uninfected cells. IL-10 was not detected in either uninfected or M. tuberculosis-infected cells at any of the time points examined either by ELISA or by intracellular cytokine staining (data not shown). However, a low level of IL-10 mRNA was detected in M. tuberculosis-infected DCs and M, and this level increased over time (Fig. 4B). In addition, IL-1 and IL-1 gene expression increased with infection in both cell types (not shown).

View larger version (22K): [in this window] [in a new window]   FIG. 4.   Cytokine secretion in response to M. tuberculosis-infected DCs and M. (A) Supernatants from M. tuberculosis-infected unactivated DCs () and M () were assayed in duplicate for production of IL-12 (ELISA for IL-12p70) or TNF. (B) Time zero represents cytokine production from the cells prior to infection. Recombinant murine cytokines were used as standards in each assay. Representative experiments of four and two total experiments are shown for IL-12 and TNF-, respectively. (B) Analysis of gene expression by RNase protection assay. A total of 6 µg of RNA from either uninfected or M. tuberculosis-infected DC () and M () at various time points was used in the assay, and specific genes of interest were standardized to GAPDH by densitometerization of the autoradiograph. A representative experiment (of two experiments) is shown.

Stimulation of T-cell proliferation and cytokine production by APCs. The differential effect of mycobacterial infection on expression of costimulatory molecules by DCs and M suggested that they might differ in mycobacterial antigen-presenting functions. To assess the ability of infected DCs and M to stimulate T-cell effector functions, T cells from M. tuberculosis-infected mice were cultured with either DCs or M as APCs. Both uninfected and live M. tuberculosis-infected APCs were used. M. tuberculosis-infected DCs induced substantial proliferation of T cells from infected mice (Fig. 5A). In contrast, uninfected DCs and M and, surprisingly, even infected M, were very poor stimulators of T cells from infected mice (Fig. 5A). T cells, DCs, or M alone did not proliferate to a significant extent (data not shown). A correlation was observed between the stimulation of T-cell proliferation and IFN- production. M. tuberculosis-infected DCs readily stimulated secretion of IFN- by T cells from infected mice, at levels comparable to that induced by ConA (Fig. 5B). However, little IFN- secretion was observed when either infected M or uninfected DCs and M were used as APCs (Fig. 5B).

View larger version (17K): [in this window] [in a new window]   FIG. 5.   Stimulation of T-cell proliferation and cytokine production by APCs. T-cell-enriched splenocytes from C57BL/6 mice were cultured for 3 days with bone marrow-derived DCs or M either infected with M. tuberculosis (MOI = 4) or left uninfected at a 1:50 APC/T-cell ratio. (A) Proliferation was measured by [3H]thymidine incorporation and is represented as the SI. Stimulation with ConA resulted in ~18,000 cpm, and stimulation with infected DCs resulted in ~16,000 cpm. DCs, M, or T cells alone incorporated fewer than 2,000 cpm. (B) Supernatants of the cultures described above were collected at the end of the culture period, and the concentration of IFN- was measured by sandwich ELISA. A representative experiment (of three experiments) is shown.

Activated DCs inhibited the proliferation of M. tuberculosis in vitro. M. tuberculosis can grow within unactivated macrophages, and this is a major site for bacterial replication in vivo. However, murine M activated by IFN- and either TNF- or LPS inhibit intracellular growth of M. tuberculosis and kill 50 to 70% of the bacteria via NOS2-dependent reactive nitrogen intermediate (RNI) production (10). It has been reported that murine DCs can produce NO in response to LPS and/or IFN- (5). We examined the ability of DCs, either unactivated or activated with IFN- and LPS, to support or inhibit the growth of intracellular M. tuberculosis. Intracellular mycobacterial proliferation was assessed by measuring [³H]uracil incorporation into M. tuberculosis, which indicates bacterial growth but does not determine whether the bacteria are killed (10, 44). M. tuberculosis in unactivated DCs or M incorporated [³H]uracil, indicating the replication of intracellular bacteria (Fig. 6A). DCs and M activated with IFN- and LPS inhibited the growth of intracellular M. tuberculosis compared to unactivated cells, and the inhibition was comparable (76% inhibition for DCs compared to 75% inhibition for M). The inhibition correlated with RNI production by activated DCs or M (Fig. 6B). Addition of an NOS2 inhibitor, aminoguanidine, to both DCs and M prior to infection abrogated the effect on intracellular M. tuberculosis (Fig. 6A). Confirmation of the importance of RNI production in the antimycobacterial activity of DCs was obtained using DCs and M generated from NOS2^/ mice. IFN-- and LPS-activated NOS2^/ DCs failed to inhibit M. tuberculosis proliferation (Fig. 6C), similar to the result obtained with activated NOS2^/ M (Fig. 6C).

View larger version (27K): [in this window] [in a new window]   FIG. 6.   Activated DCs and M are comparable in the inhibition of M. tuberculosis growth via RNI production. (A) Incorporation of [3H]uracil into M. tuberculosis within DCs and M was assessed in untreated and IFN--LPS-activated DCs and M. The lysates of [3H]uracil-labeled uninfected, M. tuberculosis-infected, or IFN--LPS-activated M. tuberculosis-infected wild-type DCs and M were TCA precipitated, collected on glass filters, and counted in a -scintillation counter. An inhibitor of NO production, aminoguanidine (AG), was added to the cells prior to infection as indicated. (B) Griess assays were performed with supernatants of M. tuberculosis-infected DCs and M and read at 570 nm. The standard curve was generated using NaNO2. (C) Lysates of [3H]uracil-labeled uninfected, M. tuberculosis-infected, or IFN--LPS-activated M. tuberculosis-infected NOS2/ DCs and M were TCA precipitated and counted using liquid scintillation. Samples were tested in triplicate, and a representative experiment from three experiments is shown. For statistical analysis for all panels, IFN--LPS or IFN--LPS plus AG samples were compared to untreated cells. *, P < 0.02.

Mycobacterial growth and killing within DCs and M. DCs encountering M. tuberculosis in the lung might not be activated initially, and thus the intracellular bacteria would be able to multiply and perhaps use this migrating cell to gain access to the lymph node. The ability of DCs to produce RNI and so inhibit bacterial replication suggests a potential role for DCs as an effector cell. Therefore, the antimicrobial effects of activated DCs were studied further. IFN--LPS-activated DCs clearly inhibited the growth of intracellular M. tuberculosis, as described above. However, the ability of DCs to actually kill the organism was unknown. To address this, intracellular CFU levels at various times postinfection were determined by plating lysates from an equal number of unactivated and activated infected DCs and M (Fig. 7). Initial (4 h postinfection) intracellular CFU levels were similar between DCs and M (P = 0.63). M. tuberculosis grew equally well within unactivated DCs and M over 60 h (Fig. 7A, Table 2), confirming the ability of DCs to support the growth of intracellular M. tuberculosis. In activated M (treated with IFN- and LPS), the number of viable bacteria was reduced by ~50% by 48 h postinfection compared to 4 h postinfection (P = 0.02) (Fig. 7B, Table 2). The number of viable bacteria in activated DCs did not increase over time, confirming that these cells can inhibit mycobacterial replication. However, in contrast to activated M, the numbers of intracellular M. tuberculosis in activated DCs were not reduced over the course of the infection in six independent experiments (4 h versus 48 h, P = 0.21), demonstrating a lack of killing of intracellular bacteria (Fig. 7B, Table 2). Examination of the viability of the DC and M cultures over the course of infection was performed by staining parallel cultures of infected cells in chamber slides and staining them with trypan blue and for acid-fast bacilli. At 48 h, there was no difference in the viabilities of the DCs and M, although at later time points (72 to 90 h) there was a marked deterioration of both cell types. For this reason, we compared intracellular killing only for up to 48 h postinfection. In addition, the culture supernatants at each time point were plated to determine the number of bacteria that had escaped the cells, perhaps due to cell death. In all cases up to 48 h, very few bacteria (<1% of the total bacterial numbers) were present in the supernatants, and there was no difference between DC and M cultures. Thus, the discrepancy in killing of intracellular organisms between DCs and M cannot be attributed simply to differences in the cell viability. The activated DCs and M both produced RNI (Table 2), which has been shown to be essential in killing intracellular M. tuberculosis (10). There is not a higher level of NOS2 at either the protein or the RNA level in activated M compared to DCs (data not shown).

View larger version (13K): [in this window] [in a new window]   FIG. 7.   DCs inhibit the growth of but do not kill intracellular M. tuberculosis. Cell pellet lysates of M. tuberculosis-infected DCs or M (MOI = 1), unactivated or activated with IFN--LPS, were serially diluted in PBS plus 0.05% Tween 80 and plated on 7H10 plates, which were incubated for 18 days at 37°C and 5% CO2. Three wells per condition were assessed at each time point, and the mean intracellular CFU values are reported at various time points after infection. (A) Intracellular CFU in resting DCs () and resting M (). (B) Intracellular CFU in activated DCs () and activated macrophages (). At each time point, <1% of the total CFU was found in the supernatant. In panel A, CFU levels within unactivated DCs and M at each time point revealed no statistical differences (P > 0.5). In panel B, bacterial numbers within DC were not significantly reduced (P = 0.2) compared to intracellular CFU at 4 h (input). Comparison of CFU within M at each time point to the CFU level at 4 h revealed statistically significant reductions over time (**, P < 0.01; *, P < 0.02). Error bars show the standard error. A representative experiment (of six experiments) is shown.

View this table: [in this window] [in a new window]   TABLE 2.   Activated M kill M. tuberculosis, whereas activated DCs inhibit the growth of M. tuberculosisa

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The interaction of DCs with infectious agents plays a vital role in the initiation of the immune response against a microbe. These potent APCs encounter antigen at the site of infection, traffic to the lymph node, and prime T cells to respond to the microbial antigen (2). The effector T cells migrate back to the site of infection, produce cytokines, activate macrophages, and lyse target cells in an effort to eliminate the microbial agent. In the studies presented here, we demonstrate that DCs interact with live M. tuberculosis bacilli in a manner different from that of M. M are the cells in which M. tuberculosis is believed to live and multiply within the host. However, M. tuberculosis grew equally well within DCs and M, and activated DCs and M were equivalent in their ability to inhibit replication of M. tuberculosis in an NOS2-dependent manner. Our data indicate that, whereas activated M could kill intracellular M. tuberculosis, activated DCs were not mycobacteriocidal and did not eliminate intracellular bacilli.

DCs readily internalized M. tuberculosis bacilli and subsequently displayed phenotypic and functional changes, including the upregulation of various cell surface molecules important in initiating immune responses and the downregulation of phagocytic ability, in addition to producing inflammatory cytokines. These DCs were superior to macrophages in stimulating proliferation and IFN- production of mycobacterium-specific T cells (47, 50; this study). Mouse bone marrow-derived DCs infected with M. tuberculosis for 48 h demonstrated increased cell surface expression of ICAM-1, B7.1, and B7.2 molecules (both as percent positive cells and the MFI) while decreasing their phagocytic ability, and this differed from the M. tuberculosis infection of macrophages. These data are consistent with our previous work with M. tuberculosis infection of human PBMC-derived DCs, although the upregulation of cell surface molecules was less profound in the murine system, and with the results reported by Tascon et al. using the tsDC cell line (50). Although the levels of B7 expression were similar between our primary DCs and the tsDC line, we also observed an increase in the level of ICAM-1 after infection, which was not observed using the tsDC cell line (50). A consistent up- or downregulation of MHC class I or II was not observed in our studies. It may be that MHC levels are susceptible to variables such as MOI, the percentage of infected cells, and the number of live (or dead) bacilli per cell. We also did not observe downregulation of MHC class II on macrophages following infection, in either this study or in our study with human DCs and M (24), in contrast to published reports (17, 22, 38).

Inflammatory cytokine production by DCs would be expected to influence the T-cell phenotype primed in the lymph node (29). M. tuberculosis-infected DCs produced low to moderate amounts of IL-12, which would skew the primed T-cell response toward a type 1 (IFN--producing) response. Production of this cytokine was lower in DCs than in infected M. IL-12 production by DCs following infection with microbial pathogens has been reported, although the downregulation of constitutive IL-12 production by DCs following infection with certain pathogens has also been reported (1, 53). DCs infected with Toxoplasma gondii have a refractory period where IL-12 production is low but, upon restimulation, IL-12 production can increase (7). TNF- production following M. tuberculosis infection was initially higher in DCs compared to M, but the M did not decrease TNF- production as quickly. This is similar to the pattern seen with M. tuberculosis infection of human DCs (24). IL-10 production was not observed in infected DCs or M for up to 60 h postinfection. Low levels of mRNA were detected and increased by 24 h (Fig. 4C). This downregulatory cytokine may be produced later in infection or may be produced primarily by T cells during this infection.

The fate of a microbe within a DC may affect the presentation of antigen by the DC to naive T cells. The bactericidal capabilities of DCs might be expected to correspond to those of M, given the similarity in the progenitor cells, but studies examining the fate of microorganisms within DCs are rare. DCs encountering a pathogen in tissue, for example, the lung, early in infection are not likely to also encounter specific T-cell-secreting cytokines, such as IFN-, to activate the DCs. M. tuberculosis bacilli replicated equally well within unactivated DCs and M. Thus, maturation of the DCs in response to M. tuberculosis infection had little effect on the intracellular survival and growth of the microbe. Although it has been reported that M. tuberculosis H37Rv did not grow within the tsDC line, we consistently observed that untreated murine bone marrow-derived DCs were as permissive for M. tuberculosis growth as bone marrow-derived M (Fig. 7). The discrepancy may be due to differences between primary DCs and an immortalized line. A mature DC would be expected to traffic to the lymph nodes, and our data suggest that such a DC would be carrying live M. tuberculosis bacilli. This may be a mechanism by which M. tuberculosis gains access to the lymph node. Studies in the Leishmania murine model suggest that DCs persist in the lymph node even when infected (36). It has been hypothesized that the host's ability to control an infection is a complex balance between a low number of persistent organisms and the specific effector cells in the immune system (19, 33, 36). Therefore, a continuous supply of antigen from a living bacterium may be advantageous for the priming and maintenance of an effective immune response. In two recent studies, mice vaccinated with BCG- or M. tuberculosis-infected DCs were shown to generate a protective immune response and to reduce the bacterial burden after challenge with M. tuberculosis (11, 50). These studies suggest that targeting DCs in vivo will be an important consideration for vaccine development and design.

In an environment in which T cells or other cells are producing IFN-, the DCs would be expected to be activated in a manner similar to the M. Activated M have been shown by various groups to inhibit the growth of intracellular M. tuberculosis bacilli (10, 12, 15) and, more importantly, to kill at least 50% of the intracellular mycobacteria (10) via NOS2-dependent mechanisms. In our studies, activation of M with IFN- and LPS also resulted in a 50% reduction in intracellular M. tuberculosis by 48 h postinfection, as determined by CFU in cell lysates. DCs treated with IFN- and LPS were capable of restricting the growth of intracellular M. tuberculosis, and this was dependent on NOS2 activity. However, in contrast to M, activated DCs were unable to reduce the intracellular bacterial numbers over time, despite the essentially similar levels of RNI production in DCs and M. In some experiments, there were differences in RNI production at 4 h postinfection (Table 2). Initially, we speculated that a higher amount of RNI early in the infection produced by activated M compared to activated DCs could be responsible for the differential ability of the cells to kill M. tuberculosis. However, analysis of data from a series of experiments indicated that, regardless of the RNI production by DCs at 4 h, including experiments in which the RNI levels were identical to those of the M, the DCs did not reduce the number of input bacteria over time. It is possible that M. tuberculosis bacilli within DCs avoid the killing effects of RNI by persisting within special vacuoles of DCs. Other possibilities for the lack of mycobacterial killing include differences in the phagosome pH, lysosomal enzymes, and reactive oxygen production in DCs compared to M. These possibilities are under investigation.

There have been relatively few studies on the effect of DCs, particularly activated DCs, on intracellular bacteria. DCs have been demonstrated to take up various microbes, including Salmonella spp. (32, 49), Escherichia coli (49), Listeria monocytogenes (21), Borrelia burgdorferi (14, 35), Bordetella bronchiseptica (20), M. bovis BCG (23, 26), Leishmania spp. (18, 36, 37), and Chlamydia spp. (25, 28, 39, 55). Salmonella was reported to survive and replicate within unactivated murine DCs (32), while Chlamydia was killed via phagolysosome fusion (39). Leishmania parasites have been shown to persist within DCs in lymph nodes, suggesting that DCs are impaired in their ability to kill Leishmania (36, 37). Recently, studies indicate that certain parasite infections, such as Trypanosoma cruzi and Plasmodium falciparum, downregulate inflammatory cytokine production and prevent maturation of the DCs (52, 53). These parasites apparently manipulate DCs as an immune evasion strategy.

We hypothesize that the ability of M. tuberculosis to survive, although perhaps not replicate, within activated DCs may be beneficial in priming a T-cell response by mature DCs in the lymph node. Live bacteria are capable of secreting antigens, which are believed to be important in the protective T-cell response against tuberculosis. In addition, it has been suggested that M. tuberculosis forms a pore within the phagosomal membrane which allows access to the cytoplasm for mycobacterial peptides; presumably, priming of MHC class I-restricted CD8 T-cell responses would be more efficient in DCs harboring live, rather than dead, tubercle bacilli (34). An alternative hypothesis is that M. tuberculosis has evolved a strategy to evade killing by the DCs and uses this cell as a vehicle for dissemination from the lung to the lymph nodes and other organs.