INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mycobacterium tuberculosis is an intracellular pathogen capable of persisting and replicating within the cells of the mononuclear phagocyte system (1). Acquired resistance against this microorganism is mediated by specific T-lymphocytes and expressed by activated macrophages. Activation of macrophages for increased mycobactericidal and/or mycobacteriostatic mechanism by lymphokines released from mycobacterial antigen-triggered T-lymphocytes is presumed to play a critical role in acquired resistance against M. tuberculosis (1).

Crowle and May (2) developed an in vitro model of macrophage phagocytosis and growth inhibition of M. tuberculosis. Their model has been used to assess the effects of differentiation of human monocytes on mycobacteriostatic activity (3). They also found that interferon- (IFN-) does not possess a macrophage-activating factor (MAF) for M. tuberculosis but actually enhances intracellular growth of the organism (4), which differs from the MAF activity of the cytokine for murine bone marrow macrophages (5). In the human system in vitro, the cytokine expressing MAF activity with regard to antituberculosis activity remains controversial (2).

Although natural killer (NK) cells, which lyse spontaneously certain tumors in vitro, have been shown to provide early defense mechanism against cancer growth (6) and viral infection (7), the role of NK cells in the host defense against M. tuberculosis remains unclear. We have reported that NK cells are activated in patients with active pulmonary tuberculosis (8).

It is becoming increasingly evident that NK cells and lymphokine-activated killer (LAK) cells may have an important role in the nonspecific defense mechanism against certain intracellular microorganisms such as Legionella pneumophila (9) and Mycobacterium avium complex (10).

However, it remains unexplored whether NK cells are involved in human defense mechanism against M. tuberculosis. The differences between NK cells and CD4⁺ T-cells in terms of the M. tuberculosis-infected monocyte activation process also remain undefined. The current studies have demonstrated that both T-cells and NK cells possess the ability to activate monocytes to kill M. tuberculosis, although the antigen specificity and genetic restriction of the cellular interactions differ.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mycobacteria

An avirulent strain, M. tuberculosis H37 Ra, was used. M. tuberculosis was cultured in Middlebrook 7H9 liquid medium (Difco Laboratories, Detroit, MI) until a density of 1 × 10⁸ to 6 × 10⁸ bacteria/ml was obtained. One-milliliter aliquots of this suspension was stored frozen at 70° C until needed. Prior to infection, a frozen stock culture was thawed in a water bath and sonicated for 20 s with a microultrasonic cell disruptor (Electro-Mechanic Instrument Co., Perkasie, PA) to disperse clumps.

Isolation of Monocytes

Peripheral blood mononuclear cells were isolated from heparinized blood of healthy donors (five purified protein derivative [PPD] skin-test-negative donors and five PPD skin-test-positive donors) by sedimentation on Ficoll-Hypaque (Pharmacia, Uppsala, Sweden) gradients. The tuberculin skin test was interpreted according to the American Thoracic Society criteria (11). The cells were suspended at a density of 1 × 10⁷/ml in RPMI 1640 medium (M.A. Bioproducts, Walkersville, MD) with 2 mM L-glutamine, 15 mM HEPES buffer (M.A. Bioproducts), and 10% heat-inactivated fetal calf serum (FCS). Five milliliters of the solution were added to 100-mm petri dishes (No. 3003; Becton Dickinson Labware, Oxnard, CA) coated with FCS and pooled human serum.

The dishes were incubated for 1 h at 37° C in 5% CO₂ in moist air for the adherence of monocytes. Nonadherent cells were aspirated for T-cell/NK cell separation. The adherent monocytes were dislodged by scraping with a cell lifter and suspended at a concentration of 1 × 10⁶/ ml in RPMI 1640 medium with 2% unheated autologous serum. A total of 100 µl (10⁵/well) of the cell suspension was added to disposable 96-well flat-bottom microtiter plates (Corning Glass, Corning, NY) and incubated for 1.5 h before addition of mycobacteria.

Infection of Monocytes

Monocyte monolayers were incubated with a sonicated M. tuberculosis (10⁸ bacteria/ml) in RPMI 1640 supplemented with 30% unheated autologous serum. After 1 h of incubation, the extracellular bacilli were washed with RPMI 1640, and the bacterial suspension was replaced with RPMI 1640 with 2% autologous unheated serum. The M. tuberculosis-infected monocytes were cultured for 16 h at 37° C.

Separation of High-density T-cells and NK Cells

High-density T-cells and NK cells were purified by sedimentation on a discontinuous Percoll (Sigma, St. Louis, MO) density gradient by a modification of the procedure described by Timonen and colleagues (12). Percoll was adjusted to isotonicity by the addition of 10% vol/vol 10-fold-concentrated Hanks' balanced salt solution (HBSS). Four concentrations of Percoll in medium were prepared (40 to 54%) with Percoll varying by approximately 3.5% increments. Two milliliters of each Percoll concentration were sequentially layered in 15-ml conical polystyrene tubes (Falcon; Becton Dickinson Labware).

The plastic-nonadherent cells were incubated on nylon-wool columns to remove adherent cells and B-lymphocytes. The eluted cells were washed and resuspended in medium with 10% FCS and loaded on the top layer of the Percoll. The tube was centrifuged at 400 g for 30 min at room temperature. The NK-cell-rich population was collected from the second (40:43.5%) and third (43.5:47%) interfaces. High-density T-cells were collected from the fourth interface (47:54%). Collected cells were washed three times and suspended in RPMI 1640 with 2% autologous serum. The NK-cell-rich population comprised 96% NK cells, as assessed, using an antihuman Leu-lla monoclonal antibody (Becton Dickinson Immunocytometry Systems, Mountain View, CA), by indirect immunofluorescence and FACS analysis. The preparation of NK cells did not react with anti-CD3 monoclonal antibody (Becton Dickinson). High-density T-cells comprised 94% CD3-reactive T-lymphocytes.

Induction of Lymphokine-activated Killer Cells

NK cells were incubated at a concentration of 2 × 10⁶/ml with 200 U/ ml of recombinant human interleukin-2 (IL-2) for 48 h in RPMI 1640 with 10% FCS at 37° C and 5% CO₂ in 24-well plates (13). The cells were then washed twice and resuspended in RPMI 1640 with 2% autologous serum.

Enrichment of CD4⁺/CD4 Lymphocytes

CD4⁺ T-lymphocytes were positively selected using Dynabeads M-450 CD4 (Dynal A.S., Oslo, Norway), which are magnetizable polystyrene beads coated with a mouse monoclonal antibody specific for the CD4 antigen (14). Briefly, nylon-wool column nonadherent cells were incubated at 4° C for 30 min with Dynabeads M-450 CD4 in a 1:3 ratio. CD4⁺ T-lymphocytes were attracted to the wall of the test tube by placing the tube in the magnetic particle concentrator (Dynal A.S.) for 2 min. CD4 T-lymphocytes were negatively selected in the supernatant depleted of CD4⁺ T-lymphocytes. Magnetic beads were detached from isolated CD4⁺ lymphocytes by cultivating the rosetted CD4⁺ T-lymphocytes overnight and pipetting with Pasteur pipettes. Isolated cells were washed two times and resuspended in RPMI 1640 with 2% autologous serum.

Intracellular Killing of M. tuberculosis within Monocytes

Purified NK cells, high-density T-cells, LAK cells, and CD4⁺/CD4 T-lymphocytes were returned to the M. tuberculosis-infected monocyte monolayers (according to the experimental design) at a 5:1 ratio, and incubated at 37° C in 5% CO₂ for as long as 10 d. Immediately, and at 4 and 7 d after infection, supernatants of the monocyte were removed and the supernatants and the plates were stored separately at 70° C. The number of adherent cells in each culture was estimated according to the method described by Nakagawara and Nathan (15).

Generation of NK Cell and High-density T-cell Supernatants

NK cells and high-density T cells were cultured with M. tuberculosis-infected monocytes at 37° C in 5% CO₂. At 4 and 7 d thereafter, the supernatant was harvested and centrifuged at 800 g for 15 min. The recovered supernatant was filtered through a 0.22-µm Millipore filter (Millipore, Bedford, MA) and stored at 70° C until use. Then 200 µl of the supernatant were added to M. tuberculosis-infected monocyte monolayers and incubated for as long as 7 d.

Evaluation of Intracellular Killing

CFU counts were performed by the method originally described by Crowle and May (2). After all samples were harvested, the adherent cells in each well were washed to collect supernatant organisms and lysed with 50 µl of solution containing 86 µl of 7H9 broth and 14 µl of 0.25% sodium dodecyl sulfate. The plates were incubated for 10 min at room temperature, and then 50 µl of 20% bovine serum albumin were added to each well. The supernatants and cell lysates were sonicated for 20 s and serially diluted 10-fold; the dilutions were plated on 7H10 agar (Difco) in 60-mm petri dishes (No. 1007; Becton Dickinson). Three 10-µl spots were plated per dilution. The agar plates were incubated for 3 wk at 37° C. Results were expressed as mean number of colony-forming units (CFU) ± standard error of the mean per milliliter of lysate and supernatant.

Statistics

Significance was tested using Student's t test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Kinetics of Intracellular Replication of M. tuberculosis

Monocytes derived from human peripheral blood were infected with M. tuberculosis and cultured for as long as 10 d with medium alone (control monocytes) and in the presence of high-density T-cells or NK cells. The time course of replication of M. tuberculosis derived from monocyte lysates and supernatant from PPD-positive donors is shown in Figure 1. Control monocytes did not display any mycobactericidal or mycobacteriostatic activity. Instead, the cells permitted intracellular replication. By 10 d after infection, intracellular M. tuberculosis had multiplied from 55 ± 4 × 10³/ml to 171 ± 7 × 10³/ml. When NK cells were added to monocyte monolayers at a ratio of 5:1 NK cells to monocytes, intracellular replication was significantly inhibited at 7 d and at 10 d after infection (p < 0.01 compared with control monocytes). The addition of NK cells was associated with 61% killing of the initial inoculum by 10 d after infection (84% compared with control monocytes). When high-density T-cells were added to monocyte monolayers, intracellular replication was also inhibited at 7 d and at 10 d after infection. The treatment with high-density T-cells was associated with 43% killing of the initial inoculum (p < 0.01, 77% compared with control monocytes). There was no significant difference in killing capability between NK-cell-exposed monocytes and high-density T-cell-exposed monocytes at Days 7 and 10. At Day 4 after infection, NK cells tended to stimulate monocytes to kill intracellular M. tuberculosis more effectively than medium alone or high-density T-cells, but the difference was not statistically significant. CFU in the supernatants were approximately 10 and 20% of those seen in monocyte lysates at Days 0 and 7, respectively. Because it was necessary to consider the possibility that NK-cell-exposed or high-density T-cell-exposed monocytes could be detached from the wells, the number of the monocytes was followed carefully by staining with nuclei-counting solution. The number of monocytes at Day 7 was 8 × 10⁴/well for both monolayer treated with NK cells and with high-density T-cells, not significantly different from the number of control monocytes (9 × 10⁴/well). These findings suggested that the decrease in number of M. tuberculosis organisms from monocyte lysates represented substantial intracellular killing within activated monocytes and was not explained simply by detachment of monocytes.

View larger version (23K): [in this window] [in a new window]   Figure 1.   Kinetics of intracellular replication of M. tuberculosis derived from monocyte lysates and supernatant. M. tuberculosis-infected monocytes were incubated with medium alone (control monocyte), high-density T-cells and NK cells. Monocytes were lysed immediately and at 4, 7, and 10 d after infection for quantification of intracellular replication, which was assessed by quantitative plating techniques as colony-forming units (CFU). Data are reported as mean ± standard deviation of five PPD-positive donors. Asterisks indicate p < 0.01 compared with control monocytes (monocyte alone).

The relationship between the ratio of NK cells or T-cells to monocytes and the percent inhibition of intracellular replication compared with that within control monocytes is shown in Figure 2. When NK cells or T-cells were added to the monocyte monolayer at a 5:1 ratio, the percent inhibition was 93 and 86%, respectively, at Day 7 after infection. The percent inhibition was shown to decrease in parallel to the decrease in the ratio of NK cells or T-cells to monocytes.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Relationship between the NK cell or high-density T-cell to monocytes ratio and percent inhibition of intracellular replication. The relationship was evaluated at Day 7 after infection.

In PPD skin-test-negative donors, control monocyte permitted intracellular replication from 25.8 ± 16.8 × 10³ to 151.3 ± 22.8 × 10³ by 10 d after infection. NK-cell-exposed monocytes were associated with 93 and 63% killing compared with control monocytes and the initial inoculum, respectively. High-density T-cell-exposed monocytes were associated with 86 and 52% killing compared with control monocytes and the initial inoculum, respectively. Almost the same magnitude of killing was observed in PPD-negative donors when monocytes were cultured in the presence of NK cells or high-density T-cells.

Effect of IL-2-activated NK Cell on Intracellular Killing of M. tuberculosis by Monocytes

The effect of NK cells cultured with or without 200 U/ml of IL-2 for 2 d on the intracellular killing of M. tuberculosis by monocytes is shown in Figure 3. LAK cells, added 24 h after infection, at a 5:1 ratio, were capable of activating monocytes to kill 93% of the initial inoculum at 7 d after infection, which was comparable to 3% of the number of M. tuberculosis found within control monocytes. In contrast, NK cells cultured with medium alone exhibited modest inhibition of intracellular replication of M. tuberculosis, which suggested that the capability of NK cells to stimulate monocytes to kill intracellular M. tuberculosis waned in the culture without IL-2 for 2 d.

View larger version (18K): [in this window] [in a new window]   Figure 3.   Effect of lymphokine-activated killer (LAK) cells on intracellular replication of M. tuberculosis. NK cells were cultured with or without 200 U/ml of interleukin-2 (IL-2) for 2 d. IL-2-activated NK cells were added to M. tuberculosis-infected monocytes 24 h after infection at 5:1 and 2:1 ratios. Monocytes were subsequently lysed at the time periods shown for quantification of intracellular replication. Data are reported as mean ± standard deviation of four PPD-negative donors.

MHC Restriction of Killing

To determine whether the interaction of T-cells or NK cells with monocytes, which activated killing of intracellular mycobacteria, was major histocompatibility complex (MHC)- restricted, autologous, or allogeneic, NK cells or T-cells were added to M. tuberculosis-infected monocytes from PPD-positive and PPD-negative donors. We actually measured human histocompatibility leukocytes (HLA)-A, B, C, D, DR, DP, and DQ loci, and it was confirmed that all loci are completely different between monocyte donors and T-cell/NK cell donors in allogeneic combination. As shown in Figure 4, allogeneic NK cells exhibited almost the same stimulatory effect on monocyte killing as autologous NK cells, which suggested that this activity was not MHC-restricted. In contrast, when allogeneic T-cells from PPD-positive donors were added to M. tuberculosis-infected monocytes, no intracellular killing was observed, although autologous T-cells stimulated monocytes to kill intracellular M. tuberculosis significantly.

View larger version (17K): [in this window] [in a new window]   Figure 4.   Major histocompatibility complex restriction in monocyte killing of M. tuberculosis when treated with NK cells or high-density (H-D) T-cells. Autologous or allogeneic NK cells or T-cells were added to M. tuberculosis-infected monocytes, which were lysed for quantification of intracellular replication. Data are reported as mean ± standard deviation of three experiments from PPD-positive donors. Asterisks indicate p < 0.01 compared with control monocytes or autologous H-D T-cells.

However, in the case of the PPD-negative donor, there was no significant difference in the stimulatory capability of autologous T-cells and allogeneic T-cells (86 and 82% killing at 7 d after infection compared with control monocytes, respectively). Both MHC-mismatched and MHC-matched NK cells activated monocytes to kill intracellular M. tuberculosis as in the case of PPD-positive donors (90 and 93% killing at 7 d after infection compared with control monocytes, respectively).

Effect of CD4⁺/CD4 Lymphocytes on Intracellular Killing

To characterize the phenotype of T-cells that endowed monocytes with the capability to kill M. tuberculosis organisms, CD4⁺ and CD4 lymphocytes were separated from nylon-wool nonadherent cells (NWNC) using the magnetic beads technique and were added to M. tuberculosis-infected monocytes. As shown in Figure 5, monocytes incubated with CD4⁺ lymphocytes killed 69% of the initial inoculum, which was significantly greater than NWNC, which stimulated monocytes to kill 14% of the initial inoculum. In contrast, monocytes incubated with CD4 lymphocytes killed 27% of the initial inoculum. These results suggested that the enriched CD4⁺ lymphocyte subpopulation mediated activation of the intracellular killing more effectively than did CD4 lymphocytes in PPD-positive donors. In PPD-negative donors, CD4 lymphocytes were found to activate monocytes more significantly than CD4⁺ lymphocytes (Figure 6). There was no significant difference in monocyte-activating capability between NWNC and CD4 lymphocytes.

View larger version (16K): [in this window] [in a new window]   Figure 5.   Effect of CD4+/CD4 lymphocytes on monocyte intracellular killing of M. tuberculosis. CD4+ T-lymphocytes were positively selected using Dynabeads M-450 CD4, which are magnetizable polystyrene beads coated with a mouse monoclonal antibody specific for the CD4 antigen (19). CD4 T-lymphocytes were negatively selected in the supernatant depleted of CD4+ T-lymphocytes. Data are reported as mean ± standard deviation of three experiments from PPD-positive donors. Asterisks indicate p < 0.001 compared with control monocytes. Double asterisks indicate p < 0.01 compared with control monocytes.

View larger version (14K): [in this window] [in a new window]   Figure 6.   Effect of CD4+/CD4 lymphocytes on monocyte intracellular killing of M. tuberculosis. Data are reported as mean ± standard deviation of three experiments from a PPD-negative donor. Asterisks indicate p < 0.01 compared with control monocytes.

In order to examine whether MHC restricts the effect of CD4⁺/CD4 lymphocytes on monocyte killing, autologous or allogeneic CD4⁺/CD4 lymphocytes were added to M. tuberculosis-infected monocytes. As shown in Figure 7, autologous CD4⁺ lymphocytes effectively stimulated monocytes to kill intracellular mycobacteria. In contrast, allogeneic CD4⁺ lymphocytes failed to stimulate monocytes to kill intracellular M. tuberculosis. These results suggested that the activating effect of CD4⁺ lymphocytes on monocyte killing was genetically restricted in PPD-positive donors. Allogeneic CD4 lymphocytes stimulated mycobacteriostatic effects in monocytes comparably to autologous CD4 lymphocytes.

View larger version (17K): [in this window] [in a new window]   Figure 7.   Relation of MHC restriction of donors to the effect of CD4+/CD4 lymphocytes on intracellular killing of M. tuberculosis. CD4+/CD4 lymphocytes were enriched using the immunomagnetic beads technique described in METHODS. Autologous and allogeneic CD4+/CD4 lymphocytes were added to M. tuberculosis- infected monocytes. Single asterisks indicate p < 0.01 compared with control monocytes. Double asterisk indicates p < 0.05 compared with control monocytes. Data are reported as mean ± standard deviation of three experiments from PPD-positive donors.

MAF Effects of Supernatants Generated from Cocultivation of High-density T-Cells or NK Cells with Monocytes

To determine whether NK cell/T-cell-dependent activation of monocytes could be mediated by a secreted soluble factor, supernatants generated from cocultivation of T-cells/NK cells and monocytes were added to monocyte monolayers. As shown in Figure 8, monocytes treated with T-cell supernatants obtained from 7 and 4 d of coculture were capable of killing 99 and 41% of the initial inoculum, respectively. Monocytes treated with NK cell supernatants obtained from 7- and 4-d cocultures also were capable of killing 91 and 49% of the initial inoculum, respectively (Figure 9). These findings suggested that activation of monocyte killing of M. tuberculosis is mediated by secreted soluble factors.

View larger version (20K): [in this window] [in a new window]   Figure 8.   Effect of supernatant generated from the coculture of high-density T-cells and monocytes on intracellular killing of M. tuberculosis. High-density T-cells were cultured with M. tuberculosis-infected monocytes at 37° C in 5% CO2. At 4 and 7 d thereafter, the supernatant was harvested and centrifuged at 800 g for 15 min. The recovered supernatant was filtered through 0.22 µm Millipore filter (Millipore) and stored at 70° C until use. Then 200 µl of the supernatant were added to M. tuberculosis-infected monocyte monolayers and incubated for as long as 7 d. Single asterisks indicate p < 0.001 compared with control monocytes. Double asterisks indicate p < 0.01 compared with control monocytes. Data are reported as mean ± standard deviation of five PPD-positive donors.

View larger version (19K): [in this window] [in a new window]   Figure 9.   Effect of supernatant generated from the coculture of NK cells and monocytes on intracellular killing of M. tuberculosis. NK cells were cultured with M. tuberculosis-infected monocytes at 37° C in 5% CO2. Then 200 µl of the supernatant were added to M. tuberculosis-infected monocyte monolayers and incubated for as long as 7 d. Single asterisks indicate p < 0.001 compared with control monocytes. Double asterisks indicate p < 0.01 compared with control monocytes. Data are presented as mean ± standard deviation of four PPD-positive donors.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study using an in vitro model of the phagocytosis of M. tuberculosis by human blood monocytes demonstrated that (1) T-cells were capable of activating monocytes to kill intracellular M. tuberculosis; (2) NK cells and LAK cells also were capable of activating monocytes to kill intracellular M. tuberculosis; (3) the stimulatory effect of T-cells on monocyte killing of M. tuberculosis, especially CD4⁺ T-cells, was genetically MHC-restricted, and, in contrast, the activity of NK cells was MHC-unrestricted; (4) T-cell/NK-cell-dependent activation was mediated by secreted soluble factors.

T-lymphocytes and Intracellular Killing

Although Mackaness (1) provided evidence that activation of murine macrophages by cytokines from antigen-triggered T-cells mediates antimycobacterial resistance, there have been few reported studies demonstrating intracellular killing within human monocytes/macrophages. Moreover, the results of studies that assess killing of intracellular mycobacteria are controversial. Conflicting findings with regard to intracellular killing may be partly ascribed to differences in the host, source of mononuclear phagocytes, strain of mycobacteria, and the method for assay of intracellular killing (16). Crowle and May (2) demonstrated killing or inhibition of M. tuberculosis by human monocyte-derived macrophages activated by crude lymphokine. The same group (4) also reported that IFN- activated human macrophages to become leishmanicidal and tumoricidal but enhanced intracellular replication of M. tuberculosis. Rook and colleagues (17) showed that IFN- activated human blood monocyte-derived macrophages to reduce intracellular replication as assessed by [³H]uracil incorporation by M. tuberculosis. However, in this study, growth inhibition was not constant, and it varied between donors and stains. In murine system, Rook and colleagues (18) demonstrated that mycobacteriostatic effects can be induced in peritoneal macrophages infected with M. tuberculosis by adding Ly 1⁺2 T-cells from in vitro cell lines derived from immunized mice. They also elucidated the mechanism through which T-cells activate macrophages. The phenomenon was shown to be dependent on compatibility at the I-A locus of the MHC.

Our present study provides evidence that activation of M. tuberculosis-infected monocytes is mainly mediated by CD4⁺ lymphocytes and the phenomenon is MHC-restricted in PPD-positive donors. CD4 lymphocytes also were shown to stimulate monocytes to increase mycobactericidal capacity, although to a lesser extent. The activation by CD4 cells was MHC- unrestricted since allogeneic lymphocytes also exhibited modest mycobacteriostatic effects. In PPD-positive donors, PPD-specific CD4⁺ lymphocytes activated infected monocytes in association with MHC products.

The relative role of CD4 lymphocytes may be more important in PPD-negative donors. We found that MHC-unrestricted CD4 lymphocytes activated monocytes more effectively than did CD4⁺ lymphocytes in PPD-negative donors. CD4 lymphocytes are known to contain multiple subpopulations, including NK cells, CD8⁺ lymphocytes, and T-lymphocytes. Although CD4, CD8, and CD16⁺ positive NK cells may be involved in the activation of monocytes, it also is possible that CD3⁺, CD4, and CD8 lymphocytes and T-lymphocytes, which are in part responsible for the activation mechanism in PPD-negative donors. This possibility is consistent with the finding presented by Kabelitz and colleagues (19) that nearly half of CD4 CD8 T-cells from a healthy PPD-negative donor were M. tuberculosis-reactive.

There are two possible mechanisms by which T-cells could be involved in the killing of intracellular M. tuberculosis. First, T-cells have been shown to secrete multiple macrophage activating factors in response to antigenic triggering (20, 21). In fact, supernatants generated from cocultures of T-cells and infected monocytes mediated the activation of intracellular killing in our experiments. Among them, IFN- (22), granulocyte-macrophage colony-stimulating factor (GM-CSF) (23) and IL-2 (24) could be responsible for the activation of monocytes to mycobactericidal capacity. The effect of IFN- on killing of M. tuberculosis within human monocytes/macrophages is still controversial (4, 5, 17). Whereas, IFN- is capable of inhibiting intracellular growth of this organism when applied to murine peritoneal (17) or bone marrow macrophages (5), our preliminary study assessment (25) suggests that IFN- has almost no mycobactericidal effect, and in some experiments, enhances intracellular growth. Denis and Ghadirian (26) showed that GM-CSF endows human monocytes with some antituberculosis activity. Our study also has suggested that tumor necrosis factor- (TNF-) activates monocytes to kill intracellular M. tuberculosis even after infection. However, the effects of individual cytokines were relatively modest compared with the effect of supernatant observed in the present study. Synergistic effects among these cytokines or other unknown cytokines may be involved in the killing mechanism of our supernatant.

Another possibility is that CD4⁺ lymphocytes function as Class II MHC-restricted cytotoxic cells, destroying monocytes infected with mycobacteria or their products. In fact, Ottenhoff and colleagues (27) reported that recombinant 65-kD heat shock protein stimulated CD4⁺ lymphocytes to express cytolytic activity against autologous monocytes. However, since in our in vitro model, only approximately 20% of the initial adherent monocytes were detached after 7-d cultures with T-cells, it is unlikely that cytotoxic T-cells have a major role in the killing demonstrated in our system.

NK Cells and Intracellular Killing

NK cells are defined phenotypically as large granular lymphocytes (LGL) that express both the antigen CD16 and the NKH-1 (28). NK cells may be defined functionally as cells that mediate non-MHC-restricted cytotoxicity against a variety of cell targets (6, 7). The present study also demonstrated that the stimulatory effect on monocyte killing was not MHC-restricted. NK cells are considered to play a role in tumor resistance and host immunity to viral infections (6, 7). Recently, the role of NK cells in bacterial infection has become the focus of investigation. Blanchard and colleagues (9) demonstrated that human NK cells could effectively lyse monocytes that are infected intracellularly with L. pneumophila, and that this cytolytic activity was enhanced by activation of NK cells with IL-2, which resulted in LAK cells. Similarly, M. avium-complex-infected monocytes were reported to be lysed by NK cells/LAK cells (10). In the present investigation, we demonstrated that NK cells stimulated M. tuberculosis-infected monocytes to kill this intracellular organism using a different system. Different mechanisms may underlie our findings in the model of M. tuberculosis infection.

Several mechanisms have been proposed for the involvement of NK cells in the defense against microbes. First, NK cells may bind and express direct cytotoxicity for microorganisms by secreting bactericidal molecules. Penarrubia and colleagues (29) demonstrated a direct bactericidal activity of NK cells for gram-positive and gram-negative bacteria. Second, NK cells could phagocytose bacteria in a manner similar to phagocytosis by the mononuclear phagocyte system (30). However, this mechanism presumably may not be involved in the intracellular killing, although it might be associated with unique mycobacteriostatic activity by NK cells.

Recent studies have demonstrated that NK cells/LAK cells mediate the lysis of cells infected with M. avium complex (10) and Legionella (9). However, such a cytolytic activity against M. tuberculosis-infected monocytes is unlikely to play a major role since at most 20% of the adherent monocytes were detached and decreased in our system. Nevertheless, it is possible that the lysis of monocytes infected with M. tuberculosis by NK cells/LAK cells may be partly involved in killing when monocytes fail to kill intracellular mycobacteria.

Finally, a fourth possible scenario is that NK cells synthesize and secrete cytokines and/or other factors capable of activating monocytes to kill intracellular pathogens. Gomez and colleagues (31) demonstrated activation of rat and human alveolar macrophage intracellular microbicidal activity by an NK cell cytokine that was rapidly released after contact with bacteria. These NK cells demonstrated immunoregulatory action on mononuclear phagocytes and other hematopoietic cells. There are a number of cytokines reported as being released by NK cells (32). Our finding that supernatants generated from cocultivation of NK cells and infected monocytes activate monocytes to kill intracellular mycobacteria supports the possibility of this mechanism. It is unknown which cytokines are involved in activating the intracellular killing in our system. Among cytokines released by NK cells, TNF- could be responsible for the activation of monocytes in our study. Our recent preliminary study (25) suggests that TNF-, but not IFN-, stimulates human monocytes to kill intracellular M. tuberculosis using the same in vitro system. Recent studies by Bermudez and Young (33) demonstrated NK-cell-dependent mycobactericidal activity in M. avium-complex-infected monocytes and that NK-supernatant-mediated activity is partially blocked with anti-TNF- antibody. Their studies suggested that TNF-, but not IFN-, is important in stimulating mycobactericidal activity in human monocytes.

Fulton and colleagues (34) have shown that H37Ra M. tuberculosis induces IL-12, originally identified as NK cell stimulatory factor (35), which is expressed by human monocytes, and that the cellular events associated with phagocytosis are potent signal for IL-12 production. NK cells might be activated by IL-12 produced from activated monocytes infected with M. tuberculosis in our system and in turn can further upregulate monocyte intracellular killing mechanisms, cooperated with CD4⁺ T-cells.

Another question concerns how NK cells share a role with T-cells in the defense mechanism against mycobacterial infection. NK cells could be responsible for the immediate clearance of M. tuberculosis before T-cell activation. Thus, NK cells could mediate a nonspecific and early defense mechanism against this microorganism.

In conclusion, the present study indicates that PPD-specific CD4⁺ lymphocytes are mainly involved in activating human monocytes to kill intracellular M. tuberculosis in an MHC- restricted manner in PPD-positive donors, whereas in PPD-negative donors, NK cells and CD4 lymphocytes may play a major role.