Lymphocyte-Dependent Inhibition of Growth of Virulent Mycobacterium tuberculosis H37Rv Within Human Monocytes: Requirement for CD4⁺ T Cells in Purified Protein Derivative-Positive, But Not in Purified Protein Derivative-Negative Subjects¹

Tuberculosis remains a world-wide public health problem of immense proportions, as one-third of the world’s population is estimated to be infected with Mycobacterium tuberculosis (M. tb)³ (1). Cell-mediated immunity, involving the interaction of lymphocytes with infected mononuclear phagocytes, can contain M. tb, resulting in the finding that the vast majority of infected individuals do not develop active tuberculosis. Clinically, the presence of specific immunity correlates with both the development of a positive skin test to purified protein derivative of M. tb (PPD) and increased resistance to subsequent exogenous reinfection (2). Even following prolonged exposure to contagious tuberculosis patients, however, some individuals remain both PPD negative and disease free (3), suggesting that innate immunity to M. tb may exist as well. Understanding the mechanisms of human immunity that mediate protection is essential for the development of immunotherapies and new vaccines for tuberculosis; to date, these mechanisms have remained elusive, in part because of the difficulty of demonstrating effective limitation of the intracellular growth of M. tb in a human in vitro system.

Studies of infection in mice have led to substantial progress in delineating the basis of protective murine immunity to M. tb. Although CD4⁺ T cells compose the essential element of the passive transfer of protective immunity (4), studies of gene-disrupted knockout mice indicate that MHC class I pathways and, by implication, CD8⁺ cells also contribute to the development of protective immunity (5). γδ T cells accumulate in tissues of naive animals exposed to M. tb and its Ags and may represent yet another line of defense (6, 7, 8). Furthermore, IFN-γ has been identified as the primary mediator of activation of murine macrophages activation, via induction of nitric oxide, to contain intracellular bacilli (9, 10, 11). Both in vitro (12) and in vivo (13) studies suggest that TNF-α also contributes to protection by potentiating the effects of IFN-γ. The predominant role of IFN-γ in protective immunity in mice has been further emphasized by the finding that IL-12 increases the resistance of mice to M. tb infection (14), but only in animals capable of producing IFN-γ (15).

M. tb and its antigenic products elicit responses from varied human T cells subsets as well (16, 17, 18, 19, 20, 21). The roles of these cell populations in conferring protective immunity to M. tb are not clear, however. Furthermore, in vitro studies have failed to identify a dominant cytokine capable of activating human mononuclear phagocytes to contain intracellular growth of M. tb. IFN-γ itself does not limit intracellular growth of M. tb within human phagocytes (22, 23). In addition, the role of nitric oxide as a mediator of human bactericidal responses remains controversial (24, 25). The discrepancies between murine and human responses to M. tb point out the necessity of developing a human in vitro system in which mechanisms of effective cell-mediated immunity to the organism can be studied.

The interpretation of existing in vitro studies of human immunity to M. tb is complicated by several factors. Reproducible infection with virulent M. tb strains is difficult to achieve because these organisms tend to clump and are toxic to heavily infected cells. On the other hand, conclusions drawn from studies of avirulent strains may not be applicable to organisms capable of causing disease. Furthermore, although the addition of individual cytokines to cultures of infected phagocytes cannot duplicate the complex sequence of events associated with T cell/monocyte (MN) interactions, attempts to recreate this complexity by adding multiple cytokines at different times during the course of a several-day culture may render the model excessively artificial.

In this study we developed a system for achieving reproducible low level infection of human MN with virulent M. tb H37Rv that did not affect cell viability. Using this model, the ability of lymphocytes to activate antimycobacterial functions of the infected MN was assessed and compared with that of cytokines.

All subjects were healthy volunteers, aged 24 to 50 yr. For some studies, subjects who had a history of a positive tuberculin skin reaction were recruited. None of these PPD-positive subjects had a history of active tuberculosis, and none had clinical signs or symptoms of active tuberculosis at the time of the study. PPD-negative subjects had undergone tuberculin skin testing within 1 yr before entry into the study and were nonreactive. Protocols for the use of human subjects for these studies were approved by the institutional review board of the Case Western Reserve University School of Medicine and University Hospitals of Cleveland.

Peripheral blood obtained by venipuncture from healthy individuals and mononuclear cells isolated by density sedimentation using Ficoll-Hypaque (Ficoll-Paque, Pharmacia, Uppsala, Sweden) and washed three times in RPMI 1640 (BioWhitakker, Walkersville, MD). MN were then isolated by adherence to tissue culture grade 100-mm polystyrene petri dishes (Falcon 3003, Becton Dickinson Labware, Lincoln, NJ) that had been precoated with 1 to 2 ml of pooled human serum. Plates were incubated for 1 h at 37°C, and NAC were removed by gently rinsing with warmed RPMI 1640 with 10% FBS (HyClone, Logan, UT). Adherent cells were covered with cold PBS (BioWhittaker) and cooled to 4°C for 15 to 30 min, after which they were dislodged with a sterile plastic scraper (Cell Lifter 3008, Costar, Cambridge, MA). This population was found to be 99% positive by nonspecific esterase staining (no. 181-B, Sigma Chemical Co., St. Louis, MO) and henceforth is referred to as human MN. The cell population obtained by the rinsing of cells from plates of adherent MN is referred to as NAC.

NAC were depleted of CD4⁺ or CD8⁺ T cells using anti-CD4 and anti-CD8 Abs attached to magnetic beads (Dynabead M-450 CD4 and M-450 CD8, Dynal, Lake Success, NY), according to the manufacturer’s recommendations. Briefly, NAC were resuspended to a density of 10 × 10⁶/ml in RPMI with 10% FCS. The number of CD4⁺ T cells was estimated to be 50% of NAC, and Dynabeads were added in a 10:1 ratio to this calculated number of CD4⁺ T cells. For depletion of CD8 T cells, the number of CD8⁺ T cells was estimated as 35% of NAC, and a 10:1 bead to target cell ratio was again used. NAC and beads were incubated on a rotating platform for 30 min at 37°C, following which bead-absorbed T cells were removed by with a magnet (magnetic particle concentrator 120-11, Dynal). The efficacy of depletion was confirmed by two-color flow cytometry following immunofluorescent staining of samples of both nonmanipulated NAC and NAC depleted of specific subsets from each subject using fluorescent-labeled anti-CD3 and anti-CD4 or anti-CD8 mAbs (347347 CD3-phycoerythrin, 7326 CD4-FITC, 347313 CD8-FITC; Becton Dickinson, San Jose, CA). In all cases, magnetic bead depletion successfully removed >98% of the targeted T cell subpopulation.

All broth cultures of M. tb were grown in sterile Middlebrook 7H9 medium with 10% Middlebrook ADC enrichment and 0.2% glycerol. Plated cultures were grown on Middlebrook 7H10 agar with 10% Middlebrook OADC enrichment (Difco, Detroit, MI).

M. tb strain H37Rv (no. 25618, American Tissue Type Collection, Rockville, MD) was initially grown as broth culture in 1.7-l expanded surface rolling bottles (no. 2528-1700, Corning, Corning, NY) at 37°C. The initial cultures were divided into aliquots and immediately stored at −70°C.

The aliquots prepared above were used to inoculate all subsequent roller bottle cultures for use in infection of MN, so that all infections were performed with organisms that had undergone only one previous laboratory passage. To minimize clumping and allow for accurate quantification, infecting cultures of mycobacteria were processed in a manner based on that described by Schlesinger (26). Briefly, 35-ml aliquots of midlog roller bottle cultures were placed into sterile 50-ml polypropylene conical tubes (Falcon 2098, Becton Dickinson) with 5 ml of washed and autoclaved 3-mm glass beads (no. 11-312A, Fisher Scientific, Pittsburgh, PA). Bacteria were vortexed for 5 min and subsequently centrifuged at 600 rpm (50 × g) for 10 min. Mycobacteria remaining suspended in the supernatant were aliquoted into sterile 1.6-ml cryotubes (Sarstedt, Newton, NC) and stored at −70°C.

When ready for use, aliquots thus processed were thawed, and three or four sterile glass beads were placed in each cryotube. Tubes were vortexed for 5 min and centrifuged at 2000 rpm (325 × g) for 10 min. Bacteria remaining in suspension were used for infections. Quantification was performed by assessment of CFU of plated serial 10-fold dilutions of this supernatant. Aliquots of bacteria prepared from a single roller bottle culture had reproducible CFU after all processing was complete. CFU assessment of one aliquot of mycobacteria was therefore used to calculate the infecting inoculum for all cryotubes of a batch.

Isolated MN were resuspended at a density of 1 × 10⁶/ml in RPMI 1640 with 1% l-glutamine, 2% HEPES buffer, 5% fresh autologous serum, and no antibiotics, and 100 μl (10⁵ MN) was aliquoted into triplicate wells of round-bottom 96-well plates (Corning) for CFU assessment for each time point to be studied. Two additional wells were plated for assessment of cell viability at each time point (see below). Following overnight incubation to allow for readherence, supernatants were removed, and bacteria were added in 50 μl/well of RPMI with 1% l-glutamine, 1% HEPES, and 30% autologous serum. Plates were returned to a 37°C incubator for 1 h, at which time the supernatants were aspirated, and each well was washed three times with RPMI 1640 and 10% FCS to remove noningested mycobacteria. Wells were then refilled with 100 μl of Iscove’s modified Dulbecco’s medium with NaHCO₃, 25 mM HEPES, 1% l-glutamine (BioWhitakker 12-722F), and 10% noninactivated autologous serum.

At 1 h, 4 days, and 7 days following infection, supernatants of the appropriate triplicate wells were aspirated and saved, and 50 μl of lysis buffer (0.067% SDS in Middlebrook 7H9/ADC) was added to each well. Plates were then incubated at 37°C for 10 min, at which time 50 μl of PBS with 20% BSA was added. From the 100 μl of lysates thus produced, four 10-fold serial dilutions in 7H9/ADC were prepared. Serial dilutions of the supernatants were made as well. Three 10-μl aliquots of each dilution of lysate and supernatant were plated onto 7H10/OADC plates. Plates were subsequently placed in a 5% CO₂ incubator at 37°C until colonies were visible and large enough to be counted (usually 2–3 wk). Results were expressed as CFU per milliliter of lysate, which corresponded to CFU per 1 × 10⁶ cultured MN.

At the same time points (1 h, 4 days, and 7 days) that CFU assays were established, two additional wells of infected MN were collected for assessment of MN viability using the method of Nakagawara and Nathan (27). Briefly, supernatants were removed from these wells and replaced with 100 μl of napthol blue-black. Following a 10-min incubation at room temperature, MN were mixed with a pipettor, 10-μl aliquots were transferred to a hemocytometer, and nuclei were counted. Percent survival was calculated as the number of viable cells divided by the initial number of MN (1 × 10⁵) added to the well at the outset × 100.

NAC were resuspended in RPMI containing 5% fresh autologous serum and stored in a 37°C CO₂ incubator during the readherence of MN to microtiter wells. The next morning, NAC were washed and resuspended at a density of 10 × 10⁶/ml in Iscove’s modified Dulbecco’s medium with NaHCO₃, 25 mM HEPES, and 1% l-glutamine with 10% noninactivated autologous serum. NAC were then added to MN cultures immediately following the rinsing of nonphagocytosed bacteria. Addition of 100 μl of NAC (for a 10:1 NAC to MN ratio) was used to reconstitute the approximate composition of PBMC. For studies involving NAC from which CD4⁺ or CD8⁺ T cells had been depleted, resuspension was performed to achieve a concentration of 10 × 10⁶ cells/ml based on the cell count before depletion to allow for the addition of the same number of all other cellular elements as that present in the initial NAC studies. In all these studies, intracellular growth was determined by CFU assay as described above.

Supernatants were harvested from cocultures of M. tb-infected MN and NAC 6, 24, and 96 h following infection. Samples were passed through 0.22-μm pore filters (Millex-GS, Millipore, Bedford, MA) and frozen at −20°C until ready for use. IFN-γ was measured using commercially available ELISA kits (Endogen, Cambridge, MA). ODs were measured on an ELISA plate reader (Molecular Devices Corp., Menlo Park, CA) at λ = 450 nm and were analyzed using SoftMAX software (Molecular Devices). Concentrations of TNF-α were determined by sandwich ELISA as previously described (28).

Recombinant human IFN-γ was obtained from Endogen (RIFN-γ-100), and recombinant human TNF-α was purchased from Genzyme (Cambridge, MA).

Neutralizing rabbit polyclonal Abs to human IFN-γ and TNF-α were obtained from Endogen (P-300A and P-700), and neutralizing polyclonal goat Ab to IL-12 was obtained from R&D Systems (Minneapolis, MN; AB-219-NA). Species-specific IgG were used as controls in blocking studies (C-100 normal rabbit IgG (Endogen) and AB-108-C normal goat IgG (R&D Systems)).

Lymphokine production was blocked by pretreating NAC with emetine, an irreversible inhibitor of protein synthesis, as previously described by Weaver and Unanue (29).

In preliminary studies the ability of emetine to inhibit protein synthesis, as measured by incorporation of [³H]leucine, was determined. Briefly, NAC were preincubated for 2 h in leucine-free RPMI (prepared from Sigma R-7130 RPMI 1640 without l-glutamine, l-leucine, l-lysine, and l-methionine, supplemented with the missing amino acids other than l-leucine) with 10% FBS at 37°C. Equal volumes of leucine-free RPMI with 10% FBS containing emetine (E-2375, Sigma) were then added to achieve final concentrations of 0, 0.1, 1, 10, and 100 μg/ml, and NAC were incubated for an additional 1 h at 37°C. Cells were then washed and resuspended in medium containing 4% [³H]leucine (1 mCi/ml; TRK 683, Amersham Life Sciences, Cleveland. OH). Following incubation for an additional 2 h at 37°C, cells were washed with PBS and precipitated with ice-cold TCA (A322-500, Fisher). Pellets were collected on glass-fiber filters, and [³H]leucine uptake was determined using a liquid scintillation counter.

These preliminary studies indicated that 10 μg/ml concentrations of emetine were effective at reducing [³H]leucine uptake by >95%, and that 100 μg/ml of emetine had only marginally greater efficacy; 50 μg/ml was therefore chosen as the concentration for use in subsequent studies of the effects of emetine-treated NAC on the intracellular growth of M. tuberculosis. Following emetine treatment, NAC were thoroughly washed and then added to cultures of M. tuberculosis-infected MN. Growth of intracellular M. tuberculosis in the presence of emetine-treated NAC was compared with growth within MN alone and within MN cocultured with untreated NAC. To confirm the continued efficacy of emetine treatment in preventing the production of lymphokines, supernatants of these cultures were collected at 4 days of culture, and IFN-γ concentrations were determined by ELISA as described above.

Initial experiments were aimed at establishing a procedure for reproducible infection of fresh human MN with virulent M. tb in which MN viability was preserved. Following infection with a 1:1 ratio of bacteria to cells, the mean viability of infected MN from five subjects on days 0, 4, and 7 following infection were 66.0, 58.3, and 54.3%, respectively (as measured by counting of napthol blue-black-stained nuclei). The mean viability of uninfected cells of the subjects at the same time points were 68.5, 53.8, and 55.0%. The differences in viability were not statistically significant (by paired t test) at any of the time points (p = 0.663 for day 0, p = 0.430 for day 4, and p = 0.893 for day 7).

As shown in Figure 1, the initial infectious burden and subsequent patterns of intracellular growth of H37Rv following this low level infection were highly reproducible. Quantification of H37Rv within MN of five subjects is indicated as CFU per milliliter of cell lysate or CFU per 10⁶ MN. Initial intracellular burden of M. tb was highly reproducible, with a mean of 6.84 ± 1.5 × 10⁴ organisms/10⁶ MN, representing a 5 to 10% infection rate. The pattern of intracellular growth also was consistent. Mean intracellular H37Rv on day 7 was 1.72 ± 0.42 × 10⁶ organisms/10⁶ MN, indicating a mean 25-fold increase in intracellular bacteria during the course of the assay.

We next determined the effect of cytokines known to mediate MN activation on the intracellular growth of virulent M. tb following low level infection. As shown in Figure 2,A, the addition of 1000 U/ml (1.2 ng/ml) of recombinant human TNF-α to infected MN from five subjects had no significant effect on intracellular growth on either day 4 (p = 0.227, by paired t test) or day 7 (p = 0.790). Addition of 10 ng/ml (66.7 U/ml) of IFN-γ to infected MN of four individuals also failed to significantly alter growth on day 4 (p = 0.842, by paired t test), but mediated a relatively small, yet statistically significant 1.9-fold reduction in intracellular H37Rv by day 7 (p = 0.017), as shown in Figure 2 B.

Given the limited effects of TNF-α or IFN-γ on intracellular growth of M. tb, we next assessed the ability of lymphocytes to inhibit the intracellular growth H37Rv. This was performed by adding NAC to cultures of infected MN. A 10:1 ratio of NAC to MN was used to reconstitute the approximate composition of PBMC. As illustrated in Figure 3, intracellular growth was reduced significantly on days 4 and 7 following addition of NAC from both PPD-positive (Fig. 3,A) and PPD-negative subjects (Fig. 3 B). NAC from both groups of subjects mediated a decrease in intracellular M. tb between days 0 and 4. Furthermore, CFU of H37Rv on day 7 following the addition of NAC was not significantly different from that at time zero for either group. As was seen in studies of MN alone, CFU of M. tb within all simultaneously harvested supernatants remained at least 10-fold lower than those in the cell lysate (data not shown), indicating that the decrease in intracellular bacteria mediated by NAC did not reflect the release of viable M. tb from infected MN.

To evaluate the specificity of the growth inhibition of H37Rv observed above, starting populations of 10⁶ NAC were depleted of CD4⁺ or CD8⁺ cells before their addition to wells containing 10⁵ infected MN. The effect of NAC depleted of specific subpopulations on intracellular growth of H37Rv was compared with that of nondepleted NAC added in 10:1 ratio relative to the number of MN. The numbers of non-CD4⁺ or non-CD8⁺ cells present in wells containing NAC and subset-depleted NAC were therefore equivalent.

As shown in Figure 4,A, addition of CD4-depleted NAC from PPD-positive donors significantly limited intracellular growth of M. tb through the first 4 days of culture (p < 0.001, by paired t test). However, the protective effects of NAC were lost over the remainder of the assay. By day 7, CD4-depleted NAC mediated an overall 1.3-fold reduction in intracellular growth compared with that observed in cultures of infected MN alone. This difference was not statistically significant (p = 0.177). CD4-depleted NAC from PPD-negative subjects also provided significant limitation of intracellular growth of M. tb through the first 4 days of the assay (p = 0.008). In contrast to the PPD-positive subjects, however, the CD4-depleted population of PPD-negative subjects continued to inhibit M. tb, so that by day 7 CFU remained 5.8-fold lower than that observed within infected MN alone. This limitation of intracellular growth of H37Rv was not significantly different from that observed following addition of 10:1 unmanipulated NAC from the same subjects (p = 0.203; Fig. 4 B).

Results of CD8 depletion studies are indicated in Figure 5. For PPD-positive subjects (Fig. 5,A), CD8-depleted NAC mediated a statistically significant 10-fold reduction in intracellular M. tb on day 4 and a 8.4-fold reduction on day 7 compared with growth within MN alone (p = 0.001 and p = 0.002, respectively). On both days 4 and 7, reduction in intracellular growth mediated by CD8-depleted NAC was not significantly different from that mediated by untreated NAC (p = 0.175 and p = 0.169, respectively). As illustrated in Figure 5 B, CD8-depleted NAC of PPD-negative subjects mediated a 6-fold reduction in intracellular M. tb on day 4 and 2.9-fold reduction on day 7 compared with growth within MN alone (p < 0.001 and p = 0.024, respectively). The degree of inhibition of intracellular growth mediated by CD8-depleted NAC of PPD-negative subjects was not significantly different from that mediated by untreated NAC of these subjects on day 4 or day 7 (p = 0.175 and p = 0.093, respectively).

To clarify why addition of NAC was more effective than exogenous TNF-α or IFN-γ in limiting intracellular growth of M. tb, we first measured the concentrations of these two cytokines in supernatants of cocultures of NAC and infected MN. Supernatants were collected 6, 24, and 96 h following the addition of NAC to cultures of H37Rv-infected MN. As shown in Table I, high concentrations of both cytokines were present in the supernatants of cell cultures of both PPD-positive and PPD-negative donors (n = 4 for each group).

TNF-α was present 6 h following infection. In PPD-positive subjects, mean TNF-α rose somewhat through 96 h, whereas mean TNF-α of PPD-negative subjects peaked at 24 h. Both the time course and actual levels of TNF-α displayed considerable subject- to-subject variation, however, and differences between mean TNF-α concentrations of PPD-positive and PPD-negative subjects were not significant at any of the time points evaluated.

The kinetics of IFN-γ production were similar in both groups of subjects. Little of the cytokine was present at 6 or 24 h, but striking concentrations were measured at 96 h, at which time the mean IFN-γ concentration was 29.8 ng/ml for PPD-positive subjects and 17.9 ng/ml for PPD-negative subjects. Again, these differences were not statistically significant, reflecting substantial donor-to-donor variation.

The finding that supernatants of cocultured infected MN and NAC contained TNF-α and IFN-γ in much higher concentrations than those that we had initially added to M. tb-infected MN suggested that either the higher levels of cytokines or their synergistic effects could account for the greater effect of NAC than cytokines in limiting intracellular growth of M. tb. To address this possibility, we collected supernatants of cocultured NAC and M. tb-infected MN and transferred these to freshly infected MN. The effects of transfer of supernatants collected at 6, 24, and 96 h following infection from studies of three PPD-positive and three PPD-negative donors (included in Table I) are illustrated in Figure 6. Supernatants collected from PPD-positive subjects at 24 h yielded a 2.2-fold reduction in intracellular growth of H37Rv over the 7-day assay (p = 0.020, by paired t test), and 96 h supernatants from these subjects mediated a 3.9-fold reduction in growth (p = 0.010). Although statistically significant, these effects remained minor compared with the 16-fold reduction in intracellular growth mediated by the addition of NAC from these same subjects (Fig. 6,A). As indicated in Figure 6 B, the greatest reduction in intracellular growth seen with supernatants from PPD-negative subjects was a 1.9-fold decrease mediated by 96 h supernatants. However, this reduction was not statistically significant (p = 0.100) and was substantially less than the 10.3-fold reduction mediated by NAC from these subjects. For both PPD-positive and PPD-negative subjects, intracellular growth following supernatant transfer was quite reproducible despite the wide variability of TNF-α and IFN-γ concentrations measured in these supernatants.

To determine the roles of cytokines in activation of M. tb-infected MN in the presence of NAC, we assessed the effects of addition of neutralizing Abs to IFN-γ, TNF-α, and IL-12 on the ability of NAC to reduce intracellular growth of M. tb in this coculture system. Each Ab was used in two concentrations. Figure 7 illustrates the resulting CFU following the addition of the higher concentration of each Ab; for anti-IFN-γ and anti-TNF-α, this was sufficient to neutralize 5 ng/ml of specific cytokine and for anti-IL-12 sufficient to neutralize 1 ng/ml.

As illustrated in Figure 7 A, addition of NAC resulted in 4.8-fold reduction in growth of H37Rv compared with growth within MN alone by day 4 of the assay. Addition of anti-TNF-α or anti-IFN-γ reduced the magnitude of this protection to 2.9- and 2.3-fold, respectively, but these effects were not statistically significant (p = 0.288 and p = 0.239). Likewise, by day 7, addition of NAC mediated a 6.2-fold reduction in intracellular H37Rv compared with growth within MN alone. The protective effect of NAC was reduced to 4.9-fold following addition of anti-TNF-α and 2.0-fold following addition of anti-IFN-γ. Again, the changes were not statistically significant (p = 0.197 and p = 0.284, respectively).

Addition of anti-IL-12 also failed to significantly affect the ability of NAC to reduce intracellular growth of M. tb. As illustrated in Figure 7 B, on day 4 of these studies, NAC mediated a 7.8-fold reduction in intracellular M. tb compared with growth within MN alone. In cultures to which anti-IL-12 was added, NAC still mediated a 6.4-fold reduction in intracellular growth. On day 7, the overall growth reduction mediated by NAC was 4.4-fold compared with 3.8-fold following addition of anti-IL-12. The alterations in NAC-mediated reduction in intracellular growth following addition of anti-IL-12 were not statistically significant at either 4 or 7 days (p = 0.126 and p = 0.738).

As stated above, Figure 7 illustrates only the effects of the higher concentration of anti-IFN-γ, anti-TNF-α, and anti-IL-12. Of these three neutralizing Abs, a dose-response effect was observed only with anti-IFN-γ. This finding as well as the observation that many cells from many subjects produced more IFN-γ than the 5 ng/ml that was neutralized suggests that neutralizing Ab to IFN-γ could result in statistically significant increases in growth of intracellular M. tb if higher concentrations of neutralizing Ab could be used. The results of using 5-fold higher concentrations of anti-IFN-γ could not be interpreted, however, as nonspecific changes in intracellular growth of H37Rv were seen with addition of control IgG at these concentrations (data not shown).

Because our findings suggested that the effects of lymphokines could not fully explain the ability of NAC to reduce intracellular growth of M. tb, we used emetine treatment to determine whether cell-to-cell contact between NAC and infected MN could directly contribute to containment of the organism. Emetine-treated NAC from two PPD-positive individuals were added to M. tb-infected MN, and subsequent growth was compared with that observed within MN alone and in infected MN cocultured with untreated NAC. The IFN-γ concentration in day 4 supernatants of cocultures containing emetine-treated NAC was compared with that in cultures containing untreated NAC to confirm the continuing effects of emetine treatment.

As illustrated in Figure 8, the IFN-γ concentration in cultures containing emetine-treated NAC were reduced by >90% for each of the two subjects. In each experiment, emetine-treated NAC were able to partially reduce the intracellular growth of H37Rv. For subject 1 (Figure 8,A), emetine-treated NAC were 45% and 43% as effective as untreated NAC at limiting intracellular growth at M.tb at days 4 and 7, respectively. For subject 2 (Figure 8 B), the limitation of growth provided by emetine-treated NAC was 28% as great as that mediated by untreated NAC on day 4, and 39% as great on day 7. These results suggest that NAC can contribute to control of growth of M. tb within human MN by contact-dependent mechanisms at both early and later phases of this assay.

In vitro studies of human immunity to M. tb have been largely modeled on the work of Crowle (30), who used a CFU-based assay to determine the ability of soluble mediators to modulate intracellular growth of the organism within infected MN. In contrast to findings in studies of murine cells, both Crowle (22) and others (23) found IFN-γ to be an ineffective at activating human MN to contain intracellular M. tb. Likewise, although vitamins D and A have some ability to activate M. tb-infected human MN (31, 32, 33), and TNF-α has been shown to contribute to in vitro containment of avirulent M. tb (34), none of these mediators limits the growth of the organism within human MN with an efficacy comparable to that of IFN-γ in the murine system. Potential reasons for the difficulty in demonstrating inhibition of intracellular growth of M. tb in human in vitro studies include 1) use of high bacteria-to-cell ratios, which may overwhelm the antimicrobial capacities of activated MN; 2) a requirement for higher concentrations of cytokines or synergistic effects of multiple cytokines to activate infected MN; or 3) the necessity of cell-cell interactions in the containment of intracellular M. tb.

We developed a model of low level infection of human MN with the virulent M. tb strain H37Rv in which the viability of infected cells was not compromised and in which both initial infectious burden and subsequent intracellular growth of the organisms were highly reproducible. This in vitro model allowed us to measure directly the capacities of both lymphocytes and their soluble mediators to inhibit the growth of M. tb within MN and thus to address the issues raised above. NAC from both PPD-positive and PPD-negative subjects were more effective at limiting the intracellular growth of H37Rv than either recombinant cytokines or transferred supernatants. Depletion of CD4⁺ T cells eliminated the growth-inhibiting capacity of NAC from PPD-positive individuals, but did not diminish the inhibitory effects of NAC from PPD-negative subjects. Depletion of CD8⁺ T cells did not affect the ability of NAC from either group of subjects to contain M. tb. NAC-mediated limitation of intracellular growth of H37Rv was not significantly reduced by neutralizing Abs to IFN-γ, TNF-α, or IL-12. Furthermore, following inhibition of protein synthesis, and thus of cytokine production, by treatment with emetine, NAC retained approximately 40% of their capacity to contain intracellular growth of M. tb.

Positive PPD skin test reactivity is considered a marker of the development of specific immunity to M. tb and is associated with protection from new infection (2). A central role for CD4⁺ T cells in this acquired immunity has been suggested by clinical observations. For example, the self-limiting course of tuberculous pleurisy is associated with local accumulation of CD4⁺ T cells (35), and HIV-infected individuals are extremely susceptible to the development of tuberculosis (36, 37). Our determination that CD4⁺ T cells were essential to the ability of NAC from PPD-positive subjects to inhibit the intracellular growth of M. tb therefore supports the relevance of this model. The finding that the CD4⁺ T cells contributed to NAC-mediated growth limitation of intracellular H37Rv predominantly during days 4 to 7 of the assay is consistent with the requirement for activation of these memory cells to fully develop Ag-specific responses.

The finding that NAC from PPD-negative individuals were also capable of mediating reduction in intracellular M. tb has some precedent, in that lymphocytes from PPD-negative subjects can display in vitro responses to M. tb and its Ags (38). The observation that some individuals remain both PPD negative and disease free despite prolonged exposure to contagious tuberculosis patients (3) further suggests that nonspecific innate immunity to M. tb may exist. Although such “natural” immunity would be anticipated to be less effective than specific, acquired immunity, it is possible that the somewhat arbitrary nature of a 7-day assay biases toward early and nonspecific defenses. CD8⁺ T cells are not the mediators of early, CD4-independent containment of H37Rv in our assay, since depletion of these CD8⁺ cells did not diminish the protective effects of NAC from either PPD-positive or PPD-negative subjects. Although M. tb can induce blastogenesis in γδ T cells from PPD-negative subjects (18, 19, 20), these in vitro responses characteristically develop over several days and therefore are unlikely to explain our observations. Another possibility is that this early protection is mediated by NK cells. In vitro cytokine responses of PPD-negative subjects to M. tb have been attributed to this cell population (21). Although NK cells can contribute to immunity against M. avium (39), their role in infection with M. tb remains unclear.

We confirmed previous observations that soluble mediators are themselves relatively ineffective activators of human MN to contain M. tb. Recombinant TNF-α had no effect, and IFN-γ had a relatively minor impact on the intracellular growth of H37Rv. Although supernatants of cocultured M. tb-infected MN and NAC from both PPD-positive and PPD-negative subjects contained nanogram quantities of both TNF-α and IFN-γ, these supernatants also were significantly less effective at limiting intracellular growth of M. tb than were NAC. Thus, the limitations of soluble mediators in containment of M. tb within human MN cannot be attributed to high (and nonphysiologic) infectious burdens. Furthermore, neither higher concentrations of TNF-α and IFN-γ nor combinations of cytokines, such as those present in transferred supernatants, are sufficient to fully mediate antimycobacterial MN activation.

The findings described above indicate that the direct interaction of lymphocytes with M. tb-infected MN is required to fully activate antimycobacterial defenses of the infected cells. The lack of effects of neutralizing Abs to IFN-γ, TNF-α, or IL-12 in the coculture system suggests that the role of cell-to-cell contact is not predominantly one of maximizing the production of these cytokines or of priming MN responsiveness to them. Alternatively, the requirement for cells could reflect a need for local release of cytokines at the cell-cell interface (40), which could be relatively shielded from the effects of neutralizing Abs. The finding that emetine-treated NAC retain a substantial proportion of their ability to contain intracellular growth of M. tb, however, suggests two other, possibly overlapping, mechanisms: direct, contact-mediated activation of antimycobacterial MN function by NAC, and NAC-mediated MN cytotoxicity.

Contact-mediated containment of intracellular organisms has been described previously, most notably with regard to the intracellular parasite Leishmania major. The most effective in vitro killing of this pathogen requires the presence of lymphocytes (41) and is mediated by the membrane-bound TNF expressed on the surface of Ag-specific CD4⁺ T cells (42, 43). Membrane-bound TNF has been implicated in antimycobacterial defenses as well (using Calmette-Guérin bacillus in an in vitro murine system) (44). Although neutralizing anti-TNF Abs did not reduce the protective effect of NAC in our studies, it remains possible that membrane-bound TNF does play a role in human immunity to M. tb, but that the ability of the neutralizing Ab to block this effect is limited by interaction of the Ab with the large amount of secreted TNF-α in our cultures. On the other hand, several other pairs of surface molecules have been implicated in contact-mediated activation of MN by lymphocytes. The interactions of lymphocyte surface CD2 with MN surface LFA-3 (45), of CD40 ligand with CD40 (46, 47, 48, 49), and of CD23 with CD11b and CD11c (50, 51) provide alternate means by which NAC may directly activate antimycobacterial MN effector functions.

Recent studies also suggest that NAC-mediated cytotoxicity could account for cytokine-independent killing of intracellular M. tb. Various populations of M. tb-reactive T cells are cytotoxic to infected MN (16, 52, 53, 54, 55, 56), and nonspecific immune responses, such as those provided by NK cells, may also be mediated by cytotoxicity. Killing of MN both by apoptosis (57) and by perforin-mediated cytotoxicity (58) may result in killing, rather than release, of intracellular mycobacteria. Either of these possibilities could be consistent with our findings, as apoptosis can be mediated by interaction of NAC surface Fas ligand with the MN surface molecule Fas. Likewise, emetine-treated NAC could still mediate MN cytotoxicity via release of preformed toxic substances, such as perforin.

The findings of this study have several potential applications to the further understanding of the nature of protective human immunity to M. tb. The quantifiable assay of growth of virulent M. tb generally allows for correlation of in vitro lymphocyte responses such as cytotoxicity with the ability of these cells to limit intracellular growth of the organism. The significant contribution of CD4⁺ T cells of PPD-positive subjects to the reduction of intracellular growth of M. tb indicates that this in vitro model may be useful in determining the contribution of Ag-specific immune responses to M. tb to this protection. For example, the ability of cell lines specific for candidate vaccine components to limit intracellular growth of M. tb may provide a direct method to evaluate the protective potential of these Ags. Further clarification of the role of CD4⁻CD8⁻ lymphocytes, particularly in PPD-negative subjects, in mediating control of intracellular H37Rv may provide insight into native immunity to M. tb. In addition, the evidence that NAC provide contact-mediated limitation of intracellular growth of virulent M. tb within human MN suggests that investigation of the contributions of specific cell-surface ligands may help clarify the mechanistic basis for protective human immunity to M. tb.

The authors thank Drs. Elizabeth Rich, Hiroe Shiratsuchi, and Zahra Toossi of Case Western Reserve University for helpful discussions regarding the establishment of the infection model. They also thank Dr. David Kaplan of Case Western Reserve University for suggestions concerning the use of emetine, and Mr. Kevin Brill for technical assistance.