Abstract
To define the immunopathologic mechanism underlying pulmonary Mycobacterium avium—intracellulare complex (MAC) disease in patients without AIDS, the ability of CD4+ and yd T cells to induce growth inhibition of MAC in monocytes was compared between patients and healthy control subjects. T cell—dependent growth inhibition and production of interferon-γ and macrophage colony—stimulating factor decreased in patients. CD4+ T and γδ T cells from patients were equally defective in inducing anti-MAC activity. The combination of these cytokines restored the ability of patients' T cells to control MAC growth. In experiments with allogeneic cocultures of γδ T cells and infected monocytes from patients and control subjects, healthy control T cells could augment growth inhibition of MAC in monocytes from patients, whereas patients' T cells could not, even in the presence of healthy control monocytes. These results indicate that the defect in T cells may be associated with impaired protective immunity against MAC in these patients.
Disseminated disease caused by Mycobacterium avium—intracellulare complex (MAC), which is often a fatal form of infection in patients with AIDS, has been on the decrease with the introduction of highly active antiretroviral therapy [1]. Although the organisms, commonly found in soil, water, and house dust, are normally regarded as opportunistic bacterial pathogens, they also demonstrate pathogenicity at a localized pulmonary site in apparently normal persons without signs of systemic immunologic diseases, including AIDS. The immunopathologic mechanism underlying pulmonary MAC disease in such patients without AIDS has not been well clarified. MAC can survive and even multiply within normal macrophages, and little is known about the mechanisms controlling these facultative intracellular bacteria. Most persons are very resistant to MAC.
There is reportedly substantial variation in the inhibitory effect of cytokines on the growth of MAC, and it remains unclear whether a single cytokine or a combination of them renders macrophages mycobactericidal in humans. Interferon (IFN)—γ induces mouse macrophages to inhibit MAC growth [2, 3], but, when tested alone in human macrophages, this cytokine may exhibit the opposite effect [4, 5]. Macrophage colony—stimulating factor (M-CSF), initially described as a hematopoietic growth factor stimulating survival, proliferation, and differentiation of cells of the mononuclear phagocyte lineage, is produced by both T cells and macrophages [6–8] and is known to play an important role in killing microbial pathogens such as Candida albicans [9, 10], Cryptococcus neoformans [11], and Listeria monocytogenes [12]. The cytokine alone or in combination with IFN-γ modulates the antimicrobial activity of human alveolar macrophages against MAC [13].
The acquired cellular immune response against mycobacteria is widely accepted to consist of T cells and macrophages. Recent reports with in vitro systems that used human cells demonstrated that normal lymphocytes activated monocytes to limit the intracellular growth of M. tuberculosis [14, 15]. Defects in T cell [16–18] and monocyte function [19, 20] have been demonstrated in human mycobacterial diseases.
Among T cell subsets, the CD4+ T cell is the most dominant in the immune response to MAC [21]. The role of CD4+ T cells in host defense against MAC infection is supported by the observation that MAC disease is much more prevalent in patients with human immunodeficiency virus (HIV) infection and advanced CD4+ T cell depletion. In addition to the CD4+ T cell, the T cell receptor (TCR) γδ—bearing T cell (γδ T cell) is another subset receiving particular attention in the immune responses to mycobacteria. Human γδ T cells are activated by M. tuberculosis [22] and may exhibit effector functions similar to those of CD4+ T cells [23], resulting in a better-integrated defense network against mycobacterial infection [24]. The role of γδ T cells in the immunopathology of MAC infection has been studied in mice [25], and these cells increase in the peripherals and liver in patients with AIDS and disseminated MAC disease [26, 27], although it is unclear how this T cell subset is involved in the human immune response against MAC.
To determine the immunopathologic mechanism underlying MAC disease in patients without AIDS, we compared the ability of highly purified CD4+ T cells and γδ T cells to activate monocytes to inhibit growth of MAC between patients and healthy control subjects. We also assessed the modulation of IFN-γ and M-CSF on T cell—dependent growth inhibition of MAC. To clarify whether T cells or monocytes were responsible for the decline in inhibiting MAC growth in patients, we conducted coculture experiments of γδ T cells and allogeneic monocytes from patients and control subjects.
Materials and Methods
Subjects
Eight patients (mean age ± SD, 62 ± 14 years old) with newly diagnosed active pulmonary MAC disease and 8 agematched healthy volunteers were studied. The diagnosis of pulmonary MAC disease was based on the criteria published by American Thoracic Society for the diagnosis of disease caused by nontuberculous mycobacteria [28]. No patients were HIV-infected or had a defined immune deficiency, history of immunosuppressive therapy, underlying pulmonary diseases, or metabolic disturbance such as diabetes mellitus. Blood samples from patients were collected before treatment for MAC was initiated. No healthy subjects had pulmonary diseases or were receiving any medications at the time of the study.
Mycobacteria
M. intracellulare 31F0393T was grown in Middlebrook 7H9 broth (Difco Laboratories, Detroit) with 10% Middlebrook albumin dextrose catalase enrichment and 0.2% glycerol, and frozen stocks were prepared as described elsewhere [29, 30]. Bacterial counts and viability were determined by light microscopy and by counting colony-forming units on 7H10 medium. MAC stocks were tested periodically for viability and with a polymerase chain reaction microwell plate hybridization kit (Amplicor Mycobacteria; Roche Diagnostics, Tokyo) to ensure purity of the MAC stocks. Before infection, bacteria were washed 3 times in RPMI 1640 and were uniformly resuspended by sonicating for 20 s, to disrupt clumps and to obtain a single-cell suspension. Adequate dispersion of bacteria was ascertained microscopically.
Isolation of peripheral blood mononuclear cells (PBMC) and monocytes
Peripheral blood was obtained from patients and healthy volunteers, and mononuclear cells were isolated by density centrifugation over sodium diatrizoate—hypaque gradients. Monocytes were isolated as described elsewhere [29]. Briefly, PBMC were incubated on plastic tissue-culture dishes precoated with pooled human serum, nonadherent cells were removed, and plasticadherent cells (mean, 96.6% monocytes by nonspecific esterase staining) were collected by scraping with a plastic policeman.
Preparation for nonstimulated and prestimulated cells
Cells directly obtained from PBMC without preexposure to MAC were referred to as nonstimulated cells. Prestimulated cells were prepared by incubating PBMC (2 × 106 cells/2-mL well) in 24-well plates (Corning Costar, Cambridge, MA) with live MAC (5 × 106 bacteria/mL) and by allowing MAC-specific T cells to expand. Culture medium consisted of RPMI 1640 supplemented with 10% pooled human serum, 20 mM HEPES, and 2 mML-glutamine. No antibiotics were used. After 7 days of culture, nonadherent cells were gently collected, and viable cells were harvested by density sedimentation over sodium diatrizoate—hypaque gradients.
Purification of CD4+ T cells and γδ T cells
Nonstimulated CD4+ T cells were negatively selected from nonstimulated cells, and prestimulated CD4+ T cells or γδ T cells were selected from prestimulated cells. Cells were enriched by negative selection with magnetic beads coated with monoclonal antibodies (Dynal, Great Neck, NY). For CD4+ T cell enrichment and CD8+ T cell, γδ T cell, and NK cell depletion, cells were treated first with TCR-51 (Endogen, Wobum, MA) and anti—human CD56 (PharMingen, San Diego), followed by goat anti—mouse IgG—coated beads and anti-CD8—coated beads. For γδ T cell enrichment and CD4+ T cell, CD8+ T cell, and NK cell depletion, cells were treated first with anti—human CD56, followed by goat anti-mouse IgG-coated beads and anti-CD4— and anti-CD8—coated beads. Antibody-coated beads were used at a 10:1 bead-to-cell ratio based on the estimated number of T cells from each T cell subset (CD4+, CD8+, CD56+, and γδ T cells) present in non- or prestimulated cells.
Purity of the negatively selected T cell populations was assessed by 2-color flow cytometry with the use of phycoerythrin-conjugated anti-CD3 and anti-CD56 and fluorescein isothiocyanate—conjugated anti-CD4, anti-CD8, anti-CD14, anti-CD19, and anti—TCR-γδ1. In all cases, the purity of the selected subsets exceeded 95% for both nonstimulated CD4+ T cells and prestimulated CD4+ T cells or γδ T cells, with <5% contamination by the depleted subsets. The 14%–19% CD3∼ cells consisted mostly of CD19+ B cells (>95%), and no statistically significant difference was seen in the percentage of B cells between non- and prestimulated T cell populations.
Infection of human monocytes with MAC
Isolated monocytes were resuspended in antibiotic-free RPMI 1640 medium with 20 mM HEPES, 2 mML-glutamine, and 5% unheated autologous serum. Monocytes (105) were aliquoted into 96-well flat-bottom plates (Corning Costar) in a volume of 100 µL and were incubated overnight. Culture supernatants were removed, and readherent monocytes were infected with bacteria in 50 µL/well at a bacterium-to-monocyte ratio of 5:1. Plates were then incubated for 1 h at 37°C and gently washed 4 times with prewarmed medium, to remove nonphagocytized bacteria. The monocytes were then cultured in triplicate wells in 100 µL of medium with 10% nonheated autologous serum for ⩾4 days in the presence or absence of cytokmes. Cytokmes were added to cultures immediately after infection and replenished on day 1.
Addition of CD4+ T cells or γδ T cells
Nonstimulated CD4+ T cells and prestimulated CD4+ and γδ T cells were resuspended in medium with 10% heat-inactivated serum and added back to monocyte cultures (at 5:1 and 10:1 T cell—to-monocyte ratios) in triplicate wells immediately after infection. At 1 and 4 days after addition of T cells, the number of bacteria in monocytes was determined. Monocyte viability at day 4 was determined in duplicate wells by the method of Nakagawara and Nathan [31]. The mean percentages of viable and adherent monocytes at day 4 in cocultures of infected monocytes with prestimulated CD4+ or γδ T cells were 69% ± 11% and 71% ± 8%, respectively, for control subjects and 73% ± 10% and 67% ± 12%, respectively, for patients, compared with the initial 105 monocytes. These percentages were not significantly different from those in cultures of MAC-infected monocytes alone(74% ± 14%forcontrol subjects and 71% ± 6% for patients).
Colony-forming unit assay
The number of MAC at each time point in monocytes was determined by using a colony-forming unit assay, as described elsewhere [19, 32]. At time 0 (1 h after infection) and at 1 and 4 days after infection, culture supernatants were aspirated and saved. Lysing solution (50 µL) containing 0.05% SDS in 7H9 medium was added to each well. Plates were incubated at room temperature for 10 min, and 50 µL of PBS with 20% bovine serum albumin was added. Cell lysates were diluted serially, and three 10-µL aliquots of each dilution wereplated onto Middlebrook 7H10 agar plates. Plates were incubated at 37°C for 2 weeks until bacterial colonies were visible. Results were expressed as colonyforming units per milliliter of cell lysate, which represented colonyforming units associated with 106 monocytes. To compare the inhibition of MAC growth in monocytes under various conditions, the percentage of inhibition was calculated by the following equation: 100 × (mean colony-forming units, infected monocytes alone at day 4) — (mean colony-forming units, infected monocytes + T cells and/or cytokines at day 4)/(mean colony-forming units, infected monocytes alone at day 4).
Recombinant human cytokines and monoclonal antibodies
Recombinant human IFN-γ and M-CSF were purchased from Genzyme (Cambridge, MA). Mouse monoclonal antibodies to human IFN-γ (1598–00) and M-CSF (MAB216) were obtained from Genzyme and R&D Systems (Minneapolis), respectively. Isotypic mouse IgG2a (PM-33031A) was obtained from PharMingen.
Cytokine production
Supernatants of cocultured MAC-infected monocytes and T cells at 96 h after infection were harvested and filtered through a 0.2-µm-pore filter (Millipore, Bedford, MA). Additional supernatants in some cultures were collected at 12, 24, and 48 h. Samples then were aliquoted and were stored at −80°C. Concentrations of IFN-γ and M-CSF in supernatants were determined by commercially available ELISA kits (R&D Systems). Absorbance was measured on a microplate reader (model 3550; Bio-Rad, Hercules, CA) at 450 nm.
Allogeneic coculture of prestimulated γδ T cells and infected monocytes from patients and age-matched healthy control subjects
Prestimulated γδ T cells from the patient were added to MAC-infected monocytes from the age-matched control subject, and, at the same time, γδ T cells obtained from the control subject were added to infected monocytes from the patient. Results are represented as mean ± SD of 8 pairs of patients and age-matched control subjects. Growth inhibition of bacteria was measured as described above. The paired patient and control subject had different HLA-A, -B, -C, and -DR alleles.
Statistical analysis
Statistical analysis was done by paired and unpaired Student's t tests, and P < .05 was considered significant.
Results
Effect of prestimulated CD4+ T cells or γδ T cells on MAC growth in monocytes from patients and healthy control subjects
Initial experiments were undertaken on growth inhibition of MAC in monocytes without T cells. No significant difference was seen in initial bacterial burden and growth patterns of MAC in monocytes between patients and control subjects (1 h after infection, 0.71 ±0.12 × 105 vs. 0.67 ± 0.13 × 105 cfu, P = .496; 1 day after infection, 0.62 ± 0.13 × 105 vs. 0.61 ± 0.18 × 105 cfu, P = .902; 4 days after infection, 2.40 ± 0.78 × 105 vs. 2.61 ± 0.81 × 105 cfu, P = .660; n = 8 for each subject group).
As shown previously, neither IFN-γ nor M-CSF inhibited MAC growth in patients' monocytes (at 4 days, 3.38 ± 1.43 × 105 cfu, P = .111, and 2.86 ± 0.99 × 105 cfu, P = .325, respectively) or in control monocytes (at 4 days, 3.60 ± 1.39 × 105 cfu, P = .106, and 2.28 ± 0.81 × 105 cfu, P = .469, respectively), compared with cultures of infected monocytes alone. The combination of IFN-γ and M-CSF modestly inhibited MAC growth in monocytes from patients (1.46 ± 0.68 × 105 cfu; 44.1% inhibition at 4 days) and control subjects (1.26 ± 0.73 × 105 cfu; 47.5% inhibition at 4 days), although the effects were not statistically significant, compared with cultures of infected monocytes alone (P = .079 and P = .061). Concentrations of IFN-γ and M-CSF added to cultures were 200 and 30 U/mL, and any other combinations of doses (100–500 U/mL for IFN-γ and 10–50 U/mL for M-CSF) resulted in less inhibition (data not shown).
We next examined the effect of purified CD4+ T cells and γδ T cells on intracellular growth of MAC in patients and control subjects. As shown in figure 1, growth of MAC in monocytes did not significantly decrease on either day 1 or 4 after the addition of either prestimulated CD4+ T cells from patients (1.62 ± 0.67 × 105 cfu; 37.9% inhibition at 4 days; P = .102) or γδ T cells from patients (1.82 ± 0.58 × 105 cfu; 30.3% inhibition at 4 days; P = .116), compared with cultures of infected monocytes alone. In contrast, both prestimulated CD4+ T cells and γδ T cells from control subjects mediated significant decreases in intracellular MAC growth on day 4 (0.50 ± 0.28 × 105 cfu; 79.2% inhibition; P = .004 and 1.01 ± 0.37 × 105 cfu; 57.9% inhibition; P = .01, respectively), compared with cultures of infected monocytes alone. The difference in inhibition between CD4+ T cells and γδ T cells for control subjects was not statistically significant (P = .057). These data were derived from experiments with a 5:1 T cell—to-monocyte ratio and were not significantly different from those with a 10:1 ratio (data not shown). We therefore conducted subsequent experiments with a 5:1 ratio.
Effect of prestimulated CD4+ T cells or γδ T cells on the growth of Mycobacterium avium-intracellulare complex (MAC) in monocytes (MN) from patients and healthy control subjects. Autologous CD4+ T cells or γδ T cells prestimulated with MAC for 7 days were added back to infected monocytes at 5:1 T cell—to-monocyte ratio immediately after infection. Both CD4+ T cells and γδ T cells from control subjects showed significant inhibitory effect on MAC growth in monocytes at day 4. aP = .004 and bP = .01, compared with cultures of infected monocytes alone at day 4. Data are mean ± SD of 8 subjects from each group.
Nonstimulated CD4+ T cells from patients and control subjects did not mediate significant growth inhibition on day 4 (1.78 ± 0.65 × 105 cfu, 31.8% inhibition, and 1.81 ±0.59 × 105 cfu, 24.6% inhibition, respectively), compared with infected monocytes alone.
The numbers of MAC in simultaneously harvested supernatants were 4%–14% of those in monocytes lysates. In general, no significant differences were seen in bacterial numbers in supernatants under different conditions (data not shown), suggesting that observed effects were not due to increased release of viable MAC into the extracellular milieu but were due to control of intracellular bacterial growth.
Effect of prestimulated CD4+ T or γδ T cells in combination with IFN-γ and/or M-CSF on growth of MAC in monocytes from patients and healthy control subjects
We next determined whether IFN-γ and/or M-CSF modulated the ability of prestimulated CD4+ T cells or γδ T cells to limit the growth of MAC in monocytes. As shown in table 1, addition of either IFN-γ or M-CSF to CD4+ T cells did not significantly reduce colonyforming units in monocytes isolated from patients. The combination of cytokines added to cocultures of patients' CD4+ T cells and infected monocytes significantly reduced colony-forming units, compared with colony-forming units of infected monocytes alone (P = .006), but also compared with infected monocytes plus CD4+ T cells (P = .022). These observations suggest that IFN-γ and M-CSF have a synergistic effect on CD4+ T cell—mediated growth inhibition. Significant growth inhibition was also observed when this cytokine combination was added to cocultures of patients' γδ T cells and infected monocytes (P = .008, compared with infected monocytes alone).
Effect of prestimulated CD4+ T or γδ T cells in combination with interferon (IFN)-γ and/or macrophage colony—stimulating factor (M-CSF) on growth of Mycobacterium avium—intracellulare complex (MAC) in monocytes from patients and healthy control subjects.
In healthy subjects, addition of either IFN-γ or M-CSF or their combination induced no further inhibition, compared with infected monocytes plus CD4+ T cells or γδ T cells. The concentrations of IFN-γ and M-CSF added were the same as above, and growth inhibition in other combinations of concentrations of the cytokines was not statistically significant (data not shown).
IFN-γ and M-CSF production in cocultures of prestimulated T cells and MAC-infected monocytes
Given the observation that the combination of IFN-γ and M-CSF augmented growth inhibition of MAC in cocultures of prestimulated T cells and infected monocytes from patients, we studied production of these cytokines in the cocultures. The production of both IFN-γ and M-CSF in patients and control subjects increased rapidly during the first 24 h and then did not change throughout the rest of the culture period (data not shown). We therefore studied the difference in cytokine production after 96 h.
As shown in table 2, IFN-γ production in cocultures of infected monocytes and either prestimulated CD4+ T or γδ T cells from patients was significantly lower than that for control subjects. Infected monocytes alone from both groups produced no significant amount of IFN-γ.
Interferon (IFN)—γ and macrophage colony—stimulating factor (M-CSF) production by Mycobacterium arium-intracellulare complex (MAC)—infected monocytes and/or prestimulated T cells.
MAC-infected monocytes produced M-CSF at a similar level for patients and control subjects. Addition of either prestimulated CD4+ T cells or γδ T cells from patients to infected monocytes induced no further production of M-CSF, in marked contrast to the significant increase in cytokine production with the addition of control CD4+ or γδ T cells.
The production of IFN-γ and M-CSF in cocultures of nonstimulated CD4+ T cells and infected monocytes from control subjects (17.0 ± 5.22 ng/mL for IFN-γ and 19.1 ± 4.99 ng/mL for M-CSF) was significantly lower than that of prestimulated CD4+ T cells and infected monocytes. These levels were not significantly different from those in patients (15.1 ± 3.82 ng/mL for IFN-γ and 16.9 ± 3.50 ng/mL for M-CSF). Prestimulated CD4+ T cells and uninfected monocytes from control subjects produced much less IFN-γ and M-CSF (1.02 ± 0.75 ng/mL and 2.27 ± 0.94 ng/mL) than did prestimulated CD4+ T cells and infected monocytes.
Effect of neutralizing antibodies to IFN-γ and M-CSF on prestimulated CD4+ T cell—mediated growth inhibition of MAC in monocytes
To confirm the effect of IFN-γ and M-CSF on prestimulated CD4+ T cell—mediated growth inhibition, we undertook a study with neutralizing antibodies to these cytokines. As shown in table 3, addition of both anti—IFN-γ and anti—M-CSF antibodies partially abrogated the inhibitory effect of prestimulated CD4+ T cells from control subjects (1.49 ± 0.60 × 105 cfu with antibodies vs. 0.50 ± 0.28 × 105 cfu without antibodies; P = .016). In contrast to control subjects, these antibodies did not show a significant effect in patients.
Effect of neutralizing antibodies to interferon (IFN)—γ and macrophage colony—stimulating factor (M-CSF) on CD4+ T cell—mediated growth inhibition of Mycobacterium avium—intracellulare complex (MAC) in monocytes.
Patients' T cells responsible for impaired growth inhibition of MAC in monocytes
Having demonstrated impaired growth inhibition of MAC in cocultures of prestimulated T cells and infected monocytes from patients, we next determined whether T cells or monocytes were responsible for this impaired growth inhibition. γδ T cells and monocytes from patients and control subjects were cocultured in various combinations, and growth inhibition of MAC was measured. γδ T cells were used for this assay because of their major histocompatibility complex—unrestricted property [30, 33, 34]. γδ T cells from patients did not provide significant growth inhibition, even in monocytes from control subjects, whereas γδ T cells from control subjects showed significant growth inhibition in monocytes from either patients or control subjects, indicating that γδ T cells from patients were defective in their ability to inhibit MAC in monocytes (figure 2).
![Patient γδ T cells did not provide significant inhibition of Mycobacterium avium-intracellulare complex (MAC) growth even in control monocytes. Prestimulated γδ T cells (T+) and allogeneic monocytes (MN) from patient (Pts) and age-matched healthy control subjects (Ctl) were cocultured for 4 days (Pts or Ctl T plus Pts or Ctl MN). Percentage of inhibition was calculated as follows: 100 × (mean colonyforming units [cfu], infected monocytes alone at day 4) — (mean cfu, infected monocytes plus T cells at day 4)/(mean cfu, infected monocytes alone at day 4). Results are mean ± SD of 8 pairs of patients and agematched control subjects. aP < .05, compared with control T cells and control monocytes; bP < .05, compared with control T cells and patient monocytes.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/jid/182/6/10.1086_317601/2/m_182-6-1664-fig002.jpeg?Expires=1709964304&Signature=Jjv67qb7d6V5IZqH3QJ8R7h0sdlZwIwnx~QzwciXoaPpQkgaT0K4ZIy8sieeKbQza57pVWQpPou1vV4OyDMEoY~050ai3C76UU8YkKr4i3je1M3pRSvNMbvYjuBcmsYC42fEttrW8K4ErQ3HaXwnxPqVSSMIwIXD-VCbtvqKthOTegkL73cXFJZ-yFbZC9PzT6tv9LYG4G5GiCXVHiwQg8DA9HPiWobGZ5vlcb1X7chwk9U7GvFGgE7D3caN-vUobcA8LFEXinRaOqTg-17I7F6qW7ZaW007nSTXDri0~yA9GMQCOpG87V9~YB~ZNhBkA6cj36dceg0Kj6-yQxaesQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Patient γδ T cells did not provide significant inhibition of Mycobacterium avium-intracellulare complex (MAC) growth even in control monocytes. Prestimulated γδ T cells (T+) and allogeneic monocytes (MN) from patient (Pts) and age-matched healthy control subjects (Ctl) were cocultured for 4 days (Pts or Ctl T plus Pts or Ctl MN). Percentage of inhibition was calculated as follows: 100 × (mean colonyforming units [cfu], infected monocytes alone at day 4) — (mean cfu, infected monocytes plus T cells at day 4)/(mean cfu, infected monocytes alone at day 4). Results are mean ± SD of 8 pairs of patients and agematched control subjects. aP < .05, compared with control T cells and control monocytes; bP < .05, compared with control T cells and patient monocytes.
Discussion
Pulmonary MAC disease may occur in persons who appear to be normal—that is, who show no signs of systemic immunologic diseases, including AIDS. Our study suggests that the decline in growth inhibition of MAC in monocytes from such patients without AIDS may be caused by an inability of their T cells to control MAC growth. Because these patients had normal CD4+ T cell counts, T cell dysfunction specific for MAC could have been a factor here.
As reported elsewhere [32, 35], adding IFN-γ and M-CSF to patients' monocytes did not significantly inhibit MAC growth. Adding these cytokines to cocultures of T cells and infected monocytes from patients markedly enhanced inhibitory activity. These results suggest that patients' T cells played a role in inhibiting MAC growth and that T cell function was only partly impaired—that is, some T cell functions remained. In this selective T cell dysfunction, it may be that production of other cytokines, including interleukin (IL)—2 and tumor necrosis factor (TNF)—α, involved in inhibiting MAC growth [35] is maintained. It was reported that IL-2 mRNA expression in PBMC is enhanced more in patients with active pulmonary tuberculosis than in healthy persons [36]. Unlike patients with AIDS, who show significant decreases in CD4+ T cell counts, patients without AIDS rarely manifest MAC disease in a disseminated form, probably, in part, because of their selective T cell dysfunction. In healthy control subjects, infected monocytes uniformly activated T cells to effectively inhibit MAC growth, and this effector function was supported, at least in part, by IFN-γ and M-CSF. The critical difference in cytokine production between patients and healthy control subjects may be one of the determinants of susceptibility to MAC infection.
In our study, γδ T cells were as significantly effective as CD4+ T cells in inhibiting MAC growth in healthy control subjects. A previous report demonstrated that, after M. tuberculosis stimulation, γδ T cells and CD4+ T cells were equivalent in effector functions, such as IFN-γ production and cytotoxic effects on M. tuberculosis—infected macrophages [23]. Accordingly, once these 2 T cell subsets were activated by MAC, they seemed to show effector functions equally. A strong correlation between the absence or loss of Vγ9/Vδ2 γδ T cells and manifestations of pulmonary tuberculosis has been demonstrated [37]. We studied patients without AIDS who had pulmonary MAC disease and demonstrated dysfunction of γδ T cells and CD4+ T cells. This result suggests that γδ T cells did not compensate for the disadvantage due to CD4+ T cell dysfunction in these patients without AIDS.
The question of whether MAC growth is inhibited by soluble mediators released in the supernatant or whether the existence of T cells in itself is indispensable to this inhibition has attracted our attention. Silver et al. [14] demonstrated that supernatants of cocultured lymphocytes and M. tuberculosis—infected monocytes were significantly less effective than lymphocytes in inhibiting bacterial growth in monocytes and suggested involvement of cell-to-cell contact in the inhibition of bacterial growth. TNF-α bound to membranes of mouse CD4+ T cells is necessary to produce antigen-specific antileishmanial [38] and antimycobacterial [39] effects. Furthermore, involvement of adhesion molecules, such as intercellular adhesion molecule—1, leukocyte function antigen—1 [40], B7-1, and CD28 [41], necessary to activate T cells and macrophages is also suggested.
In our study, neutralizing antibodies to IFN-γ and M-CSF eliminated the effect of prestimulated CD4+ T cells to inhibit MAC growth in monocytes. If cell-to-cell contact is involved in the inhibition of bacterial growth, the following mechanism appears to be active: in the process of cell-to-cell contact, cytokines and/or receptors enhancing the IFN-γ— and M-CSF—mediated inhibitory effect are up-regulated. If IFN-γ and M-CSF are neutralized, their effects are offset. We conducted an allogeneic mixed-culture experiment with γδ T cells collected from control subjects and infected monocytes from patients with different HLA alleles and found enhanced inhibition of MAC growth. This does not necessarily rule out involvement of cell-to-cell contact. However, we find that soluble mediators in the supernatant contribute significantly to growth inhibition.
We focused on whether the defect in T cells detected in these patients is a genetically determined intrinsic phenomenon or secondary to underlying disease. It is striking that patient CD4+ T cells could not be efficiently activated by infected monocytes. Several families are reported to be vulnerable to disseminated MAC disease, and their immunocompromise is caused by the decrease in IFN-γ production by PBMC associated with abnormal IL-12 production [42] or by IFN-γ receptor abnormality [43]. These defects are assumed to be hereditary. Patients with pulmonary tuberculosis showed decreased IFN-γ production by PBMC, and this decrease continued even after recovery from the disease [44]. In localized lung infection, however, an acquired reaction appears to exist in which MAC-reactive T cells are made incompetent by certain inhibitory mechanisms, such as enhanced inhibitory cytokine, IL-10, or transforming growth factor—β [45, 46]. Another mechanism is apoptosis [47], which is the selective removal of MAC-reactive T cells. In our study, we did not evaluate these possibilities. This question may be answered by analyzing results from long-term studies that include follow-up after recovery.
Nontuberculous mycobacterial diseases are difficult to cure because of a lack of effective drugs. This difficulty in treatment has driven many researchers to try such immunotherapies as recombinant Th1 cytokine therapy, especially with HIV-infected patients [48–50]. The incidence of pulmonary MAC disease in “normal” persons who have a substantial number of CD4+ T cells is increasing. Identifying such cytokines as IFN-γ and M-CSF as being capable of inhibiting MAC growth, while focusing on the T cell function, may help clarify the mechanism(s) by which host immunocompetent cells acquire resistance to this mycobacterial infection and ultimately lead to development of a new therapeutic strategy for preventing this disease.
Acknowledgments
We thank W. Henry Boom (Infectious Division, Case Western Reserve University, Cleveland) for helpful suggestions and critical review.
References
Participants gave written informed consent, and the human experimentation guidelines of Nara Medical University were followed in the conduct of this study.
