Abstract
We measured serum cytokine concentrations and Mycobacterium tuberculosis–stimulated cytokine production by peripheral blood mononuclear cells (PBMCs) obtained from persons infected with M. tuberculosis. Serum interferon-γ (IFN-γ) and interleukin-10 (IL-10) concentrations were elevated in patients with tuberculosis compared with healthy persons who had reactions to tuberculin skin tests, but IL-18 concentrations were not. In contrast, M. tuberculosis–stimulated PBMCs from patients with tuberculosis produced less IFN-γ and IL-18 but similar amounts of IL-10, compared with PBMCs from healthy subjects who had reactions to tuberculin skin tests. Pretreatment of PBMCs from healthy subjects with reaction to tuberculin with serum from patients with tuberculosis inhibited IFN-γ production in response to M. tuberculosis, and inhibition was blocked by anti–IL-10. Thus, serum concentrations of IFN-γ, IL-18, and IL-10 do not parallel M. tuberculosis–induced cytokine levels, and increased IL-10 serum levels in patients with tuberculosis inhibit IFN-γ production in response to mycobacterial antigens.
IFN-γ plays a fundamental role in the immune response to tuberculosis. Mice with a targeted deletion of the IFN-γ gene are highly susceptible to tuberculosis [1, 2], and patients with defective receptors for IFN-γ or IL-12 are highly susceptible to severe mycobacterial infections [3]. Production of IFN-γ in response to Mycobacterium tuberculosis infection is driven by the monokines IL-12 and IL-18 [4, 5], and bacterial burdens in IL-18–deficient mice are greater than those in wild-type mice, suggesting that IL-18 contributes to immunity against tuberculosis [6]. Studies of the systemic cytokine response in patients with tuberculosis have focused either on serum cytokine levels or on cytokine production by PBMCs, and these studies have yielded contradictory results. We and others have found that M. tuberculosis–stimulated production of IFN-γ and IL-18 is reduced in patients who have tuberculosis compared with findings for healthy persons who have reactions to tuberculin skin tests (hereafter referred to as “healthy TST responders”) [5, 7, 8]. In contrast, others have shown that serum concentrations of IFN-γ and IL-18 are increased in patients with tuberculosis compared with healthy persons who provided serum samples [9]. To resolve this paradox, we studied both concentrations of cytokines in serum and M. tuberculosis–induced cytokine production by PBMCs from the same persons with M. tuberculosis infection.
Materials and Methods
Patient population. Blood samples were obtained from 10 healthy TST responders and from 17 HIV-seronegative patients with culture-proven tuberculosis who had received antituberculosis therapy for <4 weeks. Seven patients had moderately advanced pulmonary disease, 9 patients had greatly advanced pulmonary disease [10], and 1 patient had tuberculous lymphadenitis.
Informed consent was obtained from all patients. All studies involving human subjects were approved by the institutional review board of the University of Texas Health Center at Tyler.
Isolation of PBMCs and cell culture conditions. PBMCs were isolated from blood samples by differential centrifugation over Ficoll-Paque (Pharmacia). PBMCs (2 × 105) were plated in flat-bottomed 96-well plates (Becton Dickinson Labware) in 200 μL of RPMI 1640 (Life Technologies) containing penicillin-streptomycin (Life Technologies) and 10% heat-inactivated human serum, in the presence or absence of 10 μg/mL heat-killed M. tuberculosis Erdman strain (provided by P. Brennan, Colorado State University, Fort Collins).
In some experiments, 4 × 106 PBMCs recovered from healthy TST responders were cultured in 2 mL of RPMI 1640 containing 10% pooled serum obtained from patients with tuberculosis or from healthy TST responders for 24 h. PBMCs were then washed 3 times, and 4 × 106 cells were recultured in the presence or absence of 10 μg/mL heat-killed M. tuberculosis Erdman strain. In some cases, PBMCs were cultured with serum and either neutralizing antibodies to IL-10 (PharMingen) or isotype control mouse IgG1 (PharMingen) before stimulation with M. tuberculosis.
Measurement of cytokine concentrations. When blood samples were obtained, serum was processed within 30 min and stored at -70°C. For measurement of cytokine levels, supernatants of cultured PBMCs were collected after 5 days and stored at -70°C. Serum and culture supernatants were thawed, and cytokine levels were measured by use of ELISA. Paired antibodies were used to detect IFN-γ, IL-10 (PharMingen), and IL-15 (R&D Systems). ELISA kits were used to measure levels of IL-18 (MBL International) and IL-12 p70 (R&D Systems). The lower limits of detection for the ELISAs were 0.7 pg/mL for IL-12, 5 pg/mL for IFN-γ, 13 pg/mL for IL-18, 8 pg/mL for IL-10, and 32 pg/mL for IL-15. Cytokine concentrations in supernatants of PBMCs cultured without M. tuberculosis were extremely low or undetectable, and all results are presented as those in supernatants of M. tuberculosis–stimulated PBMCs.
Statistical analysis. Results are shown as mean ± SE. For data that were normally distributed, comparisons between groups were done with a paired or an unpaired Student's t test, as appropriate. For data that were not normally distributed, the Wilcoxon rank sum test was used. P <.05 was considered to be statistically significant.
Results
IFN-γ in serum and in supernatants ofM. tuberculosis–stimulated PBMCs. Serum IFN-γ levels were 6-fold higher in 17 patients with tuberculosis than they were in 10 healthy TST responders (194 ± 129 pg/mL vs. 30 ± 11 pg/mL; P =.04; figure 1A). PBMCs from 13 of these 17 patients with tuberculosis and from the 10 healthy TST responders were cultured with heat-killed M. tuberculosis for 5 days. Mean IFN-γ levels in M. tuberculosis–stimulated cultures for patients with tuberculosis were less than one-third of the corresponding values for healthy TST responders (1111 ± 159 pg/mL vs. 3767 ± 217 pg/mL; P <.001).
Concentrations of IFN-γ (A), IL-18 (B), and IL-10 (C) in serum and in supernatants of Mycobacterium tuberculosis–stimulated PBMCs. Serum samples were obtained from 17 patients with tuberculosis (TB) and from 10 healthy persons with reactions to tuberculin tests (PPD+). PBMCs recovered from 13 patients with TB and from 10 healthy PPD+ persons were cultured in the presence of M. tuberculosis, and supernatants were collected after 5 days for measurement of levels of IFN-γ, IL-18, and IL-10 by ELISA. Data are mean + SE.
Concentrations of monokines that favor a type I response. Production of IFN-γ by T cells depends on the synergistic effects of the monokines IL-12 and IL-18 [11]. IL-15 may also contribute to IFN-γ production by natural killer cells and CD8+ T cells in response to mycobacteria [12]. IL-18 concentrations were similar in the serum samples obtained from 17 patients with tuberculosis and 9 healthy TST responders (147 ± 22 pg/mL vs. 186 ± 37 pg/mL; figure 1B). However, mean IL-18 concentrations in M. tuberculosis–stimulated cultures of PBMCs from patients with tuberculosis were less than one-third of the corresponding values for healthy TST responders (72 ± 15 pg/mL vs. 237 ± 62 pg/mL; P =.04). Serum IL-12 p70 levels in patients with tuberculosis were similar to those in healthy TST responders (data not shown), and IL-12 was not detected in culture supernatants of M. tuberculosis–stimulated PBMCs from either group. Serum IL-15 and M. tuberculosis–stimulated IL-15 concentrations were similar in samples obtained from patients with tuberculosis and healthy TST responders (data not shown).
IL-10 levels in serum and in supernatants ofM. tuberculosis–stimulated PBMCs. We next studied production of IL-10, which suppresses the type 1 cytokine response to M. tuberculosis and other mycobacteria [13, 14]. Mean serum IL-10 levels were elevated 15-fold in 17 patients with tuberculosis compared with levels in 10 healthy TST responders (75 ± 42 pg/mL vs. 5 ± 1 pg/mL; P <.01; figure 1C). However, M. tuberculosis–induced IL-10 production by PBMCs recovered from 13 patients with tuberculosis was similar to the production by PBMCs from 10 healthy TST responders (426 ± 149 pg/mL vs. 406 ± 144 pg/mL; P =.93).
Effect of serum obtained from patients with tuberculosis on IFN-γ production. We hypothesized that, in patients with tuberculosis, high serum levels of IL-10 reduced IFN-γ production by PBMCs cultured with M. tuberculosis in vitro. To evaluate this possibility, we first cultured PBMCs recovered from 8 healthy TST responders with 10% pooled serum from 6 patients with tuberculosis or 10% pooled serum from 6 healthy TST responders. Control cells were cultured with 10% heat-inactivated serum from healthy tuberculin reaction–negative donors. After 24 h, PBMCs were washed and stimulated with heat-killed M. tuberculosis for 5 days, and IFN-γ levels were measured in supernatants. Serum from patients with tuberculosis significantly inhibited production of IFN-γ (1423 ± 496 pg/mL vs. 2729 ± 346 pg/mL; P =.02; figure 2). In contrast, serum from healthy TST responders did not inhibit IFN-γ production. Addition of anti–IL-10 to serum from patients with tuberculosis restored M. tuberculosis–induced production of IFN-γ, but incubation of PBMCs with anti–IL-10 alone did not affect IFN-γ production. This suggests that high serum IL-10 levels in patients with tuberculosis inhibit M. tuberculosis–induced IFN-γ production. PBMCs recovered from 2 patients with tuberculosis produced a mean IFN-γ level of 1460 pg/mL when incubated with autologous serum, but this value increased to 2486 pg/mL after incubation with serum from healthy TST responders.
Effect of pooled serum from patients with tuberculosis and anti–IL-10 antibody on IFN-γ production: PBMCs from 8 healthy persons with reactions to tuberculin tests were incubated with 10% heat-inactivated serum from tuberculin-negative healthy donors (control serum), 10% pooled serum from 6 healthy persons with reaction to tuberculin (PPD+ serum), 10% pooled serum from 6 patients with tuberculosis (TB serum), TB serum and anti–IL-10, or anti–IL-10 and control serum. After 24 h, PBMCs were washed 3 times and cultured in the presence of Mycobacterium tuberculosis for 5 days, and supernatants were collected for measurement of IFN-γ levels by ELISA. Data are mean + SE. *P =.02 compared with control serum; P =.002 compared with serum from PPD+ subjects; P =.001 compared with TBS and anti–IL-10.
Discussion
To our knowledge, this is the first study to evaluate both serum levels of cytokines and M. tuberculosis–stimulated cytokine production by PBMCs in the same group of patients with tuberculosis. We found that serum IFN-γ and IL-10 levels were higher in patients with tuberculosis than in healthy TST responders. In contrast, M. tuberculosis–stimulated PBMCs recovered from patients with tuberculosis produced less IFN-γ and similar amounts of IL-10 than they did in healthy TST responders. Addition of serum from patients with tuberculosis to PBMCs recovered from healthy TST responders inhibited their capacity to produce IFN-γ in response to M. tuberculosis, and this inhibition was abrogated by anti–IL-10. These findings demonstrate that serum cytokine levels do not parallel M. tuberculosis–induced cytokine production by PBMCs and that high serum IL-10 levels in patients with tuberculosis inhibit M. tuberculosis–induced production of IFN-γ.
Although many investigators have evaluated cytokine production by PBMCs recovered from patients with tuberculosis, few have measured serum cytokine concentrations [9, 15, 16]. Our results confirm that serum levels of IFN-γ and IL-10 are abnormally high in patients with tuberculosis, whereas serum IL-12 concentrations are normal [9, 15, 16]. We found that serum levels of IL-15 and IL-18 were not elevated in patients with tuberculosis. In contrast, Yamada et al. [9] found that serum IL-18 concentrations were higher in untreated patients with tuberculosis than they were in healthy controls. This difference may be due to a rapid decrease in serum cytokine levels during antituberculosis therapy. Yamada et al. [9] studied patients with tuberculosis before they received therapy, whereas we evaluated patients who had been treated for up to 4 weeks. In untreated patients with tuberculosis, mean serum IL-18 levels are only 3-fold higher than they are in healthy subjects [9], whereas mean serum IFN-γ and IL-10 concentrations are 60–200-fold higher than are those in healthy subjects [16]. Because serum cytokine levels decrease during antituberculosis therapy [16], the relatively low serum IL-18 concentrations may decrease into the normal range during the first few weeks of antituberculosis therapy, whereas IFN-γ and IL-10 levels remain abnormal at this time.
The discordance between serum cytokine levels and M. tuberculosis–induced cytokine production by PBMCs suggests that serum cytokine concentrations do not reflect cytokine production by peripheral blood cells. We speculate that the elevated serum IFN-γ and IL-10 levels in patients with tuberculosis are due to leakage of these cytokines from tissue into the circulation, because T cells and macrophages at the site of disease in tuberculosis can both produce IFN-γ and IL-10 [17,18–19], and increased local vascular permeability is likely to favor diffusion of cytokines into the bloodstream. In addition, mean serum IFN-γ and IL-10 concentrations are higher in patients with tuberculosis with more-severe disease [9, 15, 16], who are likely to exhibit the most severe local inflammatory response and production of the highest concentrations of cytokines. Finally, the decrease in serum levels of IFN-γ and IL-10 as tissue inflammation resolves during therapy is also consistent with this hypothesis.
Mononuclear phagocytes play a central role in antigen presentation and in eliciting IFN-γ production by T cells. When they are exposed to M. tuberculosis, antigen-presenting cells produce IL-12 and IL-18, which bind to their respective receptors and induce IFN-γ production [20, 21]. IL-12 also up-regulates expression of the IL-18 receptor α chain [22], resulting in synergistic effects of these cytokines on IFN-γ production. We found that serum from patients with tuberculosis inhibited the capacity of PBMCs from healthy TST responders to produce IFN-γ in response to M. tuberculosis antigens. This inhibition was reversed by neutralization of IL-10, suggesting that the reduced capacity of PBMCs in patients with tuberculosis to produce IFN-γ in response to M. tuberculosis is due in part to exposure to high concentrations of IL-10 in vivo. IL-10 can inhibit IFN-γ production by T cells through transcriptional inhibition of IL-12 p40 mRNA production by monocytes [23]. However, M. tuberculosis–stimulated monocytes and PBMCs from patients with tuberculosis produced normal amounts of IL-12 in the present study and in prior studies [24]. Although high serum IL-10 levels do not reduce IL-12 production by PBMCs of patients with tuberculosis, they may inhibit responsiveness of T cells to IL-12. This hypothesis is supported by previous work demonstrating that IL-10 down-regulates expression of IL-12 receptor β1 and β2 chains and that T cells from patients with tuberculosis have reduced IL-12 receptor expression [25]. IL-10 also inhibits IL-18 mRNA expression [26], and we speculate that high serum IL-10 levels decrease M. tuberculosis–induced IL-18 production by PBMCs from patients with tuberculosis, contributing to reduced IFN-γ production.
In summary, we demonstrated that serum cytokine concentrations in patients with tuberculosis do not parallel M. tuberculosis–induced cytokine production by PBMCs. High serum IL-10 levels in patients with tuberculosis inhibit M. tuberculosis–induced cytokine production of IFN-γ, possibly through down-regulation of IL-12 receptor expression and inhibition of IL-18 production. Further studies are needed to understand the complex interactions of cytokines in the human immune response to tuberculosis.
References
Financial support: National Institutes of Health (grant AI-27285), Center for Pulmonary and Infectious Disease Control, the Cain Foundation for Infectious Disease Research, and the Potts Memorial Foundation. P.F.B. holds the Margaret E. Byers Cain Chair for Tuberculosis Research.
Comments