SEARCH

SEARCH BY CITATION

Keywords:

  • anti-TB therapy;
  • immune activation;
  • IFN-γ;
  • regulatory T cells;
  • tuberculosis

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Active tuberculosis (TB) is associated with prolonged suppression of Mycobacterium tuberculosis (MTB)-specific immune responses, but mechanisms involved are understood incompletely. We investigated a potential role for CD4+CD25+ regulatory T cells in depressed anti-MTB immunity by evaluating serially CD4 cell phenotype and interferon (IFN)-γ production by mononuclear cells from patients with TB. At diagnosis, frequencies of CD4+CD25+ T cells were increased in blood from TB patients compared to healthy purified protein derivative (PPD)-positive controls (with a history of prior TB exposure), and remained elevated at completion of therapy (6 months). By contrast, expression of another activation marker, CD69, by CD4 T cells was increased at diagnosis, but declined rapidly to control levels with treatment. Among CD4+CD25+ T cells from TB patients at diagnosis those expressing high levels of CD25, probably representing regulatory T cells, were increased 2·9-fold when compared to control subjects, while MTB-stimulated IFN-γ levels in whole blood supernatants were depressed. A role for CD4+CD25+ T cells in depressed IFN-γ production during TB was substantiated in depletion experiments, where CD25+-depleted CD4 T cells produced increased amounts of IFN-γ upon MTB stimulation compared to unseparated T cells. At follow-up, IFN-γ production improved most significantly in blood from TB patients with high baseline frequencies of CD4+CD25+ T cells (more than threefold higher than controls for both total and CD25hi+ CD4 T cells), who also had a significant drop in frequencies of both total and ‘regulatory’ CD4+CD25+ T cells in response to treatment. Expansion of CD4+CD25+ regulatory T cells during active TB may play a role in depressed T cell IFN-γ production.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Tuberculosis (TB) remains the leading cause of morbidity and mortality due to any one infectious agent worldwide [1]. However, understanding of its immunopathogenesis is still incomplete.

Active TB is characterized by a profound and prolonged suppression of Mycobacterium tuberculosis (MTB)-specific T cell responses, as evidenced by decreased production of the cytokines interleukin (IL)-2 and interferon (IFN)-γ[2–6]. Overproduction of immunosuppressive cytokines [IL-10 and transforming growth factor (TGF)-β] by mononuclear phagocytes has been implicated in decreased T cell function during TB [2,7–9]. TGF-β may also have a role in increased susceptibility of IFN-γ-producing T cells in the blood and at sites of MTB infection to apoptosis [10–12]. However, levels of TGF-β and IL-10 return to normal following 3 months of anti-TB treatment [2,7], while MTB-stimulated production of IFN-γ remains depressed beyond completion of anti-TB therapy [2]. Thus, additional mechanisms probably regulate T cell responses during active TB.

Recently, a subset of CD4 T cells that co-express CD25 and have immunoregulatory properties (CD4+CD25+ T cells) has been described. CD4+CD25+ T cells occur naturally in mice and man and account for 5–10% of circulating CD4+ T cells in healthy subjects [13–17]. CD4+CD25+ Tregs (which are anergic upon T cell receptor ligation themselves) suppress activation and proliferation of other (CD25) CD4+ and CD8+ T cells [13,16] through inhibition of IL-2 transcription and promoting cell cycle arrest. Antigen-specific T cells may be particularly susceptible to the immunosuppressive effects of CD4+CD25+ T cells [15,16,18,19]. In support of this premise, T cells from CD4+CD25+-depleted mice produced increased amounts of IFN-γ in response to low antigen doses compared to normal mice. Depletion of CD4+CD25+ Tregs also resulted in enhanced differentiation of IFN-γ-producing cells and decreased production of IL-4 [14].

A role for CD4+CD25+ Tregs in suppression of immune responses directed against microbes has been described recently [20,21]. In murine Leishmania major infection, CD4+CD25+ T cells recruited to the dermis of infected animals suppressed the ability of CD4+CD25 T cells to eliminate parasites from the inflammatory site [20]. Both IL-10-dependent and IL-10-independent pathways were involved in this suppression of T cell function by CD4+CD25+T cells [22]. To date, information about immunoregulatory properties of CD4+CD25+ T cells in human infectious diseases is scarce. During HIV disease frequencies of CD4+CD25+ Tregs are increased and may play a role in depressed T cell responses (both to HIV-specific and recall antigens) [23–25].

The current study was conducted to evaluate frequencies of CD4+CD25+ T cells in the blood of TB patients throughout the course of TB treatment, and to assess their role in depressed MTB antigen-stimulated IFN-γ production.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Subjects

Acid fast bacilli (AFB) smear-positive patients with newly diagnosed pulmonary TB (n = 37) and AFB smear-negative patients suspected strongly to have TB and scheduled to undergo bronchoscopy for diagnostic reasons (n = 7) were identified at the TB clinic of the Hospital Universitário Cassiano Antonio de Morais (HUCAM) in Vitoria, Brazil. Following informed consent, TB patients meeting eligibility criteria for study participation [26] were enrolled. The study protocol was reviewed and approved by the Institutional Review Boards both at CWRU and UFES, and complies with human experimentation guidelines of the US Department of Health and Human Services.

A diagnosis of TB was confirmed in all patients by positive mycobacterial culture. The mean age of TB patients was 35·3 ± 9·4 years. Thirty-one patients were male and 13 female. Eight patients had minimal disease, 16 moderately advanced and 20 far advanced TB radiographically [27]. Patients were treated with standard short-course chemotherapy as before [26]. Fourteen HIV-uninfected purified protein derivative (PPD)-positive healthy control subjects, aged 37 ± 9 years, with a history of TB exposure leading to skin test conversion but not active TB, were recruited among staff at HUCAM. Twelve controls were male and two were female. Procedures for collection and processing of sputum and blood were as described previously [26].

Analysis of cell phenotype by flow cytometry

Ethylenediamine tetraacetic acid (EDTA) blood was used for surface staining and analysis of T cell markers (three-colour analysis with combinations of fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)- and peridinum chlorophyll (PerCP)-conjugated antibodies). Antibodies used included: anti-CD3 (SK7), anti-CD4 (SK3), anti-CD8 (SK1), anti-CD25 (2A3), anti-CD69 (L78) and anti-IgG1 (X40) (Becton Dickinson Biosciences, Mountain View, CA, USA). Flow cytometry data were analysed in one session at study completion using cellquest software (Becton Dickinson), to ensure that surface marker expression could be compared directly at all study time-points.

Whole blood culture

Heparinized blood was diluted 10-fold and dispensed into 24-well tissue culture plates (1 ml aliquots per well). Wells remained unstimulated or received MTB (H37Rv) culture filtrate (CF) (5 µg/ml, provided under NIH contract). Cell-free supernatants were collected following 5-day incubation at 37OC, 5% CO2 and stored frozen at −70°C until use.

CD4+CD25+ Treg depletion assay

Peripheral blood mononuclear cells (PBMC) first were adhered to plastic dishes to separate mononuclear (MN) and non-adherent T cells (NAC). CD4+ T cells then were obtained from NAC using negative selection (CD4 T cell isolation kit II; Miltenyi Biotech, Milburn, CA, USA). A subset of CD4 T cells was depleted further of CD25+ T cells using magnetic beads (Miltenyi).

MN (5 × 104/well) and total or CD25-depleted CD4+ T cells (1·5 × 105/well) were cultured (37°C, 5% CO2) in replicate in 96-well tissue culture plates in medium alone or medium containing anti-CD3 monoclonal antibodies (mAb) (HIT3, 0·1 µg/ml, BD Pharmingen, La Jolla, CA, USA) or MTB [5 × 105 colony-forming units (CFU)/well]. Supernatants were collected following 72 h of incubation and stored frozen until assessment of IFN-γ.

Cytokine assays

IFN-γ, TNF-α and IL-10 levels in whole blood culture supernatants were assessed using commercially available enzyme-linked immunosorbent assay (ELISA) kits from Endogen (Woburn, MA, USA), R&D Systems (Minneapolis, MN, USA) and Biosource (Camarillo, CA, USA), respectively. The lower limit of detection for the assays was <  2 pg/ml, <  5 pg/ml and < 11 pg/ml, respectively. Cytokine content in sputum was assessed as described previously [26].

Statistics

Normally distributed data sets were analysed by Student's t-test, paired t-test, analysis of variance (anova) and linear regression and correlation analysis (using ‘Primer for Biostatistics’). The Wilcoxon two-sample test and Kruskall–Wallis test were used for data sets that were not normally distributed (using SAS). P ≤ 0·05 was considered significant.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Frequencies of CD4+CD25+ T cells in blood from TB patients and healthy PPD+ controls

We evaluated expression of CD25 by CD4 and CD8 T cells in blood from TB patients and healthy PPD+ control subjects. Expression of the early activation marker CD69 was assessed in parallel. Expression of CD25 (calculated as the ratio of frequencies of CD4+CD25+ T cells/total CD4+ T cells) was found on 41·2 ± 2·1% (mean ± s.e.m., range 17·3–75·8%) of CD4 T cells from TB patients, and on 12 ± 1·3% (range 8·9–5·1%) of CD4 T cells from healthy control subjects, accounting for a 3·6-fold difference in frequencies of CD4+CD25+ T cells (P ≤ 0·001, n = 34; Fig. 1). Expression of CD25 by CD8 T cells from TB patients and controls was low (< 10%) and comparable between both groups (data not shown). Expression of CD69 at diagnosis was also increased on CD4 (but not CD8) T cells from TB patients (7·8 ± 0·9%) compared to control subjects (2·7 ± 0·4%; 2·7-fold, P ≤ 0·05) (Fig. 1). Total numbers of CD4 and CD8 T cells were decreased among PBMC from TB patients compared to healthy controls (data not shown). This confirms previous findings [28], and indicates that increased expression of CD25 and CD69 on CD4+ cells during newly diagnosed active TB is not secondary to expansion of the CD4 T cell population.

image

Figure 1. Frequencies of CD4+CD25+ T cells during the course of tuberculosis (TB). Whole blood collected from TB patients (n = 34) throughout the course of TB treatment and healthy purified protein derivative (PPD)-positive control subjects (n = 10) were stained with fluorochrome conjugated antibodies to CD3, CD4 and CD25 or CD69 and frequencies of CD4+CD25+ and of CD4+CD69+ T cells were assessed in one setting following completion of this study. Data are expressed as mean ± s.e.m. of the ratio of percentage of CD4+CD25+ T cells and total CD4+ T cells.

Download figure to PowerPoint

It has been shown previously that among CD4+CD25+ T cells only those with high fluorescence intensity of CD25 may have a regulatory function [13], whereas expression of low levels of CD25 by CD4 T cells may be indicative of T cell activation. To ascertain that increased frequencies of CD4+CD25+ T cells observed in the current study were not simply a result of excess immune activation, we next determined frequencies of CD4+CD25+ T cells expressing high levels of CD25 (CD4+CD25hi+ T cells). The flow cytometry algorithm used for analysis is shown in Fig. 2 for a representative TB patient. Interestingly, CD4+CD25hi+ T cells were enriched among CD4 T cells from TB patients compared to control subjects (2·9-fold: 8·0 ± 0·9%versus 2·7 ± 0·1%, respectively; P < 0·001 (Student's t-test; data not shown).

image

Figure 2. Algorithm for detection of Tregs based on intensity of CD25 expression. To begin to address whether CD4+CD25+ T cells in the peripheral blood from tuberculosis (TB) patients may indeed represent regulatory T cells, we next established frequencies of CD4 T cells expressing only the highest levels of CD25 (> log 2). For this purpose CD4 T cells were first identified among total peripheral blood mononuclear cells (PBMC) stained with antibodies to CD3 and CD4 (without prior gating of T cells based on forward scatter (FSC) and side scatter (SSC) characteristics). Fluorescence intensity of CD25 then was established by histogram analysis in the fluorescence 2 channel. Data for one representative TB patient are shown.

Download figure to PowerPoint

Next, frequencies of CD4+CD25+ and CD4+CD69+ T cells were evaluated during anti-MTB therapy. After initiation of TB treatment, frequencies of CD4+CD69+ T cells declined rapidly and by week 12 of treatment were no longer significantly different from values of controls (4·5 ± 2·8%, P = 0·5; Fig. 1). By contrast, the percentage of CD4+CD25+ T cells remained increased throughout the course of TB therapy, and at completion of treatment was threefold higher than that of control subjects (41·0 ± 2·6%; P < 0·001; Fig. 1). Frequencies of CD4+CD25hi+ T cells followed a pattern similar to that of overall CD4+CD25+ T cells, exceeding frequencies among PBMC from control subjects by 2·3-fold (6·51 ± 0·49%, P < 0·0001; Student's t-test) by week 24 of follow-up (data not shown).

Thus, it appears that both overall frequencies of CD4+CD25+ T cells and those of CD4+CD25hi+ T cell are increased at the time of TB diagnosis, and remain elevated at its completion. This observation provides a potential mechanism for sustained depression of T cell responses during active TB [2].

Frequencies of CD4+CD25+ T cells are particularly accentuated at sites of active MTB infection in the lung during human TB

We wanted to ascertain that findings of increased numbers of CD4+CD25+ T cells in the peripheral blood reflected accurately events at sites of active MTB infection, such as the lung. For this purpose we recruited seven consecutive patients undergoing bronchoscopy to establish a diagnosis of TB. We assessed simultaneously frequencies of CD4+CD25+ T cells overall and of CD4+CD25hi+ T cells among mononuclear cells recovered from radiographically involved and uninvolved lung segments and autologous PBMC. As shown in Fig. 3, both CD4+CD25+ T cells and CD4+CD25hi+ T cells were compartmentalized in the TB-involved lung (*P < 0·02 and **0·04, respectively, when compared to the uninvolved lung; n = 7). Further, frequencies of CD4+CD25+ and of CD4+CD25hi+ T cells from both involved and uninvolved lung compartments exceeded those in the peripheral blood by a minimum of twofold (***P < 0·001 for both; Fig. 3). These data indicate that assessment of frequencies of CD4+CD25+ T cells overall and of CD4+CD25hi+ T cells in the peripheral blood provides an accurate, albeit attenuated, picture of events at sites of active MTB infection in the lung.

image

Figure 3. Frequencies of CD4+CD25hi+ T cells are increased at sites of active Mycobacterium tuberculosis (MTB) infection in lungs of patients with pulmonary tuberculosis (TB). Frequencies of CD25+ CD4 T cells overall and of CD4+CD25hi+ T cells in involved (BAC I) and uninvolved (BAC U) lung and among peripheral blood mononuclear cells (PBMC) were assessed in patients with pulmonary TB undergoing diagnostic bronchoscopy (n = 7) by flow cytometry. Frequencies of both CD4+CD25+ T cells and of CD4+CD25hi+ T cells were highest in the TB-involved lung (P < 0·02 and 0·04, respectively, when compared to the TB-uninvolved lung). Further, both lung compartments contained significantly greater frequencies of CD4+CD25+ T cells and of CD4+CD25hi+ T cells when compared to PBMC (minimum difference twofold, P < 0·001 for all).

Download figure to PowerPoint

Depletion of CD4+CD25+ Tregs improves CD4 T cell responses of TB patients

To determine if CD4+CD25+ T cells have a role in down-modulation of MTB-specific T cell responses, we assessed the functional capacity of unfractionated CD4 T cells and CD4 T cells depleted of CD4+CD25+ T cells. For this purpose, we recruited an additional three TB patients and four control subjects. CD4 T cells and CD4 T cells depleted of CD4+CD25+ T cells were prepared. Purities of CD4+CD25 T cells were >  95% for both TB patients and control subjects.

Unfractionated CD4 T cells and purified CD4+CD25 and CD4+CD25+ T cell subsets were then combined with autologous blood MN in 96-well tissue culture plates and cultured with MTB or anti-CD3 for 72 h prior to assessment of IFN-γ levels in supernatants. Depletion of CD4+CD25+ T cells resulted in increased MTB and anti-CD3 induced production of IFN-γ in supernatants of TB patients only (Fig. 4). Differences between IFN-γ levels in supernatants of MN co-cultured with CD4 T cells as a whole and T cells depleted of CD4+CD25+ subsets were statistically significant only following stimulation with MTB (*P ≤ 0·001; Fig. 4b). Amounts of IFN-γ produced by anti-CD3 stimulated cultures exceeded those induced by MTB by more than 10-fold. Supernatants of co-cultures of MN and CD4+CD25+ T cells contained negligible amounts of IFN-γ following stimulation with anti-CD3 or MTB (data not shown). Data presented here support further the notion that during active TB CD4+CD25+ T cells may have regulatory properties.

image

Figure 4. Depletion of CD4+CD25+ T cells improves T cell interferon (IFN)-γ responses of CD4 T cells from tuberculosis (TB) patients. CD4 T cells and CD4 T cells depleted of CD4+CD25+ T cells from TB patients (n = 3) and healthy purified protein derivative (PPD)-positive control subjects (n = 4) were purified using magnetic beads. T cells were added to autologous blood monocytes to assure optimal antigen presentation and incubated for 72 h in the presence or absence of a-CD3 (0·1 mg/ml) (a) or Mycobacterium tuberculosis (MTB) [5 × 105 colony-forming units (CFU)] (b). IFN-γ levels in culture supernatants were determined by enzyme-linked immunosorbent assay (ELISA). Differences between cytokine levels in culture supernatants of CD4 T cells and CD4 T cells depleted of CD4+CD25+ T cells were assessed by paired t-test. *P < 0·001 when compared to total CD4 T cells.

Download figure to PowerPoint

CD4+CD25+ T cell expression profile may differ among subsets of TB patients

As shown above, frequencies of CD4+CD25+ T cells in blood from TB patients were not different at baseline and completion of TB treatment. However, the range of CD25+ expression by CD4+ T cells from individual TB patients was somewhat large. To evaluate whether analysis of mean CD4+CD25+ T cell frequencies may mask differing patterns of CD25 expression in subsets of TB patients, we next plotted frequencies of CD4+CD25+ T cells during therapy for each patient individually. Interestingly, the TB patient population could be separated readily into two groups. Group 1 (56% of enrolees), were subjects whose baseline CD4+CD25+ T cell counts were more than threefold higher than those of control subjects, and decreased during TB treatment. Group 2 (44% of enrolees) were patients with more modest (two- to threefold) increases in baseline CD4+CD25+ T cell counts, but whose numbers of CD4+CD25+ T cells increased during TB treatment. Baseline frequencies of CD4+CD25+ T cells in the blood from patients in group 1 were significantly higher (1·5-fold) than those of patients in group 2 (*P < 0·001; Fig. 5). By completion of treatment, however, frequencies of CD4+CD25+ T cells were comparable between both groups (Fig. 5). We next assessed whether the two patient groups also differed with respect to their frequencies of CD4+CD25hi+ T cells. As shown in Table 1, at TB diagnosis frequencies of CD4+CD25hi+ T cells in the blood from TB patients in group 1 exceeded those in group 2 by 1·6-fold (*P < 0·0001), then declined rapidly, and from week 4 of treatment remained comparable to those of CD4+CD25hi+ T cells of patients in group 2.

image

Figure 5. Frequencies of CD4+CD25+ T cells during the course of tuberculosis (TB) differs among subsets of TB patients. Frequencies of CD4+CD25+ T cells were assessed for individual patients by immunostaining and flowcytometry (as described above). Based on the patterns of CD25 expression throughout TB treatment, two subsets of TB patients could be identified: group 1 whose CD4+CD25+ T cells were increased more than threefold compared to controls and whose T cell numbers decreased with treatment and group 2 with only moderately increased CD4+CD25+ T cell counts, whose T cells increased during TB treatment. *P < 0·001 when compared to group 1; **P < 0·04 compared to baseline frequencies of CD4+CD25+ T cells; ***P < 0·005 compared to baseline frequencies of CD4+CD25+ T cells (all assessed by Student's t-test).

Download figure to PowerPoint

Table 1.  Frequencies of CD4+CD25hi+ T cells over the course of tuberculosis (TB) treatment in group1 and group 2 TB patients.
 Time 0Week 4Week 24
  1. Frequencies of CD4 T cells with high fluorescence intensity for CD25 were assessed for patients in group1 and group 2 at TB diagnosis and during TB treatment. Results are expressed as mean ± s.e.m. percentage of CD4+CD25hi+ T cells among total CD4 T cells. *P < 0·0001 when compared to group 2 baseline frequencies.

Group 19·66 ± 0·42*7·44 ± 0·416·42 ± 0·24
Group 26·29 ± 0·67·01 ± 0·366·0 ± 0·24

By contrast to CD4+CD25+ T cells, frequencies of CD69+ CD4 T cells did not differ between groups at any time-point studied.

CD4+CD25+ expression correlates with immunological, but not clinical profile

We next examined whether the two groups of TB patients differed in their immunological or clinical make-up.

First, we assessed levels of the macrophage activating cytokine IFN-γ in MTB CF-stimulated whole blood cultures. IFN-γ was depressed in supernatants from TB patients as a whole compared to those from controls (data not shown). Interestingly, IFN-γ levels were much lower in culture supernatants from group 1 compared to group 2 TB patients (**P ≤ 0·05; Fig. 6a). With treatment, IFN-γ production increased in the whole blood of group 1 patients, and by week 4 of follow-up (when CD4+CD25hi+ T cell numbers had decreased by 35%) was more than twofold higher than baseline (*P ≤ 0·001; Fig. 6a). IFN-γ levels in supernatants of group 2 TB patients did not change significantly with therapy (Fig. 6a).

image

Figure 6. Cytokine levels in Mycobacterium tuberculosis culture filtrate (MTB-CF)-stimulated whole blood supernatants from tuberculosis (TB) patients in groups 1 and 2. Whole blood was collected at defined time intervals throughout TB treatment and cultured for 5 days in the presence or absence of MTB-CF. Interferon (IFN)-γ and interleukin (IL)-10 levels in culture supernatants were assessed by enzyme-linked immunosorbent assay (ELISA). (a) IFN-γ levels in supernatants collected at baseline, week and week 24 (end of TB treatment). *P < 0·001 when compared to baseline; **P < 0·05 compared to group 1 (Kruskall–Wallis). (b) IL-10 levels in supernatants collected at baseline, week and week 24; *P < 0·005 when compared to baseline; **P < 0·001 when compared to group 1 (Kruskall–Wallis).

Download figure to PowerPoint

IL-10 and TGF-β, cytokines produced in excess by monocytes during active TB [2,7], may play a role in the immunosuppressive properties of CD4+CD25+ Tregs[17]. Therefore, we then investigated the relationship between frequencies CD4+CD25+ T cells and levels of IL-10 in MTB antigen-stimulated whole blood culture supernatants from patients in groups 1 and 2. Levels of TGF-β were not assessed, because high background levels of TGF-β in whole blood cultures (released by platelets during degranulation) preclude meaningful analysis of antigen-induced changes in production of the cytokine. MTB-stimulated IL-10 levels were determined in supernatants from group 1 (n = 10) and group 2 (n = 6) patients. Levels of IL-10 were higher (1·8-fold, **P < 0·001) in supernatants of patients from group 1 compared to group 2 (Fig. 6b). Interestingly, by week 4 antigen-stimulated production of IL-10 had declined significantly in whole blood cultures from group 1 patients (*P < 0·005) and in parallel with declining frequencies of CD4+CD25+ Tregs.

To investigate if differences in frequencies of circulating CD4+CD25+ T cells had an impact on local immune responses at sites of active MTB infection in the lung, we next measured IFN-γ and TNF-α immunoreactivities in sputum from TB patients in groups 1 and 2. At baseline, median levels of IFN-γ and TNF-α were higher in sputum from group 2 compared to group 1 patients (1·5-fold and 1·7 fold, respectively, Fig. 7). Differences between groups, however, were not statistically significant (Wicoxon's two-sample test and Kruskall–Wallis test). IL-10 was undetectable in sputum from either group. To evaluate whether TB patients in groups 1 and 2 also differed in their clinical profiles we then evaluated microbiological data and clinical records from all study participants. Mycobacterial load (expressed as log10 CFU) at the time of TB diagnosis was increased slightly in patients in group 1 compared to those in group 2 (log10 CFU: 6·2 ± 0·1 and 5·8 ± 0·1 (mean ± s.e.m.), respectively; P ≤ 0·002, Student's t-test). No differences between subjects in groups 1 and 2 with respect to duration of symptoms (cough, fever and weight-loss), haematological parameters or PPD skin test size (data not shown) were observed. The extent of radiographic disease [27] was also not significantly different between groups. However, when parameters used to compute the overall radiographic score were evaluated individually, some subtle differences were observed: the number of disease-involved zones at the time of TB diagnosis was significantly higher in subjects in group 1 (P ≤ 0·04, Student's t-test; data not shown). By contrast, lung fibrosis was more pronounced in subjects in group 2 (P ≤ 0·04, Student's t-test; data not shown). Numbers of cavities did not differ significantly between groups. Thus, only subtle differences in clinical characteristics separated patients with high and low frequencies of CD25+CD4+ Tregs at baseline.

image

Figure 7. Cytokine levels in sputum from tuberculosis (TB) patients in groups 1 and 2. Interferon (IFN)-γ and tumour necrosis factor (TNF)-α levels in sputum from TB patients in group 1 and group 2 were assessed by enzyme-linked immunosorbent assay (ELISA) (expressed as median values). Both IFN-γ and TNF-α levels were higher (approximately 1·5-fold) in sputum from group 2 patients. However, differences were not statistically significant (Wilcoxon and Kruskall–Wallis).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Research over the last decade has identified a multitude of mechanisms for depressed MTB antigen-stimulated production of IFN-γ at the time of diagnosis of active TB [2,3,7–10]. However, factors responsible for the prolonged delay in recovery of MTB-specific T cell IFN-γ production by PBMC of TB patients [2] remain unidentified. Here, we report that frequencies both of CD4+CD25+ T cells overall and of CD4+CD25hi+ T cells, probably representing true regulatory T cells [29], are increased during active TB and that this increase is sustained even at completion of TB treatment. Depletion of CD4+CD25+T cells improved T cell IFN-γ production by CD4 T cells from newly diagnosed TB patients, indicating that increased frequencies of CD4+CD25+ T cells are not simply a reflection of the excessive immune activation during TB [30] but that subsets, presumably those with a CD4+CD25hi+ phenotype, indeed have immunoregulatory properties. Expression of the classic activation marker CD69 by CD4 T cells was increased at the time of diagnosis of TB, but corrected rapidly with TB treatment. Thus, CD4+CD25+ T cells, whose frequencies remain expanded even after completion of TB treatment (as shown here), may play a role in the prolonged suppression of MTB-induced IFN-γ production by PBMC from TB patients.

In contrast to previously published studies [18,19,31], depletion of CD4+CD25+ T cells did not improve IFN-γ production in healthy control subjects. One possible explanation for this discrepancy is that most previous studies evaluated T cell proliferation, rather than IFN-γ production, which may be regulated differently. Another, more probable explanation relates to the fact that frequencies of CD4+CD25hi+ (true regulatory) T cells were extremely low (< 3%) among CD4 T cells from all but one control subject, and therefore may not have had any significant impact on IFN-γ levels. IFN-γ levels actually increased (by 30%) in supernatants of CD25-depleted CD4 T cells from the one control subject with frequencies of CD4+CD25hi+ T cells in excess of 5%.

While CD4+CD25+ Tregs are involved prominently in the regulation of fundamental immune functions such as self-tolerance [32,33], autoimmunity and anti-tumour responses [34–36], recent evidence also supports a role for CD4+CD25+T cells in regulation of immune responses during infection [23–25,37]. The specific role of CD4+CD25+ T cells during infections may be to limit ‘strong’ Th1 responses induced by microbial antigens and to prevent excessive inflammation and tissue damage [14,20–22]. Thus, immunoregulatory circuits involving Tregs may be potentially ‘beneficial’ to the host in diseases such as TB, whose hallmark is long-term persistence of the pathogen and strong immune responses aimed at prevention of reinfection. In fact, such Treg-mediated down-regulation of T cell responses may have been the basis for decreases of frequencies of MTB antigen-responsive IFN-γ producing T cells following successful chemotherapy of patients with active TB and of latently infected PPD-positive people as reported in a recent study [38]. To date, CD4+CD25+ Tregs have been shown to be important regulators of immune responses during HIV infection [23–25,37,39] and to lead to immunosuppression in chronic onchocerciasis [40].

In HIV disease, depletion of CD4+CD25+ Tregs from PBMC augmented HIV- and CMV-specific CD4 and CD8 T cell responses [23]. Conversely, it has been suggested that the beneficial effect of IL-2 therapy during HIV infection may stem from expansion of CD4+CD25+ T cells and down-regulation of the excess immune activation implicated in progression of HIV disease [39]. Expansion of CD4+CD25+ T cells at the time of diagnosis of TB in the current study may, similarly, suggest a regulatory role for this cell type. Specifically, expansion of CD4+CD25+ T cells may be an adaptive host response to counteract the intense inflammatory response at sites of active MTB infection, such as the lung. The same immunoregulatory properties may then, paradoxically, depress anti-MTB T cell responses necessary to control infection (such as IFN-γ production). This hypothesis is supported by our recent data, which indicate that frequencies of Tregs at the time of TB diagnosis are increased more than twofold in the lung of TB patients compared to the peripheral blood (Fig. 3). A similar scenario also appears to be operant in patients with river blindness (due to onchocerciasis), where expansion of CD4+CD25+ T cells in the skin may prevent dermatitis [40].

An interesting finding of the current study was that, even though overall frequencies of CD4+CD25+ T cells were markedly increased in the blood from TB patients when compared to healthy tuberculin skin test positive subjects, TB patients could be subdivided further into two groups. These two groups were distinguishable both based on their baseline frequencies of CD4+CD25+ T cells overall and of CD4+CD25hi+ Tregs and on their patterns of expression throughout TB therapy. Interestingly, there were striking immunological differences distinguishing the two groups. Group 1 comprised patients with high baseline CD4+CD25+ T cell frequencies and production of little IFN-γ and more IL-10 compared to group 2. By contrast, frequencies of CD4+CD25+ Tregs were increased only modestly in blood from patients in group 2, whose culture supernatants contained significantly higher amounts of IFN-γ and less IL-10.

At least two distinct subsets of CD4+CD25+ Tregs have been identified to date [13,17,41], namely naturally occurring and induced Tregs. Induced Tregs develop from CD4+CD25 CD4 T cells in response to activation [13,17,41,42]. This differentiation is critically dependent on the differentiation status of the antigen-presenting cell and the cytokine milieu during activation [13,17,34,35,41,43], with IL-10 and TGF-β acting as important promoters of CD4 T cell development into CD4+CD25+ (regulatory) T cells. Previous studies and our own observations indicate that levels of IFN-γ[44] and of immunosuppressive cytokines correlate with the extent of TB lung disease. A role for IL-10-producing T cells in the suppression of immune responses has also been described recently in anergic TB patients [45]. Thus, increased levels of immunosuppressive cytokines could result in the accentuated expansion of CD4+CD25+ Tregs during TB, particularly in patients with more extensive disease. In the current study, we found increased IL-10 levels in MTB-stimulated supernatants from TB patients with initial high frequencies of CD4+CD25hi+ T cells probably representing Tregs (group 1). Support for a causal role of IL-10 in the expansion of CD4+CD25+ Tregs during human TB stems from the fact that with the initiation of TB treatment IL-10 levels in supernatants of group 1 patients declined rapidly and in parallel with declining frequencies of CD4+CD25+ T cells. As stated, TGF-β was not assessed in the current study due to high background activities of the cytokine in whole blood cultures as a result of platelet degranulation.

Whereas expansion and the function of CD4+CD25+ Tregs at the time of diagnosis of active TB has been addressed in the present study, how this expansion of CD4+CD25+ Tregs is sustained at 6-month follow-up (completion of anti-TB therapy) remains unclear, particularly as our previous findings indicate that levels of TGF-β and IL-10 normalize even before the end of treatment [2]. One possible explanation is that MTB-responsive CD4 T cells present in the blood at the time of diagnosis of TB may be recruited to the lung once anti-TB therapy is initiated, to counteract the intense local inflammatory response in situ and to prevent the ensuing tissue damage. These cells then may recirculate back to the blood as the inflammation at the site of infection decreases. Alternatively, CD4+CD25+ T cells found in peripheral blood of TB patients may represent mutually exclusive populations; one preferentially representing immunoregulatory CD4+CD25+ T cells (found at TB diagnosis and early during follow-up) and a second population (found at end of therapy) without immunosuppressive activity. In support of this premise, Baecher-Allan et al. [13] demonstrated that human peripheral blood contained CD4+CD25+ cells, which could be separated into two major groups: CD25highCD4+ and CD25lowCD4+ cells. Only the CD25highCD4+ population appears to exhibit a strong regulatory function [13,41]. Interestingly, true Tregs which are of the CD4+CD25hi+ phenotype can be differentiated from activated CD4+CD25low+ T cells without regulatory properties by their constitutive expression of Foxp3 [32,46,47], human leucocyte antigen D-related (HLA-DR) and anti-cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) [41]. Foxp3, which was thought initially to be a specific marker predominantly for murine Tregs, also has been validated recently as a marker for human Tregs[47]. Thus, future studies may be directed at characterizing more extensively the phenotype of the subset of CD4+CD25hi+ T cells identified in the current study during the course of active TB.

In summary, data presented here indicate that active TB is associated with expansion of CD4+CD25+T cells overall, and specifically the CD4+CD25hi+ subset of CD4 T cells (probably representing true regulatory T cells), accounting, at least in part, for suppressed MTB-specific T cell responses during active TB. Further, persistence of increased frequencies of CD4+CD25hi+ Tregs beyond completion of anti-TB therapy may contribute to a delay in recovery of MTB-specific IFN-γ production. Further studies should be directed at identifying the mediators and mechanisms involved in the immunoregulatory properties of CD4+CD25+ T cells as a whole and to characterize more extensively the phenotype of CD4+CD25hi+ T cells during the course of human MTB infection.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We would like to thank the doctors (Drs Sa and Hadad) and nursing staff at the general clinical research centre (CPC) at UFES, Vitoria, Brazil for their involvement in patient recruitment, follow-up and specimen collection. In addition, we appreciate the assistance by the personnel at the NDI Microbiology laboratory with sputum processing and culture. Most importantly, however, we would like to acknowledge all the TB patients and healthy volunteers in Vitoria enrolled in the present study. Without their participation the project would not have been possible. This study was supported by funding from NIAID, NIH (RFP AI-45244/AI-95383 Tuberculosis Research Unit).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References