− regulatory T cells from peripheral blood lymphocytes of healthy Mycobacterium tuberculosis-infected humans

Microbes and Infection 9 (2007) 1325e1332 www.elsevier.com/locate/micinf

Original article

In vitro expansion of CD4þCD25highFOXP3þCD127low/ regulatory T cells from peripheral blood lymphocytes of healthy Mycobacterium tuberculosis-infected humans Jean-Michel Hougardy a, Virginie Verscheure a, Camille Locht b,c, Franc¸oise Mascart a,d,* a

Laboratory of Vaccinology and Mucosal Immunity, Universite´ Libre de Bruxelles (U.L.B.), Belgium b INSERM, U629, Lille, France c Institut Pasteur de Lille, F-59019 Lille Cedex, France d Immunobiology Clinic, Hoˆpital Erasme, 808 route de Lennik, B-1070 Brussels, Belgium Received 4 September 2006; accepted 13 June 2007 Available online 30 June 2007

Abstract CD4þCD25highFOXP3þ regulatory T (Treg) cells have recently been found at elevated levels in the peripheral blood of tuberculosis patients, compared to Mycobacterium tuberculosis latently infected (LTBI) healthy individuals and non-infected controls. Here, we show that CD4þCD25highFOXP3þ T lymphocytes can be expanded in vitro from peripheral blood mononuclear cells (PBMC) of LTBI individuals, but not of uninfected controls by incubating them with BCG in the presence of TGF-b. These expanded cells from the PBMC of LTBI subjects expressed CTLA-4, GITR and OX-40, but were CD127low/ and have therefore the phenotype of Treg cells. In addition, they inhibited in a dose-dependant manner the proliferation of freshly isolated mononuclear cells in response to polyclonal stimulation, indicating that they are functional Treg lymphocytes. In contrast, incubation of the PBMC with BCG alone preferentially induced activated CD4þ T cells, expressing CD25 and/or CD69 and secreting IFN-g. These results show that CD4þCD25highFOXP3þ Treg cells can be expanded or induced in the peripheral blood of LTBI individuals in conditions known to predispose to progression towards active tuberculosis and may therefore play an important role in the pathogenesis of the disease. Ó 2007 Elsevier Masson SAS. All rights reserved. Keywords: Tuberculosis; Human regulatory T cells; FOXP3; CD127

1. Introduction With nearly 2 million deaths and more than 8 million new cases annually [1], tuberculosis (TB) remains one of the most important infectious diseases worldwide, despite the fact that most Mycobacterium tuberculosis-infected individuals remain healthy and show no sign of disease during their entire lifetime. These latently infected (LTBI) individuals may thus be considered as protected against disease by an appropriate * Corresponding author. Immunobiology Clinic, Hoˆpital Erasme, 808 route de Lennik, B-1070 Brussels, Belgium. Tel.: þ32 2 555 34 67; fax: þ32 2 555 44 99. E-mail address: [email protected] (F. Mascart). 1286-4579/$ - see front matter Ó 2007 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.micinf.2007.06.004

immune response induced through natural infection. Immunological differences between LTBI individuals and patients with active disease have therefore been the subject of intense investigations, as they may reveal correlates of protection, the identification of which should be helpful for the development of new vaccines against TB. T cell responses are known since long to be essential for protection against TB [2], and several differences in T cell responses between subjects with LTBI and patients with active TB have already been identified. For example, T cell responses to certain protective antigens are often stronger during LTBI than during active disease [3e5], whereas T cell responses to other mycobacterial antigens, such as ESAT-6, are similar in the two groups [6]. Regulatory cytokines, such as

1326

J.-M. Hougardy et al. / Microbes and Infection 9 (2007) 1325e1332

TGF-b [7] and IL-10 [8], have been implicated in decreased T cell function during active TB, especially in anergic patients [8]. However, these cytokines display no antigen-specific suppressive properties. More recently, regulatory T (Treg) cells with the CD4þCD25highFOXP3þ phenotype have been detected at elevated levels among the peripheral blood mononuclear cells (PBMC) of TB patients compared to uninfected control subjects [9], or to LTBI individuals [10]. Ex vivo depletion of the CD4þCD25high T cells from PBMC of TB patients has been shown to result in enhanced IFN-g production upon stimulation with mycobacterial antigens, whereas this was not observed for PBMC of LTBI subjects [9e11], suggesting that these cells down-modulate T cell responses only during active TB. At the same time, they may also limit T cell-mediated immunopathology at the site of disease, even though their suppressive activity may not be sufficient to prevent it. These cells should thus be regarded as both friends and foes of the infected host [12]. Although apparently linked to TB disease, it is not known whether elevated frequencies of CD4þ CD25highFOXP3þ T cells in the PBMC may be the cause of progression from latent infection to active disease or a consequence of the disease. We reasoned that if expansion of CD4þCD25highFOXP3þ T cells predisposed LTBI subjects, thus M. tuberculosis-infected, yet still healthy, to progress to active TB and therefore occurred prior to the onset of the disease, it should be possible to generate or expand them in vitro from their PBMC, under conditions known to favour disease development. Here, we show that these cells can indeed be expanded or generated in vitro from PBMC of LTBI individuals, but not from those of uninfected control subjects, by incubation with BCG in the presence of TGF-b. 2. Materials and methods 2.1. Study subjects Venous blood samples were collected from seven LTBI and from four uninfected subjects. The latter were tuberculin-skin test negative, had not been vaccinated with BCG and had no known exposure to M. tuberculosis. The LTBI individuals were selected according to the recommended criteria from the Belgian national foundation against respiratory affections (FARES e www.fares.be): a risk stratified induration of 72 h after intradermal injection of 2 U tuberculin PPD-RT23 (Statens Serum Institute, Copenhagen), normal chest radiographs, and no clinical sign of TB. The study was approved by the local Ethical Committee of the hospital where the subjects were enrolled, and informed consent was given by all individuals. 2.2. CD25þ cell depletion and in vitro induction of CD25highFOXP3þ T cells PBMC were obtained by density gradient centrifugation of the blood on Lymphoprep (Nycomed Pharma, Oslo, Norway).

CD25þ T cells were then depleted from the PBMC by positive immunomagnetic selection using anti-CD25 antibody under conditions recommended by the supplier (EasySep Stemcell, Grenoble, France). The depleted PBMC were then cultured in the presence or absence of BCG at a multiplicity of infection of 5:1 with or without 5 ng/ml TGF-b1 (R&D, Mineapolis, MN) or IL-10 (R&D) during 96 h. IFN-g concentrations were measured in the cell culture supernatants, and the phenotype of the cells was characterised by flow cytometry. 2.3. Phenotypic analyses The surface expression of CD25, CD69, CD127 by the CD3þCD4þ lymphocytes, as well as intracellular expression of FOXP3, GITR, OX-40 and CTLA-4 were analysed on a 6-color Canto Flow Cytometer (BD Biosciences, San Jose´) after staining with PerCP-labelled anti-CD3, FITC-labelled or APC-Cy7-labelled anti-CD4, PE-labelled anti-CD25, APC-labelled anti-CD69, PE-labelled anti-CD127, PElabelled anti-CD152 (CTLA-4), PE-labelled anti-GITR, PElabelled anti-CD134 (OX-40), and APC-labelled anti-FOXP3 antibodies. All reagents were from BD Biosciences, except for the anti-FOXP3 antibodies that came from eBioscience (San Diego, CA) and for the anti-GITR antibodies purchased from R&D. The reagents were all used according to the manufacturer’s instructions. The data were analysed using the Flow JO software (v 6.3.4 Tree Star Inc. Ashland, OR). 2.4. IFN-g concentrations IFN-g concentrations were measured in the cell culture supernatants by ELISA, as previously described [13]. The capture and detection antibodies were, respectively, anti-human IFN-g IgG1 (clone 350B 10 G6) and biotin-labelled antihuman IFN-g IgG1 (clone 67F 12A8); both antibodies were purchased from BioSource International (Camarillo, CA). The standard curves were generated using serial dilutions of recombinant IFN-g (BioSource International). 2.5. Functional immuno-suppressive assays To analyse the suppressive function of the CD4þCD25highFOXP3þ T cells induced or expanded in vitro by incubation with BCG and TGF-b, mixed lymphocyte cultures were performed using these cells together with freshly isolated PBMC from the same LTBI subject. Fresh PBMC were labelled for 10 min with 0.5 mM CFSE (Vybrant CFDA SE-Cell Tracer kit, Molecular Probes, Eugene, OR) in PBSe BSA 0.1%, and then washed with fresh culture medium, before the addition of anti-CD3 (1 ng/ml, Orthoclone OKT3, Muromonab, Janssen-Cilag, Anvers, Belgium) and antiCD28 (5 mg/ml, Immunotech, Marseille, France) monoclonal antibodies. The PBMC containing the in vitro induced/expanded CD4þCD25highFOXP3þ cells were then added at different ratios, and the mixed lymphocyte cultures were incubated during 72 h at 37  C in a 5% CO2-containing atmosphere. The CD3þCD4þ T cell proliferation was then analysed

J.-M. Hougardy et al. / Microbes and Infection 9 (2007) 1325e1332

1327

by flow cytometry using the Flow JO software (v 6.3.4 Tree Star Inc., Ashland, OR). 2.6. Statistical analyses Statistical differences were assigned by the non-parametric Wilcoxon signed rank test (paired values). A value of P < 0.05 was considered to be significant. All results were obtained with the Graphpad Prism Software version 4.0b. 3. Results 3.1. In vitro expansion of CD4þCD25highFOXP3þ T cells in the presence of BCG and TGF-b PBMC from LTBI or uninfected subjects contain usually 1e5% CD4þCD25high T cells, although the frequency of these cells is increased in patients with active disease [9e11]. Therefore, the CD4þCD25high T cells present in the isolated PBMC were first depleted using immunomagnetic beads to allow us to analyse the CD4þCD25highT cell expansion/induction within a cell suspension containing only a very low proportion of CD4þCD25highT lymphocytes. Whereas the CD25highFOXP3þ T cell frequencies were 1.99% (median: 1.76e2.37; 25the75th percentiles) in the PBMC, residual frequencies of these cells after depletion were always lower than 0.25% of the CD4þCD3þ lymphocytes (median: 0.18%), as estimated by flow cytometric analysis. Incubation of the depleted PBMC from LTBI subjects (n ¼ 6) with BCG resulted in the appearance of CD4þCD25highFOXP3þ T cells after 96 h of incubation (Figs. 1 and 2, upper panels), since the percentages of CD4þ T lymphocytes expressing both CD25high and FOXP3 were significantly higher when the cells were incubated in the presence of BCG compared to culture medium alone (P ¼ 0.0313). However, the addition of TGF-b resulted in a further significant increase of the frequencies of these cells (P ¼ 0.0313), whereas TGF-b alone had no effect (Figs. 1 and 2, upper panels). In contrast, the addition of IL-10 to the culture medium had no effect, regardless of the presence of BCG (Fig. 1). When the depleted PBMC from uninfected individuals (n ¼ 4) were cultured in the presence of BCG with or without TGF-b, no significant induction of CD4þCD25highFOXP3þ cells was noted (P ¼ 0.350; Fig. 2, lower panels). The differences observed between non-infected controls and LTBI indicate that the CD4þCD25highFOXP3þ cells generated or explanded in vitro from the PBMC of LTBI subjects came from peripheral precursor cells induced by M. tuberculosis infection during latency.

Fig. 1. In vitro induction of CD4þCD25highFOXP3þ T cells in the PBMC from LTBI. CD25þ T cell-depleted PBMC from six LTBI were incubated for 96 h in the presence or absence of BCG, TGF-b, or IL-10 as indicated, and analysed by flow cytometry for the induction of CD4þCD25highFOXP3þ T cells. The results represent the percentages of the CD25highFOXP3þ cells among the CD4þCD3þ lymphocytes induced in the CD25 PBMC. The different conditions were analysed for the six samples except those with IL-10 that were only performed for three of them. The boxes represent 25th and 75th percentiles, vertical bars, minimum and maximum frequencies, and horizontal bars the medians. *P < 0.05.

they are Treg cells. To confirm that in the presence of TGFb, BCG mostly induces/expands Treg cells in the PBMC from LTBI individuals, rather than activated T cells, we analysed the expression of the early activation marker CD69 by CD4þ T cells and the secretion of BCG-induced IFN-g by the cultured PBMC from LTBI subjects. Significantly more CD4þ T cells expressed the T cell activation marker CD69 after the incubation with BCG alone compared to BCG and TGF-b (P ¼ 0.0312). As a consequence, the ratios of the CD4þCD25highFOXP3þ to the CD4þCD69þ T cell frequencies were significantly higher after incubation of the PBMC from LTBI subjects with BCG and TGF-b compared to BCG alone (Fig. 3A). Consistent with the surface expression of CD69, the T cells induced/expanded from the PBMC of LTBI subjects by BCG alone secreted high concentrations of IFN-g in the cell culture supernatants, whereas no significant amounts of IFN-g were detected when these cells were incubated with BCG in the presence of TGF-b (Fig. 3B). These results indicate that BCG alone predominantly induces activated T cells among the PBMC from LTBI individuals, whereas the addition of TGF-b to the BCG strongly tips the balance towards the CD25highFOXP3þCD4þ Treg cells.

3.2. CD69 expression and IFN-g secretion Although CD25 is a marker of regulatory T cells [14], it is also expressed on activated T cells, albeit usually at lower levels. However, the expression of FOXP3 in the CD4þCD25high T cell population induced/expanded in vitro from the PBMC of LTBI subjects, strongly suggests that

3.3. Further phenotypic characterisation of the in vitro induced/expanded CD25highFOXP3þCD4þ T cells Although the intracellular expression of FOXP3 is currently considered as the best and most specific marker of Treg cells in humans [15], other antigens have also been reported to be

1328

J.-M. Hougardy et al. / Microbes and Infection 9 (2007) 1325e1332

Fig. 2. Representative experiments of the in vitro induction of CD4þCD25highFOXP3þ T cells in the PBMC from one out of six LTBI subject (upper panel) and from one out of four uninfected control (lower panel). CD25þ T cell-depleted PBMC were incubated for 96 h in the presence or absence of BCG and TGF-b as indicated. The PBMCs were gated on the CD3þCD4þ T lymphocytes, and analysed for the surface expression of CD25 and for the intracellular expression of FOXP3. Criteria to define CD25high and FOXP3 positivity were determined on total PBMC. As human CD25high CD4þ lymphocytes are characterised by a slightly lower fluorescence intensity for the expression of CD4, the cut-off value of the fluorescence intensity defining the CD25high lymphocytes was defined by the lowest value of the CD25 fluorescence for the CD4lowCD25þ T cells’ population. The FOXP3 positivity was defined on the CD3þCD4 lymphocytes among PBMC. For further details, see Ref. [10]. The numbers in the dot plot indicate the percentage of the CD3þCD4þ T cells expressing the relevant markers.

Fig. 3. (A) Ratios between the percentages of the CD25highFOXP3þ and the CD69þ cells among the CD3þCD4þ lymphocytes. CD25þ T cell-depleted PBMC from six LTBI were incubated for 96 h in the presence of BCG with or without TGF-b as indicated. The results represent the ratios between the percentages of the CD25highFOXP3þ and the CD69þ cells among the CD4þCD3þ lymphocytes induced in the CD25 PBMC. The boxes represent 25th and 75th percentiles, vertical bars, minimum and maximum ratios, and horizontal bars the medians. *P < 0.05. (B) IFN-g concentrations secreted by the CD25 PBMC from six LTBI incubated for 96 h in the presence of BCG with or without TGF-b as indicated, prior to the measurement of the IFN-g concentration in the culture supernatants. The columns represent the medians and the vertical bars the 25th and 75th percentiles of the results. *P < 0.05.

expressed on Treg cells, such as CTLA-4, the glucocorticoidinduced tumor necrosis factor receptor family-related proteins (GITR) and OX-40 [16,17]. However, these markers are also expressed on potent effector T cells, but their use in combination with other Treg cell markers has been proposed to help defining the CD4þ Treg cell population [18]. We have therefore used these markers to further characterise the phenotype of the CD4þCD25highFOXP3þ T cells induced/expanded in vitro in the presence of BCG and TGF-b. As shown in Fig. 4, the in vitro induced/expanded CD4þCD25highFOXP3þ T cells also preferentially express CTLA-4, GITR and OX-40 when compared to the CD4þCD25 or the CD4þCD25highFOXP3 T cells. These results further argue that in the presence of mycobacterial antigens and TGF-b, Treg cells are expanded or induced from the CD25 PBMC of LTBI individuals, with pre-existing memory T cells specific for mycobacterial antigens. Very recently, the IL-7 receptor (CD127) has been reported to inversely correlate with FOXP3 expression on human CD4þ Treg cells, and may constitute the so far best cell surface marker for the identification of CD4þFOXP3þ Treg cells by the CD4þCD25þCD127low/ phenotype [19]. The CD4þCD25þ CD127low/ cell subset accounts for w80% of the CD4þ FOXP3þ cells among the PBMC of normal individuals [19]. We therefore analysed CD127 expression on the in vitro induced/expanded CD4þCD25highFOXP3þ T cells. As shown in

J.-M. Hougardy et al. / Microbes and Infection 9 (2007) 1325e1332

1329

Fig. 4. Expression of CTLA-4, GITR, and OX-40 on the different T lymphocyte subsets. After gating the CD3þ CD4þ lymphocytes for their CD25 and FOXP3 expression, the indicated subsets were analysed for their CTLA-4, GITR, or OX-40 expression. The numbers in the histograms indicate the percentages of gated cells expressing the marker indicated in the left margin. For each marker analysed, the cell positivity was determined by reference to a negative population for this marker, the CD3 lymphocytes.

Fig. 5, the CD4þCD25highFOXP3þ T cells showed low expression of CD127, as evidenced by their low median of CD127 fluorescence compared to that of the CD4þCD25FOXP3 T cells (Fig. 5, upper right panel). In contrast, the median of CD127 fluorescence of the CD4þCD25highFOXP3 T cells was similar to that of the CD4þCD25FOXP3 T cells (Fig. 5, lower right panel). These observations provide further evidence that the CD4þCD25highFOXP3þ T cells expanded/induced in vitro from the PBMC of LTBI subjects by incubation with BCG in the presence of TGF-b are indeed Treg cells. 3.4. Suppressive function of the PBMC from LTBI subjects cultured in the presence of BCG and TGF-b Mixed lymphocyte cultures were performed using freshly isolated PBMC from an LTBI subject incubated at different ratios with PBMC containing the in vitro induced/expanded CD4þCD25highFOXP3þ T cells from the same LTBI subject

to demonstrate the suppressive function of the induced/expanded CD4þCD25highFOXP3þ T cells on the proliferative capacity of the fresh cells. As shown in Fig. 6 (upper right panel), in vitro stimulation with anti-CD3 and anti-CD28 of CFSE-labelled PBMC freshly isolated from a subject with LTBI induced an important proliferation of the CD3þ T cells compared to non-stimulated cells (upper left panel). However, when these cells were stimulated with the same antibodies in the presence of PBMC from the same individual containing the CD4þCD25highFOXP3þ T cells induced/expanded in vitro by BCG and TGF-b, the proliferation was strongly inhibited (Fig. 6, lower panels). This inhibition was proportional to the amount of the CD4þCD25highFOXP3þ T cell-containing PBMC added in the assay, confirming thereby that these in vitro induced/expanded cells are functional Treg cells. When the CD4þCD25high cells were depleted from the in vitro induced/expanded population before addition in the mixed lymphocyte cultures, no suppressive effect on the proliferative

1330

J.-M. Hougardy et al. / Microbes and Infection 9 (2007) 1325e1332

Fig. 5. Expression of CD127 on different T lymphocyte subsets. Among the CD3þCD4þ lymphocytes, CD25high lymphocytes and FOXP3þ (upper histogram) or FOXP3 (lower histogram) cells were analysed for their CD127 expression. Results obtained for the gated cells are represented by the filled histograms compared to those obtained for the CD3þCD4þCD25low/ FOXP3 lymphocytes represented by the open histograms.

Fig. 6. Proliferation of anti-CD3 and anti-CD28 antibody-stimulated PBMC in the presence or absence of PBMC containing CD4þCD25highFOXP3þ T cells induced in vitro by BCG and TGF-b. The CD3þ T cell proliferation was measured using the CFSE labeling method. The two upper panels show the number of proliferating PBMC in the absence or presence of the stimulating agent (left and right, respectively). The lower panels show the numbers of proliferating PBMC in response to the polyclonal stimulation, in the presence of CD4þCD25highFOXP3þ T cell-containing PBMC at the indicated suppressor to target PBMC ratios. The numbers on the panels indicated the percentages of dividing cells.

J.-M. Hougardy et al. / Microbes and Infection 9 (2007) 1325e1332

capacity of the fresh PBMC was observed (data not shown). These results indicate that the CD4þCD25highFOXP3þ T cells generated or expanded in vitro from the PBMC isolated from LTBI subjects are functional Treg cells. 4. Discussion TB patients have recently been reported to be characterised by a high proportion of circulating Treg cells with the CD4þCD25highFOXP3þ phenotype, in comparison to LTBI subjects or uninfected controls [9,10]. In the diseased patients these cells have been shown to suppress T cell responses to mycobacterial antigens, whereas the CD4þCD25highFOXP3þ cells of non-diseased individuals do not suppress the IFN-g secretion induced by protective mycobacterial antigens [10,11]. Elevated levels of functional CD4þCD25highFOXP3þ Treg cells in the peripheral blood of TB patients may be a consequence of the disease. Alternatively, they may be its cause, which would imply that they are generated or expanded during latent infection prior to the onset of disease. Although CD4þCD25highFOXP3þ Treg cells are naturally generated in the thymus [20,21], they can also be induced in the periphery, in both mice [22] and humans [23,24], suggesting that peripheral Treg cells may arise from antigenic challenge during the course of an immune response [24]. We demonstrate here that, in the presence of mycobacterial antigens and TGF-b, CD4þCD25highFOXP3þ T cells are expanded or induced in vitro from the CD25 PBMC of LTBI individuals. These cells were only generated from the PBMC of infected individuals, and not from those of naive controls, suggesting that they are generated from antigenic-specific precursor cells, which are most likely memory cells primed by the first encounter with the mycobacteria during primary infection. Several lines of evidence, based on their phenotype, strongly suggest that the induced/expanded CD4þCD25high FOXP3þ T cells are Treg cells, rather than activated T cells. The lymphocytes induced/expanded in the presence of BCG and TGF-b do not secrete antigen-specific IFN-g, and the balance between the expression of markers of regulatory (CD25high and FOXP3) and recently activated cells (CD69) is strongly in favour of the Treg cells. These cells have first been identified by a high expression of CD25 [14], and until very recently, FOXP3 was considered as the best and most specific marker of Treg cells, although some studies have questioned whether all Treg cells are FOXP3þ and whether all FOXP3þ T cells are Treg cells [15]. We have therefore further characterised the phenotype of these cells and found that they preferentially express CTLA-4, GITR, and OX-40, but show low expression of the IL-7 receptor, a very recently reported, most discriminatory marker of functional Treg cells [19]. These phenotypic characteristics argue strongly in favour of the induced/expanded CD4þCD25highFOXP3þ T cells as being Treg cells. In addition, the suppressive function of the in vitro induced/expanded cells was demonstrated in mixed lymphocyte assays. The addition of PBMC from an LTBI subject containing the in vitro induced/expanded CD4þCD25high FOXP3þ T cells to fresh PBMC from the same individual

1331

strongly inhibited their proliferative response in a regulatory to target cell ratio dependant manner, whereas no inhibition was seen when the induced/expanded CD4þCD25high T cells were depleted prior to co-culturing with the fresh PBMC. The presence of TGF-b together with BCG was important during the in vitro induction/expansion of the CD4þ CD25highFOXP3þ T cells. In the absence of TGF-b, BCGinduced essentially activated T cells expressing the activation marker CD69 and were able to secrete antigen-specific IFN-g. However, when TGF-b was included, the induced/expanded cells were essentially of the Treg cell phenotype. TGF-b alone did not induce these cells, nor did the addition of IL-10. The essential role of TGF-b for the peripheral induction/expansion of CD25þFOXP3þ Treg cells has been shown in mice, and it was suggested that the induction of FOXP3 expression requires TGF-b and TCR costimulation [22]. The critical involvement of TGF-b in the generation of the mycobacteria-induced CD4þCD25highFOXP3þ Treg cells has important implications for the pathogenesis of TB. It provides a possible mechanistic explanation of the fact that pathologies known to result in high levels of circulating TGF-b, such as diabetes [25], anthracosilicosis [26] and certain nephropathies [27], are associated with a high risk to develop active TB. In addition, the demonstration that BCG plays a role in the induction/expansion of Treg cells in subjects that have already encountered mycobacteria may have practical consequences for strategies of the prevention of TB. BCG is the only currently available vaccine against TB and it has shown limited effectiveness, especially against pulmonary TB in adolescents and adults [28]. In fact, it has been shown to increase severity of TB when given after exposure to M. tuberculosis [29], and repeated vaccinations with BCG have been associated with increased risk of developing active TB in highly endemic regions, where pre-exposure to M. tuberculosis is very likely [30]. In light of the results presented in this study, it is thus likely that in preexposed but healthy individuals an additional antigenic charge by repeated BCG vaccination may result in the generation of peripheral CD4þCD25highFOXP3þ Treg cells that down-regulate the protective anti-mycobacterial immune response and therefore favours progression towards active disease. Peripheral CD4þCD25highFOXP3þ Treg cells may thus play a role in the early stages of the pathogenesis of TB, as they may be generated/expanded from PBMC of LTBI and as they have been previously shown to depress cellular immune responses in non-anergic patients with pulmonary TB [9,10]. Although we cannot formally exclude that local lymphocytes may respond differently, this leads us to propose the following most likely scenario for the progression from latent infection to active disease. After M. tuberculosis infection, in the vast majority of cases, a protective cell-mediated immune response controls the pathogen over many years, most often during lifetime, without clinical consequences. When for any reason circulating TGF-b levels rise, even transiently, and the infected individual is exposed to mycobacterial antigens, either because of the latent infection, of re-exposure to M. tuberculosis or of post-exposure BCG vaccination, CD4þCD25highFOXP3þ Treg cells are generated or expand.

1332

J.-M. Hougardy et al. / Microbes and Infection 9 (2007) 1325e1332

These cells then suppress the effective immune responses to protective antigens, allowing the mycobacteria to escape from the control of the immune system, to proliferate and to colonise the vulnerable zones, including the lungs. The resulting increased bacterial load can further stimulate the CD4þCD25highFOXP3þ Treg cell expansion, induce the production of TGF-b [7], and finally the production of IL-10 [8] evolving into late-stage anergy. The generation of CD4þCD25highFOXP3þ Treg cells that suppress T cell responses to protective antigens could thus represent a major strategy early in the pathogenesis to evade the host protective immunity against M. tuberculosis and to ultimately trigger active disease. Acknowledgements This work was supported by a grant from the ‘‘Fond de la Recherche Scientifique Me´dicale’’ and by the European Commission within the 6th Framework Program, contract no. LSHP-CT-2003-503367. JMH was supported by a fellowship from the ‘‘Fond pour la Formation a` la Recherche dans l’Industrie et dans l’Agriculture’’ (FRIA). We thank S.T. Temmerman for helpful discussions. References [1] C. Dye, Global epidemiology of tuberculosis, Lancet 367 (2006) 938e940. [2] J.L. Flynn, J. Chan, Immunology of tuberculosis, Annu. Rev. Immunol. 19 (2001) 93e129. [3] K. Huygen, J.P. Van Vooren, M. Turneer, M. Bosmans, R. Dierckx, J. De Bruyn, Immunoproliferation, g-interferon production, and serum immunoglobulin G directed against a purified 32 kDa mycobacterial protein antigen (P32) in patients with active tuberculosis, Scand. J. Immunol. 27 (1988) 187e194. [4] C. Masungi, S. Temmerman, J.P. Van Vooren, A. Drowart, K. Pethe, F.D. Menozzi, C. Locht, F. Mascart, Differential T and B cell responses against Mycobacterium tuberculosis heparin-binding hemagglutinin adhesion in infected healthy individuals and patients with tuberculosis, J. Infect. Dis. 185 (2002) 513e520. [5] S. Temmerman, K. Pethe, M. Parra, S. Alonso, C. Rouanet, T. Pickett, A. Drowart, A.S. Debrie, G. Delogu, F.D. Menozzi, C. Sergheraert, M.J. Brennan, F. Mascart, C. Locht, Methylation-dependent T cell immunity to Mycobacterium tuberculosis heparin-binding hemagglutinin, Nat. Med. 10 (2004) 935e941. [6] A. Demissie, E.M. Leyten, M. Abebe, L. Wassie, A. Aseffa, G. Abate, H. Fletcher, P. Owiafe, P.C. Hill, R. Brookes, G. Rook, A. Zumla, S.M. Arend, M. Klein, T.H. Ottenhoff, P. Andersen, T.M. Doherty, the VACSEL Study Group, Recognition of stage-specific mycobacterial antigens differentiates between acute and latent infections with Mycobacterium tuberculosis, Clin. Vaccine Immunol. 13 (2006) 179e186. [7] C.S. Hirsch, R. Hussain, Z. Toossi, G. Dawood, F. Shahid, J.J. Ellner, Cross-modulation by transforming growth factor beta in human tuberculosis: suppression of antigen-driven blastogenesis and interferon gamma production, Proc. Natl. Acad. Sci. USA 93 (1996) 3193e3198. [8] V.A. Boussiotis, E.Y. Tsai, E.J. Yunis, S. Thim, J.C. Delgado, C.C. Dascher, A. Berezovskaya, D. Rousset, J.M. Reynes, A.E. Golfeld, IL-10-inducing T cells suppress immune responses in anergic tuberculosis patients, J. Clin. Invest. 105 (2000) 1317e1325. [9] V. Guyot-Revol, J.A. Innes, S. Hackforth, T. Hinks, A. Lalvani, Regulatory T cells are expanded in blood and disease sites in patients with tuberculosis, Am. J. Crit. Care Med. 173 (2006) 803e810.

[10] J.M. Hougardy, S. Place, M. Hildebrand, A. Drowart, A.S. Debrie, C. Locht, F. Mascart, Regulatory T cells depress immune responses to protective antigens in active tuberculosis, Am. J. Respir. Crit. Care Med. 176 (2007) 409e416. [11] R. Ribeiro-Rodrigues, T. Resende Co, R. Rojas, Z. Toossi, R. Dietze, W.H. Boom, E. Maciel, C.S. Hirsch, A role for CD4þ CD25þ T cells in regulation of the immune response during human tuberculosis, Clin. Exp. Immunol. 144 (2006) 25e34. [12] K.H. Mills, Regulatory T cells: friend or foe in immunity to infection? Nat. Rev. Immunol. 4 (2004) 841e855. [13] F. Mascart, V. Verscheure, A. Malfroot, M. Hainaut, D. Pie´rard, S. Temerman, A. Peltier, A.S. Debrie, J. Levy, G. Del Giudice, C. Locht, Bordetella pertussis infection in 2-month-old infants promotes type 1 T cells responses, J. Immunol. 170 (2003) 1504e1509. [14] C. Baecher-Allan, J.A. Brown, G.J. Freeman, D.A. Hafler, CD4þCD25high regulatory cells in human peripheral blood, J. Immunol. 167 (2001) 1245e1253. [15] S.F. Ziegler, FOXP3: of mice and men, Annu. Rev. Immunol. 24 (2006) 209e226. [16] T. Takahashi, T. Tagami, S. Yamazaki, T. Uede, J. Shimizu, N. Sakaguchi, T.W. Mak, S. Sakaguchi, Immunologic self-tolerance maintained by CD25(þ)CD4(þ) regulatory T cells constitutively expressing cytotoxic T lymphocyte-associated antigen 4, J. Exp. Med. 192 (2000) 303e310. [17] R.S. McHugh, M.J. Whitters, C.A. Piccirillo, D.A. Young, E.M. Shevach, M. Collins, M.C. Byrne, CD4(þ) CD25(þ) immunoregulatory T cells: gene expression analysis reveals a functional role for the glucocorticoid-induced TNF receptor, Immunity 16 (2002) 303e323. [18] H. Jiang, L. Chess, Regulation of immune responses by T cells, N. Engl. J. Med. 354 (2006) 1166e1176. [19] W. Liu, A.L. Putman, Z. Xu-yu, G.L. Szot, M.R. Lee, S. Zhu, P.A. Gottlieb, P. Kapranov, T.R. Gingeras, B.F. de St. Groth, C. Clayberger, D.M. Soper, S.F. Ziegler, J.A. Bluestone, CD127 expression inversely correlates with FOXP3 and suppressive function of human CD4þ T reg cells, J. Exp. Med. 203 (2006) 1701e1711. [20] M. Itoh, T. Takahashi, N. Sakagushi, Y. Kuniyasu, J. Shimizu, F. Otsuka, S. Sakaguchi, Thymus and autoimmunity: production of CD25þCD4þ naturally anergic and suppressive T cells as a key function of, the thymus in maintaining immunologic self-tolerance, J. Immunol. 162 (1999) 5317e5326. [21] I. Apostolou, A. Sarukhan, L. Klein, H. von Boehmer, Origin of regulatory T cells with known specificity for antigen, Nat. Immunol. 3 (2002) 756e763. [22] W.J. Chen, W. Jin, N. Hardegen, K.J. Lei, L. Li, N. Marinos, G. McGrady, S.M. Wahl, Conversion of peripheral CD4þCD25 naive T cells to CD4þCD25þ regulatory T cells by TGF-ß induction of transcription factor Foxp3, J. Exp. Med. 198 (2003) 1875e1886. [23] M.R. Walker, J.D. Kasprowicz, V.H. Gersuk, A. Benard, M. Van Landeghen, J.H. Buckner, S.F. Ziegler, Induction of FoxP3 and acquisition of T regulatory activity by stimulated human CD4þCD25 T cells, J. Clin. Invest. 112 (2003) 1437e1443. [24] T.A. Chatila, Role of regulatory T cells in human diseases, J. Allergy Clin. Immunol. 116 (2005) 949e959. [25] L. Flores, S. Naf, R. Hernaez, I. Conget, R. Gomis, E. Esmatjes, Transforming growth factor-ß at clinical onset of Type 1 diabetes mellitus. A pilot study, Diabet. Med. 21 (2004) 818e822. [26] J. Jagirdar, R. Begin, A. Dufresne, S. Goswami, T.C. Lee, W.N. Rom, Transforming growth factor-ß (TGF-ß) in silicosis, Am. J. Respir. Crit. Care Med. 154 (1996) 1076e1081. [27] T. Yamamoto, N.A. Noble, A.H. Cohen, C.C. Nast, A. Hishida, L.I. Gold, W.A. Border, Expression of transforming growth factor-ß isoforms in human glomerular diseases, Kidney Int. 49 (1996) 461e469. [28] P.E. Fine, Variation in protection by BCG: implications of and for heterologous immunity, Lancet 346 (1995) 1339e1345. [29] J. Turner, E.R. Rhoades, M. Keen, J.T. Beliste, A.A. Frank, I.M. Orme, Effective preexposure tuberculosis vaccines fail to protect when they are given in an immunotherapeutic model, Infect. Immun. 68 (2000) 1706e1709. [30] Karonga Prevention Trial Group, Randomised controlled trial of single BGG, repeated BCG, or combined BCG and killed Mycobacterium leprae vaccine for the prevention of leprosy and tuberculosis in Malawi, Lancet 348 (1996) 17e24.