Cutting Edge: Regulatory T Cells Prevent Efficient Clearance of Mycobacterium tuberculosis

Tuberculosis remains one of the most threatening bacterial infections, which is responsible for high incidences of morbidity and mortality, reaching ∼2 million deaths annually (1). However, mortality figures are only one facet of this disease. Approximately one-third of the world’s population is latently infected with Mycobacterium tuberculosis, and 90% of these individuals will never develop active disease, indicating that the human immune system is capable of controlling M. tuberculosis infection effectively. Generally, host immunity, however, fails to eradicate tubercle bacilli successfully. In the vast majority of cases, a chronic infection develops and a balance between pathogen persistence and immune response is maintained, perpetuating the risk of uncontrolled reactivation. The mechanisms underlying this fine-tuned balance, which is a likely consequence of the coevolution of both organisms, are incompletely understood. Probably, they involve sophisticated survival and immune escape strategies of M. tuberculosis and aberrant immunoregulatory host mechanisms, which intervene during chronic infection to prevent exacerbated immunopathology.

M. tuberculosis is generally transmitted via inhalation of mycobacteria-containing droplets. Within the lung alveolar space, bacilli are engulfed by alveolar macrophages and are either killed immediately or end up in deeper lung tissue where they reside within local macrophages. In an attempt to isolate the tissue site of infection, the immune response induces granuloma formation. Although various cells contribute to immunity against M. tuberculosis, T cells, notably effector CD4⁺ T cells of the Th1 type, dominate protective immunity (2, 3, 4). Upon activation, CD4⁺ T cells secrete IFN-γ and TNF-α, which in turn induce antimycobacterial mechanisms in macrophages. These include hydrogen peroxide and other reactive oxygen intermediates generated via the oxidative burst, as well as NO and related reactive nitrogen intermediates, produced by the inducible form of NO synthase (NOS2) (2, 5, 6). These and additional macrophage effector mechanisms effectively control M. tuberculosis in vitro, and studies using gene knockout mice indicate that reactive nitrogen intermediates and reactive oxygen intermediates are critical in vivo (7, 8, 9, 10). Thus, it is unclear why the immune system only restricts mycobacterial growth and fails to achieve sterile eradication of this pathogen.

Critical involvement of regulatory T cells (T_reg)³ in control of immune responses to self-Ags and in immune homeostasis is well established, and there is increasing evidence for a role of T_reg in the regulation of immunity to infection (11, 12). It is believed that these T_reg down-modulate immune responses after pathogen eradication to avoid exacerbated pathology. Although this mechanism is generally to the benefit of the host in acute infections, it poses problems in chronic infections, notably when pathogen persistence is sustained in the face of an active immune response.

T_reg represent 5–10% of the CD4⁺ T cell population and are characterized by the expression of the transcription factor Foxp3. A large fraction of the T_reg population is CD25⁺, and T_reg represent up to 90% of the CD4⁺CD25⁺ T cell subset in naive animals (13). The mechanisms by which T_reg control immune responses are incompletely understood, but there is evidence for a central role of the inhibitory cytokines IL-10 and TGF-β and the surface molecule CTLA-4 (12, 14). The role of T_reg in tuberculosis is thus far unknown and is central in the current study.

C57BL/6 and Rag1^−/− mice on a C57BL/6 background were purchased from the Federal Institute for Health Protection of Consumers and Veterinary Medicine in Berlin and bred in our facilities. All animal experiments were conducted according to German animal protection law.

Mice were infected by aerosol with 200 CFU of M. tuberculosis H37Rv using an aerosol chamber (Glas-Col). Inocula were confirmed at day 1 postinfection by plating complete lung homogenates onto Middlebrook 7H11 ampicillin plates. CFU in the lung and spleen were determined after mechanical disruption of organs in water supplemented with 1% albumin and 0.5% Tween 80, and plating serial dilutions onto Middlebrook 7H11 ampicillin agar at day 30 postinfection.

Cells from spleen and peripheral lymph nodes of naive C57BL/6 donor mice were incubated with purified rat Ig and anti-CD16/CD32 mAb (clone, 2.4G2). After 5–10 min, biotinylated anti-CD25 mAb (clone PC61) and magnetic antibiotin microbeads (Miltenyi Biotec) were added, and CD25⁺ T cells were purified with a magnetic column (autoMACS; Miltenyi Biotec). T cells bound and recovered from the column were >75% CD25⁺, >90% CD4⁺, and >80% Foxp3⁺. CD25-depleted cells were incubated with FITC-conjugated anti-CD4 mAb (GK1.5) and subsequently with anti-FITC mAb-coated magnetic microbeads (Miltenyi Biotec). CD4⁺ T cells were isolated using the autoMACS and were >98% CD4⁺, <0.5% CD25⁺, and <4.0% Foxp3⁺. Purified CD4⁺CD25⁻ T cells (2 × 10⁶) or CD4⁺CD25⁺ T cells (1 × 10⁶) or a combination of both cell populations were adoptively transferred (i.v.) into Rag 1^−/− recipient mice. At day 1 after cell transfer, mice were infected with M. tuberculosis H37Rv by aerosol.

Spleens were removed, and single-cell suspensions were prepared using an iron mesh sieve. RBC were lysed, and spleen cells were washed twice with RPMI 1640 medium supplemented with glutamine, Na-pyruvate, 2-ME, penicillin, streptomycin, and 10% heat-inactivated FCS (complete RPMI 1640). Lungs were cut into small pieces and incubated for 30 min with complete RPMI 1640 supplemented with Collagenase D (Roche) and collagenase type VIII (Sigma-Aldrich). Single-cell suspensions of lungs were then prepared in a manner similar to those of spleens, using an iron mesh sieve and RBC lysis. Foxp3 staining was performed using the PE anti-Foxp3 staining kit (eBioscience) according to the manufacturer’s recommendations. For the determination of cytokine expression, 4 × 10⁶ cells were cultured for 6 h in a volume of 1-ml complete RPMI 1640. Cells were stimulated by adding 5 μg/ml anti-CD3 mAb and 5 μg/ml anti-CD28 mAb or as a negative control, with no supplementary additives. During the final 5 h of culture, 10 μg/ml brefeldin A were added. Cultured cells were washed and incubated for 10 min with rat IgG Abs and anti-CD16/CD32 mAb to block nonspecific Ab binding. Subsequently, cells were stained with PE-Cy5-conjugated anti-CD4 mAb and FITC-conjugated anti-Thy1.2. After 20 min on ice, cells were washed with PBS and fixed for 20 min at room temperature with PBS 2% paraformaldehyde. Cells were washed with PBS/0.2% BSA, permeabilized with PBS/0.1% BSA/0.5% saponin, and incubated in this buffer with rat IgG Abs and anti-CD16/CD32 mAb. After 10 min, Cy5-conjugated anti-IFN-γ or TNF-α mAb and PE-conjugated anti-IL10 mAb or isotype control mAb were added. After a further 20 min at room temperature, cells were washed with PBS and fixed with PBS/1% paraformaldehyde. Cells were analyzed using a FACS-Canto II and Diva software (BD Biosciences).

Rat IgG Abs, anti-CD16/CD32 mAb, anti-IFN-γ mAb, and anti-TNF-α mAb were purified from rat serum or hybridoma supernatants with protein G Sepharose. Abs were Cy5-conjugated according to standard protocols. Rat IgG1 isotype control mAb was purchased from BD Pharmingen. The anti-CD25 and anti-Foxp3 mAb were obtained from eBioscience, and PE-Cy5-conjugated anti-CD4 (RM4-5), FITC-conjugated anti-Thy-1.2, and PE-conjugated anti-IL-10 mAb were obtained from BD Biosciences.

Total RNA was extracted from tissues using TRIzol reagent (Invitrogen Life Technologies) as recommended by the manufacturer. RNA from animals of similarly treated groups was pooled, and samples were treated with DNase I (Invitrogen Life Technologies) to eliminate genomic DNA contamination. For semiquantitative real-time RT-PCR, 5 μg of DNase I-digested RNA-pools were used for reverse transcription using 200 ng of random hexamers as primers for Superscript II (Invitrogen Life Technologies) according to the manufacturer’s recommendation. All real-time PCR were run for 45 cycles with 20 s 94°C and 60 s 60°C in the ABI Prism 7000 Sequence Detection System (Applied Biosystems) using ABI PRISM optical 96-well plates (Applied Biosystems). Primers were designed to span large introns and to produce product sizes between 100 and 200 bp (β-actin-forward (fw), TGG AAT CCT GTG GCA TCC ATG AAA C; β-actin-reverse (rev), TAA AAC GCA GCT CAG TAA CAG TCC G; GAPDH-fw, GCA ACT CCC ACT CTT CCA CCT TC; GAPDH-rev, CCT CTC TTG CTC AGT GTC CTT GCT; Foxp3-fw, GCC TCT CCA GAG AGA AGT GGT G; Foxp3-rev, CCC TGA GTA CTG GTG GCT ACG ATG; IFN-γ-fw, ACG GCA CAG TCA TTG AAA GCC TA; IFN-γ-rev, CTC ACC ATC CTT TTG CCA GTT CC; IL-10-fw, GGA CAA CAT ACT GCT AAC CGA CTC CT; IL-10-rev, CTG CTC CAC TGC CTT GCT CTT ATT; inducible NO synthase (iNOS)-fw, GAC GAG ACG GAT AGG CAG AGA TTG; iNOS-rev, CCT GGG AGG AGC TGA TGG AGT AG). Reaction mixtures were set up to a final volume of 30 μl using 15 pmol of each primer and 15 μl of 2× SYBR Green PCR Master mix (Applied Biosystems). Quantification was performed at least two times with independent cDNA samples and in triplicates for each cDNA and primer pair. Data analyses were performed using the ABI Prism 7000 SDS Software (Applied Biosystems) and Excel (Microsoft). The threshold cycles were determined for each sample, and fold differences relative to the expression level in one of the analyzed cDNA samples was calculated for each cDNA sample and primer pair (fold-difference = 2^−ΔCt). Resulting fold differences for Foxp3, IFN-γ, IL-10, and iNOS expression levels were corrected for different amounts of cDNA by multiplication with the average fold difference of GAPDH and β-actin expression within the same sample. DNase I-digested, not reverse-transcribed RNA, was used as a template in separate reactions, to control for genomic DNA contamination.

Statistical significance of results was determined with the statistics program included in the GraphPad Prism software (GraphPad). Bacterial titers were analyzed using the Mann-Whitney U test for a nonparametric sample distribution (p < 0.05).

In the first set of experiments, we determined the number of Foxp3-expressing CD4⁺ T cells in lung and spleen at different time points after M. tuberculosis infection. During the first 60 days postaerosol M. tuberculosis infection, a consistent increase of FoxP3-expressing CD4⁺ T cells was detected in spleen as well as in lung, indicating infiltration of T_reg into infected tissue sites (Fig. 1 A).

To evaluate the impact of T_reg in protection against M. tuberculosis, we purified CD4⁺ T cells from naive C57BL/6 mice, separated CD4⁺CD25⁻ cells (Th) from CD4⁺CD25⁺ cells (T_reg), and adoptively transferred these T cell subsets into mice lacking T and B cells (Rag1^−/− mice). At day 30 postaerosol infection with M. tuberculosis, we analyzed the mycobacterial burden in lungs and spleens (Fig. 1 B). Rag1^−/− mice reconstituted with selected Th controlled M. tuberculosis more efficiently than Rag1^−/− controls. Notably, Rag1^−/− mice reconstituted with Th harbored markedly lower bacterial titers in the order of several magnitudes as compared with C57BL/6 wild-type (wt) mice. In some Th-reconstituted Rag1^−/− mice, the bacterial burden was below the detection limit of our assay, suggesting sterile eradication. In contrast, protection was reduced to that seen in C57BL/6 wt mice when Th and T_reg were cotransferred into Rag1^−/− animals. Hence, Th can effectively control M. tuberculosis but are severely restricted in their function by T_reg.

To gain insights into the mechanism underlying protection, we analyzed the composition of the T cell compartment at day 30 after M. tuberculosis infection in C57BL/6 mice and Rag1^−/− mice that received either Th alone or Th and T_reg before infection using flow cytometry and intracellular cytokine staining (Fig. 2). Transfer of Th without T_reg led to high numbers and frequencies of IFN-γ-producing CD4⁺ T cells, whereas cotransfer of Th and T_reg not only prevented the development of IFN-γ-producing CD4⁺ T cells among the cotransferred Th cells, but also suppressed Th proliferation (Fig. 2 B). However, despite higher frequencies, the total number of IFN-γ-producing CD4⁺ T cells was still higher in less well-protected wt mice as compared with Rag1^−/− mice that received selected Th. Notably, the level of protection correlated with the frequency, but not with the number of IFN-γ producers among CD4⁺ T cells.

Because IFN-γ is a critical mediator of the protective immune response against M. tuberculosis (15, 16), we aimed to determine whether enhanced protection in Rag1^−/− mice that received selected Th was due to increased IFN-γ production despite lower numbers of IFN-γ-producing CD4⁺ T cells. To compare total IFN-γ mRNA levels, we performed real-time RT-PCR analyses of lung and spleen tissue from naive C57BL/6 and infected C57BL/6 mice, and T cell reconstituted Rag1^−/− mice (Fig. 3). Notably, IFN-γ mRNA levels did not correlate with protection against M. tuberculosis. Spleen and lung levels of IFN-γ mRNA were equally enhanced in infected Th recipient Rag1^−/− mice and C57BL/6 mice. Expression of iNOS is induced by IFN-γ and represents one of the key effector mechanisms of macrophages against M. tuberculosis (9, 10). We thus analyzed iNOS expression in the spleen and lung. Similar to IFN-γ expression, we observed strong induction of iNOS in infected C57BL/6 wt mice as well as in Th-reconstituted Rag1^−/− mice. However, after infection, Rag1^−/− mice without T cell reconstitution showed enhanced IFN-γ and iNOS expression as well. Therefore, our results are inconsistent with a direct correlation between strength of protection and abundance of IFN-γ and iNOS, particularly after Th reconstitution of Rag1^−/− mice. This observation is in accordance with results previously described by Scanga et al. (17), who found that CD4⁺ T cell-depleted mice were more susceptible to M. tuberculosis infection but did not show decreased IFN-γ and iNOS expression.

Transfer of T_reg into Rag1^−/− mice did not contribute to protection against M. tuberculosis (Fig. 1,B). When cotransferred with Th, T_reg even suppressed protection mediated by effector CD4⁺ T cells (Fig. 1,B). To better understand the mechanisms of T_reg-mediated suppression of Th, we performed intracellular IL-10 staining of CD4⁺ T cells recovered from M. tuberculosis-infected wt and T cell-reconstituted Rag1^−/− mice as well as real-time RT-PCR analyses of IL-10 mRNA in tissue samples from these mice. Flow cytometric analysis revealed that the suppressive effect of T_reg on Th was not correlated by increasing numbers or frequencies of IL-10-producing CD4⁺ T cells, neither among the transferred T_reg nor among the Th (Fig. 2, A and D). Moreover, total IL-10 expression did not differ significantly between infected C57BL/6 wt and Th- or Th- plus T_reg-reconstituted Rag1^−/− mice (Fig. 3). Thus, our results reveal no correlation between IL-10 and probably also TGF-β (preliminary data) abundance and T_reg-mediated suppression of protection against tuberculosis. Both cytokines represent critical mediators of T_reg-controlled immune responses (12). In contrast to our results, reactivation of M. tuberculosis in mice was found to correlate with increased IL-10 expression after treatment with TNF-α-neutralizing Abs (18). However, reactivation of M. tuberculosis due to TNF-α neutralization probably encompasses different mechanisms of regulation as compared with our T cell transfer model. In humans, active tuberculosis correlates with increased IL-10 blood serum levels (19). Although, at first sight, this observation might seem to contradict our findings, it is essentially in line with our results, given the fact that murine tuberculosis represents a chronic disease, rather than latent infection, and that IL-10 expression was observed in all M. tuberculosis-infected mice harboring T cells. More precisely, we observed a correlation between IL-10 and the presence of T cells, although not specifically with T_reg. A possible explanation for this observation could be that other cells, like M. tuberculosis-infected dendritic cells and macrophages, express IL-10, particularly in the presence of effector T cells (20, 21).

Despite significant differences in protection and T cell responses, histological analyses revealed similar granuloma shape, size, and numbers between infected C57BL/6 wt and T cell-reconstituted Rag1^−/− mice (data not shown).

In conclusion, our experiments focused on protective immunity conferred by CD4⁺ Th. In this adoptive transfer system, CD4⁺ Th were capable of controlling M. tuberculosis with high efficiency only in the absence of T_reg. We do not exclude the contribution of other cells to protection such as CD8⁺ T cells, which have been found to participate in control of tuberculosis, particularly at later stages of disease (22). It remains to be established whether CD8⁺ T cells are also under the control of T_reg during natural infection. So far, it is unclear why T_reg are activated during tuberculosis if they prevent sterile eradication and thus increase the risk of disease reactivation. It is possible that T_reg are programmed by specific microbial stimuli, ensuring survival of tubercle bacilli as suggested for malaria parasites (23). Suppression by T_reg could also serve as a means of the immune system to hedge a constant antigenic load within the host, thus maintaining a constant pool of mycobacteria-specific T cells ready to defend the host against reinfection. Such a mechanism has been described for infection with Leishmania major (11, 24, 25, 26). It is likely that latent infection provides a constant stimulus for Th as well as T_reg. A balanced response of both T cell populations can curtail both active disease and extensive pathology. This immune response is sufficiently fine-tuned for the majority of M. tuberculosis-infected individuals and only a minority experiences disease outbreak. However, in the face of HIV, latent M. tuberculosis infection has become a more serious issue. Coinfection with HIV debilitates the immune response, which controls M. tuberculosis, thus increasing the risk of disease reactivation by several 100-fold (4). A better understanding of the mechanisms by which T_reg suppress protective immunity against tuberculosis could promote the design of novel intervention strategies aimed at generating highly efficacious T cell responses unimpaired by concomitant generation of T_reg.

The authors have no financial conflict of interest.