CD4+ T-cell activation by antigen-presenting cells infected with urease-deficient recombinant Mycobacterium bovis bacillus Calmette-Guérin

Authors


  • Editor: Patrick Brennan

Correspondence: Masahiko Makino, Department of Microbiology, Leprosy Research Center, National Institute of Infectious Diseases, 4-2-1 Aobacho, Higashimurayama, Tokyo 189-0002, Japan. Tel.: +81 42 391 8059; fax: +81 42 391 8212; e-mail: mmaki@nih.go.jp

Abstract

We constructed a recombinant Mycobacterium bovis bacillus Calmette-Guérin (BCG-ΔUT) that lacks urease, providing acidic intraphagosomal conditions to drive an effective human immune T-cell response. BCG-ΔUT-infected macrophages stimulated autologous CD4+ T cells more efficiently than parent BCG-infected macrophages. For further T-cell activation, BCG-ΔUT-infected macrophages required pretreatment with exogenous recombinant granulocyte-macrophage colony-stimulating factor or costimulation with either CD40 ligand or interferon-γ. By contrast, BCG-ΔUT-infected dendritic cells induced significant activation of naïve CD4+ T cells without costimulating signals. C57BL/6 mice intradermally inoculated with BCG-ΔUT more efficiently produced memory T cells that responded to recall antigen. Therefore, the depletion of urease from BCG is useful for the activation of T cells.

Introduction

Mycobacteria, such as Mycobacterium leprae and Mycobacterium tuberculosis, are representative parasitic intracellular pathogens. Mycobacterium leprae is a causative agent of human leprosy, in cases of which skin lesions and chronic progressive peripheral nerve injury are usually observed (Stoner, 1979; Job, 1989). At present, around one-third of individuals are infected with M. tuberculosis and several millions die as result of tuberculosis each year (Dye et al., 2005; World Health Organization, 2006). Mycobacterium bovis bacillus Calmette-Guérin (BCG) has been used as a vaccine against leprosy, although its efficacy is quite limited (Andersen & Doherty, 2005; Setia et al., 2006). The emergence of multidrug-resistant strains of these mycobacteria is of concern (Maeda et al., 2001; Kai et al., 2004; Kaufmann, 2005), and therefore the urgent development of a new vaccine, including a more efficacious recombinant BCG, is desired (Kaufmann, 2005).

Among various immunocompetent cells, CD4+ T cells, especially IFN-γ-producing cells, play an extremely important role in inhibiting the multiplication of mycobacteria, killing them in the early stages of infection, and keeping the bacterial load at a stable level (Orme et al., 1993; Dockrell et al., 1996; Hashimoto et al., 2002). CD4+ T cells that can respond quickly to pathogenic mycobacteria and produce IFN-γ are known as memory T cells. The efficient production of such memory T cells needs pre-exposure to antigenic vaccinating molecules, which share their antigenicity with that of pathogenic mycobacteria (Kaufmann, 2006). BCG has been considered a good candidate for a vaccine against M. leprae in this respect, however its efficacy is limited in several aspects, including the ability to activate T cells (Kaufmann & McMichael, 2005). BCG resides in the phagosomes of macrophages and thus attenuates the trafficking of antigenic molecules to the macrophage cell surface (Grode et al., 2005). One possible strategy for improving the ability of BCG to stimulate T cells is to enhance its ability to fuse with the lysosomes. To this end, we knocked out the urease gene from BCG. The urease-deficient recombinant BCG (BCG-ΔUT) is expected to allow phagosomal acidification in the host cells, and induce efficient phagosome maturation for cytolytic activity of the antigenic molecules of BCG (Schaible et al., 1998; Honerzu Bentrup & Russell, 2001).

In the present study, we evaluated the ability of BCG-ΔUT to activate IFN-γ-producing type 1 CD4+ T cells through antigen-presenting cells (APCs), and to produce memory CD4+ T cells. When used as a target of BCG-ΔUT, macrophages fully stimulated CD4+ T cells in the presence of costimulatory agents such as CD40 ligand (L) and IFN-γ. In addition, BCG-ΔUT-infected monocyte-derived dendritic cells (DCs) activated type 1 CD4+ T cells more efficiently than parent BCG-infected cells in the absence of these costimulators. Therefore, BCG-ΔUT was found to be a useful T-cell-stimulating agent.

Materials and methods

Preparation of blood cells

Peripheral blood was obtained from healthy purified protein derivative (PPD)-positive individuals with informed consent. PPD-negative individuals provide more information, however, as healthy individuals are PPD-positive, due to compulsory BCG vaccination for children in Japan (0–4 years old). Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll-Paque Plus (Pharmacia, Uppsala, Sweden) and cryopreserved in liquid nitrogen until use, as previously described (Makino & Baba, 1997). For the preparation of peripheral monocytes, CD3+ T cells were removed from either freshly isolated heparinized blood, or cryopreserved PBMCs using immunomagnetic beads coated with anti-CD3 monoclonal antibody (mAb) (Dynabeads 450, Dynal, Oslo, Norway). The CD3 PBMC fraction was plated on collagen-coated plates and nonadherent cells were removed by extensive washing. The remaining adherent cells were used as monocytes (Makino & Baba, 1997). Macrophages were generated by culturing monocytes in the presence of 20% fetal calf serum and recombinant (r) macrophage colony-stimulating factor (M-CSF) (R&D Systems, Abingdom, UK) (Makino et al., 2007). Macrophages were pulsed with rBCGs on day 5 of culture, and were used as a stimulator of T cells on day 7 (Makino et al., 2007). Monocyte-derived DCs were differentiated as described previously (Makino et al., 1999). Briefly, monocytes were cultured in the presence of 50 ng recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF; Pepro Tech EC Ltd, London, UK) and 10 ng of recombinant interleukin (rIL)-4 (Pepro Tech) per millilitre (Makino et al., 1999). On day 3 of culture, immature DCs were infected with rBCGs at the indicated multiplicity of infection (MOI), and on day 5 of culture, DCs were used for further analyses of surface antigens and for mixed-lymphocyte assays.

BCG culture and DNA manipulation

The mycobacterial strain, BCG substrain Tokyo, for DNA manipulation was grown in Middlebrook 7H9 broth (Difco Laboratories) with 0.05% Tween 80 or Middlebrook 7H10 agar (Difco) with 0.5% glycerol, each supplemented with 10% albumin–dextrose–catalase enrichment (Difco). DNA manipulations including isolation of DNA, transformation and PCR, were carried out as described previously (Miyamoto et al., 2004). Escherichia coli strain DH5α was used for routine manipulation and the propagation of plasmid DNA. Escherichia coli strain STBL4 was used for the construction of plasmid vectors derived from phAE87. Antibiotics were added as required: hygromycin B, 150 μg mL−1 for E. coli and 75 μg mL−1 for Mycobacterium smegmatis (mc2155) and M. bovis BCG. A recombinant BCG that lacks a urease gene was constructed. The sequence of the targeted gene, ureC (BCG 1886), was obtained from the BCG list (http://genolist.pasteur.fr/BCGList/). The ureC gene was inactivated by inserting a hygromycin-resistance cassette (hyg) using a specialized transducing phage system for homologous recombination (Bardarov et al., 2002; Miyamoto et al., 2006). To construct the disrupted sequence, fragments of around 0.9 kb both upstream and downstream of ureC were amplified from BCG-Tokyo genomic DNA using the following two pairs of primers: F UureC and R UureC for upstream of ureC, and F DureC and R DureC for downstream of ureC. The PCR products were digested with each restriction enzyme and cloned into the corresponding site flanking hyg of pYUB854 to give pYUB854-ureC-UD. This plasmid was used for packaging into the phasmid vector phAE87 to construct a specialized transducing mycobacteriophage for gene disruption as described previously (Bardarov et al., 2002; Miyamoto et al., 2006). BCG-Tokyo infected with the mycobacteriophage at an MOI of 50 was incubated at 37 °C for 3 h in 7H9 broth without Tween 80. Harvested bacterial cells were then plated and cultured on 7H10 agar containing hygromycin B (75 μg mL−1) for 3 weeks. The hygromycin B-resistant colonies were selected and evaluated with a conventional urease assay. A change in the color of the assay medium from yellowish to red was scored as urease-positive. Furthermore, genomic DNA obtained from these colonies was subjected to PCR to confirm the disruption of the gene using primers F ureC and R ureC (Fig. 1). The colony which tested negative in the urease assay was named BCG-ΔUT, while the parental BCG substrain Tokyo is referred to as BCG-Tokyo.

Figure 1.

 Disruption of the ureC gene. (a) Schematic diagram of the ureC region on the chromosome of the wild-type Mycobacterium bovis BCG Tokyo strain and its gene disruptant, ΔUT. The shaded boxes indicate the regions included in the recombinant phage for gene disruption. The black arrow represents the coding region of the ureC gene. The gray box represents the hygromycin-resistace cassette (hyg). The primers used for PCR analysis are indicated by small arrows. (b) PCR analysis of the wild-type and the disruptant using the primers indicated above.

Preparation of M. leprae

Mycobacterium leprae (Thai-53) was isolated from the footpads of BALB/c-nu/nu mice (McDermott-Lancaster et al., 1987). The isolated bacteria were counted by Shepard's method (Charles & Shepard, 1960). The MOI for infection to host cells was determined based on the assumption that macrophages and DCs were equally susceptible to infection with BCG or M. leprae (Hashimoto et al., 2002).

Preparation of mycobacterial antigen

The cytosolic fraction of BCG-Tokyo (BCC) was obtained as described previously (Maeda et al., 2003). Briefly, the mycobacterial suspension containing the protease inhibitors was mixed with zirconium beads at a ratio of c. 1 : 1 (v/v) and homogenized using a Beads Homogenizer Model BC-20 (Central Scientific Commerce, Tokyo). The suspension was centrifuged at 10 000 g to remove the cell-wall fractions. The supernatant was then ultracentrifuged at 100 000 g and the resulting supernatant was taken as the cytosolic fraction. For preparation of the M. leprae membrane (MLM) fraction, M. leprae was used instead of BCG and treated similarly. The pellet obtained by ultracentrifugation (100 000 g for 1 h) was used as a membrane fraction (MLM). The optimal concentration of BCC and MLM for stimulating T cells was determined in advance.

Analysis of cell surface antigens

The expression of cell surface antigens on macrophages and DCs, either untreated or treated with exogenous rIFN-γ (R&D Systems), was analyzed using a FACSCalibur flow cytometer. Dead cells were eliminated from the analysis by staining with propidium iodide (Sigma Chemical Co., St. Louis, MO), and 1 × 104 live cells were analyzed. For the analysis of cell surface antigens, the following mAbs were used: fluorescein isothiocyanate (FITC)-conjugated mAbs against HLA-ABC (G46-2.6), HLA-DR (L243), CD14 (M5E2), CD40 (5C3) and CD86 (FUN-1). These mAbs were obtained from BD PharMingen (San Diego, CA).

APC function of rBCG-infected macrophages and DCs

The ability of rBCG-infected macrophages to stimulate T cells was assessed using an autologous mixed-lymphocyte assay as previously described (Wakamatsu et al., 1999; Hashimoto et al., 2002). The responder CD4+ T cells were purified from freshly thawed PBMCs by using a CD4-negative isolation kit (Dynabeads 450; Dynal) (Wakamatsu et al., 1999). The purity of CD4+ T cells was more than 95% as assessed by fluorescence-activated cell sorting (FACS) analysis. Naïve CD4+ T cells were produced by further treatment of CD4+ T cells with an mAb to CD45RO antigen, followed by incubation with beads coated with goat antimouse IgG. Memory-type T cells were similarly produced by the treatment of cells with an mAb to CD45RA antigen. The purified responder cells (1 × 105 well−1) were plated in 96-well round-bottom tissue culture plates and macrophages or DCs were added to give the indicated APC/CD4+ T-cell ratio. Supernatants of the cocultures were collected on day 4 and the concentration of cytokines was determined. In some cases, macrophages were treated with the indicated dose of exogenous rGM-CSF (Pepro Tech) in advance of infection with rBCGs. Further, macrophages were infected with rBCGs in the presence of neutralizing mAb to IL-10 (JES3-9D7; Rat IgG, BD PharMingen) or control normal rat IgG. Macrophages infected with BCGs were further costimulated with either rCD40L (Pepro Tech) or rIFN-γ (R&D Systems), and in some cases, the macrophages were stimulated with rIFN-γ in the presence of anti-IFN-γ receptor α chain (CD119) (GIR-208, mouse IgG1, BD PharMingen) or control normal mouse IgG. In other cases, macrophages infected with BCG-ΔUT in the presence of exogenous rIFN-γ were treated with either mAb to HLA-DR (L243, mouse IgG2a), CD86 (IT2.2, mouse IgG2b, BD PharMingen) or control normal mouse IgG, and subsequently cocultured with responder CD4+ T cells. The concentration of IFN-γ produced by CD4+ T cells was quantified using an enzyme assay kit [OptEIA Human enzyme linked immunosorbent assay (ELISA) Set; BD Biosciences].

Production of IL-12p70 and IL-1β by DCs

The ability of DCs to produce IL-12p70 and IL-1β on stimulation with BCG-Tokyo or BCG-ΔUT was assessed. The DCs were stimulated with BCGs at the indicated MOI for 24 h, and the concentration of these cytokines was quantified using the Opt EIA Human ELISA Set.

Animal studies

For inoculation into mice, BCG-Tokyo and BCG-ΔUT were cultured in Middlebrook 7H9 to log phase and stored at 108 CFU mL−1 at −80 °C. Before aliquots were used for inoculation, the concentration of viable bacilli was determined by plating cells on the Middlebrook 7H10 agar plate. Three 5-week-old C57BL/6J mice per group were inoculated intradermally with 0.1 mL phosphate-buffered saline (PBS) containing 1 × 102 or 1 × 103 BCG-Tokyo or BCG-ΔUT. The animals were kept under specific pathogen-free conditions and were supplied with sterilized food and water. Four weeks after injection, the spleens were removed, and the splenocytes were suspended at a concentration of 2 × 106 cells mL−1 in culture medium, and stimulated with the indicated concentration of BCC or MLM in triplicate in 96-well round-bottomed microplates. The individual culture supernatants were collected 3 days after stimulation, and IFN-γ and IL-2 were measured using an OptEIA mouse ELISA set.

Statistical analysis

The Student's t-test was applied to determine statistical differences.

Results

Induction of the fusion of BCG-ΔUT-infected phagosomes with lysosomes

The efficacy with which BCG-ΔUT-infected phagosomes fused with lysosomes in macrophages was examined using confocal microscopy. Lysosomes were stained with anti-LAMP1 mAb after treatment of THP-1 cells with FITC-labeled BCG-Tokyo or BCG-ΔUT for 24 h. The parental BCG colocalized with lysosomes less efficiently than BCG-ΔUT (data not shown). Therefore, BCG-ΔUT may at least partially enhance the ability to induce phagosomal maturation.

T-cell-stimulating activity of BCG-ΔUT

The activity of BCG-ΔUT to stimulate IFN-γ-producing CD4+ T cells, when infected to macrophages, was assessed (Fig. 2). BCG-ΔUT-infected macrophages activated unseparated CD4+ T cells to release IFN-γ substantially more efficiently than parent BCG-infected macrophages. Although BCG-ΔUT-infected macrophages also induced production of IL-2 from CD4+ T cells (data not shown), the extent of IFN-γ (<50 pg mL−1) and IL-2 production was not as high as expected. Furthermore, BCG-ΔUT did not induce the activation of naïve CD4+ T cells (data not shown). As the activation of T cells is largely influenced by the cytokine milieu, in which T cells and their stimulators are present, we determined the level of cytokines produced from macrophages on stimulation with BCG-ΔUT. BCG-ΔUT produced significantly more cytokines, such as IL-10, GM-CSF, TNFα and IL-1β, than the parental BCG (data not shown). It has been reported that IL-10 inhibits the APC-mediated activation of T cells (Granelli-Piperno et al., 2004) and GM-CSF regulates the function of macrophages (Makino et al., 2007). To examine the role of IL-10 on T-cell activation, macrophages were infected with BCGs in the presence of a neutralizing mAb to IL-10 (Fig. 3a). The IFN-γ production by stimulated CD4+ T cells was not influenced by the treatment of macrophages with control IgG; however, a significantly higher level of IFN-γ was produced on treatment with the neutralizing mAb to IL-10. The up-regulation by IL-10 mAb treatment was observed in both BCG-Tokyo and BCG-ΔUT in a similar fashion. Furthermore, the pretreatment of macrophages with exogenous GM-CSF also significantly upregulated the antigen-presenting function of macrophages, although the effect of GM-CSF was more pronounced in BCG-ΔUT-infected macrophages (Fig. 3b).

Figure 2.

 Production of IFN-γ by CD4+ T cells. Macrophages, differentiated by 5 days of culture with rM-CSF from monocytes, were infected with BCG-Tokyo (parental BCG) or BCG-ΔUT at the indicated MOI, and cultured for another 2 days in the presence of rM-CSF. These macrophages were used as a stimulator of autologous CD4+ T cells (1 × 105 cells well−1) at the indicated T-cell/macrophage ratio in a 4-day culture. A representative example of three separate experiments is shown. Assays were performed in triplicate and the results are expressed as the means±SD. Titers were statistically compared using Student's t-test.

Figure 3.

 Effect of IL-10 and GM-CSF on IFN-γ production. (a) Macrophages differentiated from monocytes by using rM-CSF were infected with either BCG-Tokyo or BCG-ΔUT at an MOI of 0.25 on day 5 of culture and cultured for another 2 days in the presence of rM-CSF. The BCG-infected macrophages were treated with neutralizing mAb to IL-10 or isotype-matched control IgG (10 μg mL−1), and used as a stimulator of CD4+ T cells, at a T-cell/macrophage ratio of 10 : 1, and cultured for another 4 days. The optimal concentration of mAb was determined in advance. (b) Macrophages obtained by 4 days of culture with rM-CSF were treated with the indicated dose of rGM-CSF. The macrophages pretreated with rGM-CSF were infected with BCG-Tokyo or BCG-ΔUT at an MOI of 0.25, cultured for another 2 days in the presence of rM-CSF used as a stimulator of CD4+ T cells on day 8, at a T-cell/macrophage ratio of 10 : 1 (4 days of stimulation). A representative example of three separate experiments is shown. Assays were performed in triplicate and the results are expressed as the means±SD. Titers were statistically compared using Student's t-test.

Next, we phenotypically assessed the effect of BCG-ΔUT on macrophages (Fig. 4a). BCG-ΔUT induced enhanced expression of both CD14 and CD40 on macrophages compared with BCG-Tokyo. Based on these results, we treated BCG-infected macrophages with CD40L to examine its role as a costimulator of macrophages (Fig. 4b). The CD40L treatment upregulated the T-cell activation by BCG-infected macrophages, but it more efficiently affected BCG-ΔUT-infected macrophages. Similarly, there was a significant difference between parent BCG and BCG-ΔUT in sensitivity to IFN-γ (Fig. 5a). However, other cytokines such as TNFα and IL-1β did not enhance the T-cell-stimulating activity of rBCG-infected macrophages. The IFN-γ treatment was effective against both BCG-Tokyo- and BCG-ΔUT-infected macrophages; however, more than a 10-fold increase in the production of IFN-γ from T cells was achieved only when BCG-ΔUT-infected macrophages were stimulated with exogenous IFN-γ. The optimal stimulation of T cells induced the production of more than 500 pg mL−1 IFN-γ. The exogenous IFN-γ seems to contribute directly to the enhancement of APC function, as the IFN-γ-mediated enhancement was cancelled out by the pretreatment of BCG-ΔUT-infected macrophages with mAb to IFN-γ receptor α-chain (Fig. 5b). Furthermore, IFN-γ significantly enhanced the expression of HLA-DR and CD86 on BCG-ΔUT-infected macrophages (Fig. 5c), while the phenotypic alteration of BCG-Tokyo-infected macrophages by IFN-γ was minimum (data not shown). When BCG-ΔUT-infected, IFN-γ-treated macrophages were treated with mAb to either HLA-DR or CD86 in advance of being cocultured with CD4+ T cells, IFN-γ production by the T cells was significantly inhibited, while normal murine IgG treatment did not have any effect (Fig. 5d).

Figure 4.

 (a) Expression of CD14 and CD40 molecules on macrophages. Macrophages produced by using rM-CSF were infected with BCGs at an MOI of 0.25, and cultured for another 2 days in the presence of rM-CSF. The macrophages on day 7 of culture were gated and analyzed. Dotted lines, isotype-matched control IgG; solid lines, the indicated test mAb. The number in the top right-hand corner of each panel represents the difference in mean fluorescence intensity between the control IgG and the test mAb. Representative results of three separate experiments are shown. (b) IFN-γ production by CD4+ T cells stimulated with BCG-infected macrophages. Macrophages differentiated from monocytes using rM-CSF were infected with BCGs at an MOI of 0.25 on day 5 of culture, further treated with CD40L (1 μg mL−1) on day 6, and used as a stimulator of CD4+ T cells (T-cell/macrophage ratio of 10 : 1, 4 days of stimulation). A representative of three separate experiments is shown. Assays were performed in triplicate and the results are expressed as the means±SD. Titers were statistically compared using Student's t-test.

Figure 5.

 (a) Effect of exogenous IFN-γ on CD4+ T-cell activation. Macrophages produced by 5 days of culture with rM-CSF from monocytes were infected with BCGs at an MOI of 0.25 and simultaneously treated with the indicated dose of exogenous IFN-γ. The macrophages were used as a stimulator of CD4+ T cells (T-cell/macrophage ratio of 10 : 1, 4 days of stimulation). A representative of three separate experiments is shown. Assays were performed in triplicate and the results are expressed as the means±SD. Titers were statistically compared using Student's t-test. (b) Involvement of IFN-γ receptor in T-cell activation. Macrophages produced as in (a) were infected with BCG-ΔUT (MOI of 0.25), stimulated with exogenous IFN-γ (100 U mL−1) in the presence of mAb to IFN-γ receptor α-chain (CD119) or isotype matched control IgG (10 μg mL−1), and cultured for another 2 days in the presence of rM-CSF. The macrophages were used as a stimulator of CD4+ T cells (T-cell/macrophage ratio of 10 : 1, 4 days of stimulation). A representative of three separate experiments is shown. Assays were performed in triplicate and the results are expressed as the means±SD. Titers were statistically compared using Student's t-test. (c) Surface expression of various molecules on BCG-ΔUT-infected, IFN-γ-treated macrophages. Macrophages produced as in (a) were infected with BCG-ΔUT (MOI of 0.25), stimulated with exogenous IFN-γ (100 U mL−1) and cultured for another 2 days in the presence of rM-CSF. The macrophages on day 7 of culture were gated and analyzed. Dotted lines, isotype-matched control IgG; solid lines, the indicated test mAb. The number in the top right-hand corner of each panel represents the difference in mean fluorescence intensity between the control IgG and the test mAb. Representative results of three separate experiments are shown. (d) Involvement of surface antigens of BCG-ΔUT-infected, IFN-γ-stimulated macrophages in T-cell activation. Macrophages produced as in (a) were infected with BCG-ΔUT (MOI of 0.25), treated with exogenous IFN-γ (100 U mL−1) and cultured for another 2 days in the presence of rM-CSF. These macrophages were cocultured with autologous CD4+ T cells at a T-cell/macrophage ratio of 10 : 1 in a 4-day culture in the presence of the indicated mAb (10 μg mL−1). A representative of three separate experiments is shown. Assays were performed in triplicate and the results are expressed as the means±SD. Titers were statistically compared using Student's t-test.

CD4+ T-cell activation by BCG-ΔUT-infected DCs

As BCG-ΔUT significantly but less efficiently activated CD4+ T cells through macrophages in the absence of costimulation, the potency of BCG-ΔUT-infected DCs as a T-cell activator was evaluated. Expression of surface molecules on DCs infected with either BCG-Tokyo or BCG-ΔUT was examined (Fig. 6a). Expression of HLA-ABC, HLA-DR, CD86 and CD83 was more significantly upregulated by the infection with BCG-ΔUT than with BCG-Tokyo. Higher levels of IL-12p70 and IL-1β were produced by BCG-ΔUT stimulation (Fig. 6b). Furthermore, we assessed whether BCG-ΔUT activated naïve and memory CD4+ T cells through DCs by using various MOI titers and multiple T/DC ratios (Fig. 6c). IFN-γ levels were significantly higher following stimulation with BCG-ΔUT than with parent BCG in both naïve and memory CD4+ T cells. Also, a higher level of CD40L was expressed on CD4+ T cells after stimulation with BCG-ΔUT-infected DCs (data not shown). These results indicate that the infection of DCs with BCG-ΔUT alone was sufficient, as compared with macrophages which required costimulators to drive a strong T-cell response.

Figure 6.

 (a) Expression of various molecules on BCG-infected DCs. Monocyte-derived immature DCs were infected with either BCG-Tokyo or BCG-ΔUT at an MOI of 0.25 and cultured for another 2 days in the presence of rGM-CSF and rIL-4. The DCs from day 5 were gated and analyzed. Dotted lines, isotype-matched control IgG; solid lines, the indicated test mAb. The number in the top right-hand corner of each panel represents the difference in mean fluorescence intensity between the control IgG and the test mAb. Representative results of three separate experiments are shown. (b) Cytokine production from DCs stimulated by BCG. Monocyte-derived DCs from 4 days of culture in the presence of rGM-CSF and rIL-4 were stimulated with the indicated dose of either BCG-Tokyo or BCG-ΔUT for 24 h. The concentration of the indicated cytokine was determined by the ELISA method. A representative of three separate experiments is shown. Assays were performed in triplicate and the results are expressed as the means±SD. Titers were statistically compared using Student's t-test. (c) IFN-γ production by naïve CD4+ T cells and memory CD4+ T cells. DCs obtained from monocytes infected with either BCG-Tokyo or BCG-ΔUT were used as a stimulator of naïve and memory CD4+ T cells in a 4-day culture. A representative of three separate experiments is shown. Assays were performed in triplicate and the results are expressed as the means±SD. Titers were statistically compared using Student's t-test.

Memory T-cell production by BCG-ΔUT

Another important aspect of using BCG as a vaccine is the production of memory T cells in vivo. We examined the response of splenic T cells obtained from BCG-infected C57BL/6 mice to mycobacterial recall antigen (Fig. 7). We used BCC as a recall antigen. At 4 weeks following infection, splenic T cells from BCG-ΔUT-infected mice produced more IFN-γ than those from mice infected with BCG-Tokyo by responding to BCC. The lymphocyte population producing IFN-γ was found to be CD4+ T cells by intracellular staining (data not shown). Furthermore, upon stimulation with MLM, which contains immunodominant antigens of M. leprae, CD4+ T cells from BCG-ΔUT-infected mice produced significantly higher levels of IFN-γ than those from uninfected or BCG-Tokyo-infected mice (Fig. 7).

Figure 7.

 IFN-γ production by splenic T cells obtained from C57BL/6 mice infected with BCG-Tokyo or BCG-ΔUT. Five-week-old C57BL/6 mice were infected with the indicated dose of BCG intradermally. Four weeks after the inoculation, splenocytes (2 × 105 cells well−1) were stimulated with the indicated dose of either BCG-derived cytosol protein or Mycobacterium leprae-derived membrane protein for 4 days. Assays were performed in triplicate for each mouse, and the results for three mice per group are given, expressed as the means±SD. Representative results for two separate experiments are shown. Titers were statistically compared using Student's t-test.

Discussion

To date, BCG is the only suitable vaccine against leprosy; however, its efficacy is quite limited. Overall efficacy in one meta-analysis was reported to be only 26% (Setia et al., 2006). Several reasons might explain why BCG cannot block multiplication of M. leprae or inhibit the development of leprosy. The most important defect of BCG is that it is retained in phagosomes of macrophages, avoiding phagosomal acidification and hence interfering in the efficient fusion of BCG-containing phagosomes with lysosomes (Clements et al., 1995; Reyrat et al., 1995; Grode et al., 2005). The lack of phagosome–lysosome fusion inhibits the trafficking of BCG-derived antigens through the major histocompatibility class (MHC) II pathway, which is enrolled for preferential stimulation of CD4+ T cells, the most important cells involved in inhibition of M. leprae growth (Sendide et al., 2004). Further, macrophages produce abundant amounts of IL-10 on infection with BCG, which, in turn, inhibits the activation of CD4+ T cells (Mochida-Nishimura et al., 2001; Granelli-Piperno et al., 2004).

In the present study, we constructed a recombinant BCG (BCG-ΔUT) that lacks a urease gene through allelic exchange of chromosomal DNA. As urease is involved in the maintenance of intraphagosomal pH at neutral (Grode et al., 2005) or slightly alkaline values (Sendide et al., 2004), lack of this enzyme may contribute to the induction of phagosomal acidification (Sendide et al., 2004), thereby promoting the fusion of BCG-containing phagosomes with lysosomes. The efficient colocalization of BCG-ΔUT with lysosome was observed, leading us to expect an efficient enhancement of T-cell activation by BCG-ΔUT-infected macrophages. Previously, rBCG deficient in urease C was produced by a similar system and found to be superior to parental BCG in producing acidic conditions (pH 4.5–5.5) in BCG-infected phagosomes in murine macrophages (Reyrat et al., 1995; Grode et al., 2005). However, it was not demonstrated whether the rBCG deficient in urease C promoted the MHC class II trafficking pathway and actually activated human CD4+ T cells through APCs. The newly constructed BCG-ΔUT lacked urease activity and in vitro studies confirmed that it could not degrade urea to ammonia. When BCG-ΔUT was infected to macrophages, it activated human CD4+ T cells more efficiently than the parental BCG. However, the amount of IFN-γ released from the T cells was not as high as expected (<50 pg mL−1). These results suggest that improvement of intraphagosomal pH milieu for efficient phagosome–lysosome fusion was not sufficient for the induction of full T-cell activation as far as macrophages were concerned. Thus, we further searched for factors which might be helpful in inducing full activation of T cells. First, we examined the influence of endogenously produced IL-10, as abundant IL-10 was produced from macrophages by infection with BCG-ΔUT (data not shown). The neutralization of IL-10 from macrophages drastically enhanced T-cell activation (Fig. 3a). Furthermore, pretreatment of macrophages with GM-CSF, which is normally produced from activated CD4+ T cells, monocytes and macrophages (data not shown), and inhibits IL-10 production (Makino et al., 2007), was also quite efficient in enhancing the BCG-ΔUT-mediated T-cell activity. Therefore, the unexpectedly weak activation of CD4+ T cells by BCG-ΔUT seemed to be at least partly due to the immunosuppressive effect of IL-10. Secondly, we focused on the costimulating factors capable of actively up-regulating the T-cell-stimulating function of macrophages, and found that both CD40L and IFN-γ were quite efficient. It was previously reported that both CD40L and IFN-γ were needed to costimulate macrophages infected with M. leprae (Makino et al., 2007); however, in the present study, the sole treatment of BCG-ΔUT-infected macrophages with either CD40L or IFN-γ was enough to confer a sufficient effect (Figs 4 and 5). The high sensitivity of BCG-ΔUT-infected macrophages to CD40L may be due to the ability of rBCG to induce greater expression of CD40 (Fig. 4a). The exogenous IFN-γ may contribute to increased production of IFN-γ from T cells by activating macrophages, as it enhanced the surface expression of HLA-DR and CD86 on BCG-ΔUT-infected macrophages, which facilitated antigen-specific T-cell activation. As reported, M. leprae is less sensitive to IFN-γ (Makino et al., 2007), and also parental BCG was found to be clearly less sensitive to IFN-γ than BCG-ΔUT. These results indicate that each mycobacterium may have differential sensitivity to IFN-γ (Verreck et al., 2004). Although the molecular mechanism responsible for the difference in sensitivity remains unexplained, it is well known that IFN-γ facilitates the digestion of intracellular mycobacteria in macrophages, and thus the following speculation may be possible: in the present system, the alteration of the pH milieu of BCG-containing phagosomes caused by the depletion of urease activity may help to establish circumstances where cell activation as well as enhanced trafficking of mycobacterial antigens to the surface by the MHC class II pathway can be induced by IFN-γ treatment. The urease gene of pathogenic mycobacteria may be a good target for combination immunotherapy/chemotherapy as urease depletion downregulates the growth of mycobacteria (data not shown) and upregulates the immunoactivity of intracellular digestion of bacteria in host cells.

In contrast to macrophages, DCs were highly activated by the sole infection with BCG-ΔUT in terms of phenotype and cytokine production, and BCG-ΔUT-infected DCs efficiently activated both naïve and memory CD4+ T cells in the absence of additional costimulation. The activated T cells produced abundant amounts of both IFN-γ (Fig. 5c) and GM-CSF, and induced CD40L expression (data not shown). Therefore, DCs can inherently provide the critical factors needed by BCG-ΔUT-infected macrophages. As BCG infects both macrophages and DCs in vivo, we evaluated the efficacy of BCG-ΔUT as a T-cell activator by using C57BL/6 mice. BCG-ΔUT was superior to BCG-Tokyo in the production of murine memory CD4+ T cells, which can respond to BCG-derived recall antigen and also proteins derived from pathogenic M. leprae. Just 100 BCG-ΔUT bacilli were sufficient to produce such memory T cells. These findings indicate that BCG-ΔUT convincingly stimulated CD4+ T cells in vivo. As the C57BL/6 strain is a T helper (Th)1 response-prone mouse, further study using Th2 response-prone mice would provide further insight into how memory T cells are generated by inoculation with BCG-ΔUT.

Taking our data together, BCG-ΔUT is more potent than the parental BCG in the activation of macrophages, DCs and CD4+ T cells. The depletion of urease from BCG may be useful in upregulating the potency of BCG as an immunostimulator.

Acknowledgements

We acknowledge the contribution of Ms N. Makino in the preparation of the manuscript. We thank Ms Y. Harada for her technical support and the Japanese Red Cross Society for kindly providing blood from healthy donors. We also thank Professor William R. Jacobs, Jr, Howard Hughes Medical Institute, for providing the plasmids, pYUB854 and phAE87. This work was supported in part by a Grant-in-Aid for Research on Emerging and Re-emerging Infectious Diseases from the Ministry of Health, Labour, and Welfare of Japan.

Ancillary