OUP user menu

Recombinant BCG coexpressing Ag85B, ESAT-6 and mouse-IFN-γ confers effective protection against Mycobacterium tuberculosis in C57BL/6 mice

Ying Xu, Bingdong Zhu, Qingzhong Wang, Jiazhen Chen, Yaqing Qie, Jiuling Wang, Hongyan Wang, Baolin Wang, Honghai Wang
DOI: http://dx.doi.org/10.1111/j.1574-695X.2007.00322.x 480-487 First published online: 1 December 2007


In this study, the protective efficacy of a novel recombinant bacille Calmette Géurin (BCG) strain (rBCG-AEI) expressing fusion protein the antigen 85B (Ag85B)- the 6-kDa early secreted antigen target (ESAT-6)-IFN-γ against Mycobacterium tuberculosis H37Rv in mice was evaluated. The immunogenicity study showed that rBCG-AEI could induce higher specific antibody titers and significantly increase cellular immune response than BCG, or rBCG-A strain (expressing Ag85B), or rBCG-AE strain (expressing fusion protein Ag85B-ESAT-6). The protective experiment demonstrated that rBCG-AEI could confer similar or even better protective efficacy against M. tuberculosis infection compared with others in organ bacterial loads, lung histopathology and net weight gain or loss. The results suggested that rBCG-AEI is a potential candidate for further study.

  • Mycobacterium tuberculosis
  • rBCG
  • IFN-γ


Tuberculosis (TB) remains a leading cause of mortality and morbidity worldwide, with c. eight million new cases and two million deaths annually (Dye et al., 1999). Mycobacterium bovis bacille Calmette Géurin (BCG) vaccine, an attenuated strain of M. bovis, is the only vaccine currently available against Mycobacterium tuberculosis and the most widely used vaccine, with over three billion administered doses (Bloom & Fine, 1994; Andersen & Doherty, 2005). BCG protects children efficiently against miliary and meningeal TB, but the protective efficiency against adult pulmonary TB ranges from 0% to 80% based on large, well-controlled field trials (Fine, 1995). Therefore, it is urgently needed to develop more effective and satisfactory BCG vaccines against TB.

During the past decade, progress has been made in searching for the explanations of BCG failure and modified BCG strains as new anti-TB vaccine candidates. So far, a wide range of recombinant BCG (rBCG) vaccine candidates containing foreign viral, bacterial, parasite or immunomodulatory genetic material have been primarily developed and evaluated in animal models, for immune response to the foreign antigen (Dennehy & Williamson, 2005). However, the most important strategies are as follows: first, to stimulate more potent immune responses against M. tuberculosis. rBCG strains have been constructed that express cytokines such as IFN-γ, IFN-α or IL-2, IL-12 and granulocyte-macrophage colony-stimulating factor (GM-CSF) (Murray et al., 1996; Wangoo et al., 2000; Luo et al., 2001). Second, on the basis of the hypothesis that BCG has been weakened by continuing attenuation and gene loss, adding deleted genes back to BCG or increasing the expression of immunodominant genes might improve the effects of BCG vaccination, such as the rBCG strains overexpressing the lacked 6-kDa early secreted antigen target (ESAT-6) protein or the protective antigen the antigen 85B (Ag85B) (Horwitz et al., 2000; Pym et al., 2003). Moreover, the studies of Horwitz and Pym all show better protective efficacy than BCG in experimental models of TB. In this study, a rBCG strain secreting Ag85B-ESAT-6-IFN-γ (mouse) fusion protein was developed, and the protective efficacy of this novel vaccine against M. tuberculosis H37Rv in mice was examined.

Materials and methods

Bacterial strains and cultures

Mycobacterium bovis BCG (obtained from Shanghai Biological Products Institute), rBCG and M. tuberculosis strain H37Rv were grown on Middlebrook 7H9 medium (Difco Laboratories, Detroit, MI) supplemented with 0.5% glycerol, 0.05% Tween 80 and 10% ADC or on solid Middlebrook 7H11 medium (Difco Laboratories) supplemented with OADC. When required, the antibiotic kanamycin was added at a concentration of 25 µg mL−1. Escherichia coli DH5α was grown in Luria–Bertani medium and used for cloning.

Construction of rBCG

Coding sequences for ag85b (containing signal sequence) esat6 and mouse ifn-γ were amplified from the M. tuberculosis H37Rv genomic DNA and mouse ifn-γ cDNA (a kind gift from Dr Zhongming Li, Haigui Company, Shanghai, China), respectively, by PCR using the primers shown in Table 1. The ag85b-, esat6- and ifn-γ-coding regions were cloned into the mycobacterial–E. coli shuttle vector pMV261(Fig. 1), in which gene expression is under the control of the strong M. bovis HSP60 promoter. Inserted genes were sequencing confirmed. The three recombinant plasmids P261-A, P261-AE and P261-AEI were transformed into BCG by electroporation as described previously (Stover et al., 1993). The transformed BCG cells were plated on 7H11 medium supplemented with 25 µg mL−1 kanamycin and grown at 37 °C for 3 weeks; individual colonies were picked and grown in Sauton medium (0.25 g MgSO4·7H2O, 0.25 g K2HPO4, 1 g citric acid, 4 g sodium glutamate, 30 mL glycerol, 5 mg ZnSO4 and 25 mg ferrum-ammonium citrate in 500 mL) containing 25 µg mL−1 kanamycin. After 2 weeks' growth, protein expression was induced by heating at 45 °C for 60 min (Bao et al., 2003). The bacterial cells were centrifuged at 8000 g for 20 min. The culture supernatants were concentrated as described previously (Bao et al., 2003). Twelve micrograms of the concentrated culture supernatants were analyzed for expression of recombinant antigens by immunoblotting using anti-Ag85B and anti-ESAT-6 rabbit polyclonal anti-sera and the monoclonal anti-mouse IFN-γ antibody (R&D Systems, Minneapolis, MN).

View this table:
Table 1

Primers used

Figure 1

Mycobacterial–Escherichia coli shuttle vector pMV261 containing ag85b, ag85b-esat6 and ag85b-esat6-ifn-γ gene. For construction of the P261-AEI fusion molecule, the coding region (with the secretory signal sequence) of Ag85B was amplified from Mycobacterium tuberculosis H37Rv chromosomal DNA using primers A1 and A3. The PCR product was digested by BalI and EcoRI and cloned into pMV261. The PCR-amplified product of esat-6 using primers E1 and E3 was digested with EcoRI and SalI, followed by subcloning in-frame with the predigested pMV261-ag85b plasmid. Coding sequences for ifn-γ were amplified from mouse ifn-γ cDNA using primers I1 and I2, and cloned into the predigested pMV261-ag85b-esat6 plasmid with SalI and HpaI. The P261-A and P261-AE recombinant plasmids were constructed in a similar way, except that the PCR products were obtained by different primers.

The recombinant vaccines were grown in parallel in 100 mL Middlebrook 7H9 medium supplemented with 0.5% glycerol, 0.05% Tween 80, 10% ADC and 25 µg mL−1 kanamycin. Bacteria were collected by centrifugation at 8000 g for 20 min and washed once with 50 mM phosphate-buffered saline (PBS) (pH 7.0) before resuspension in 2 mL of the above PBS supplemented with 25% glycerol. The bacterial suspensions were divided into aliquots and frozen at −80 °C for later use. A single aliquot was defrosted for quantification of each vaccine lot.

Bioassay for IFN-γ activity of rBCG-AEI

The mouse IFN-γ Quantikine ELISA Kit (R&D Systems, Minneapolis, MN) was used to determine the IFN-γ activity of rBCG-AEI following the manufacturer's instructions. Finally, according to the standard curve, the bioactivity of IFN-γ produced by the rBCG-AEI strain was calculated. The experiment was performed twice, and the data were analyzed using a Student's t-test.

Mice and immunization

Female, 6–8-week-old C57BL/6 mice used in this study were obtained from the Animal Center of Second Military Medical University. The mice were maintained under specific pathogen-free conditions. The mice (nine per group) were immunized s.c. at the base of the tail with 5 × 106 CFU of BCG or rBCG in 100 µL PBS. They were sacrificed after 4, 8 and 12 weeks to prepare sera and splenocytes. The experiment was repeated twice.

Enzyme-linked immunosorbent assay (ELISA) analysis

ELISA plates (Maxisorb, type 96F; Nunc, Roskilde, Denmark) were coated overnight at 4 °C with Ag85B (0.5 µg well−1), ESAT-6 (1 µg well−1). The plates were blocked with 200 µL well−1 PBS containing 1% bovine serum albumin for 30 min at 37 °C and washed with PBS containing 0.05% Tween 20 three times. Sera were added at serial twofold dilutions (beginning at a 1/500 dilution) for 2 h at 37 °C and washed, followed by addition of 150 µL well−1 horseradish peroxidase-conjugated rabbit anti-mouse IgG (Dingguo Biotechnology, Beijing, China), IgG1 and IgG2b diluted at 1/10 000, 1/1000 and 1/1000 in PBS, respectively. Plates were incubated for 1 h at 37 °C, washed and developed with 0.1 M citrate–phosphate buffer, pH 5.0, containing 1 mg mL−1o-phenylenediamine and 0.03% hydrogen peroxide. Antibody titers are expressed as reciprocal end point titers. Reactions were stopped by addition of 50 µL well−1 of 1 N H2SO4 and were read on an ELISA plate reader at 492 nm.

Enzyme-linked immunospot (ELISPOT) assay

Mice were euthanized and their spleens were removed aseptically in RPMI-1640 medium containing 10% fetal calf serum, 2 mM glutamine, 50 µM β-mercaptoethanol, 100 µg mL−1 streptomycin and 100 U mL−1 penicillin. Lymphocytes from spleen cells obtained and diluted in culture medium containing an appropriate stimulus (10 µg mL−1 PPD, 2 µg mL−1 Ag85B, 6 µg mL−1 ESAT-6) were brought to the wells of the ELISPOT plate at 5 × 105 cells well−1. The mouse IFN-γ ELISPOT kit (U-Cytech biosciences, CT317-PR5, Utrecht, the Netherlands) was used to determine the relative number of IFN-γ-expressing cells in the single-cell spleen suspensions following the manufacturer's instructions. Finally, the spots were counted microscopically. Wells with fewer than 10 spots were not used for calculations.

Protective efficacy of rBCG in C57BL/6 mice

All of the animals were kept under controlled conditions in the P3 High Security Laboratory of the Animal Facility at Wuhan University. Groups of C57BL/6 mice (nine per group) were vaccinated subcutaneously at the base of the tail with 5 × 106 CFUs of either BCG or rBCG, and the control mice were immunized with saline. At 8 weeks postvaccination, all mice were challenged via the lateral tail vein (intravenous) with 1.8 × 106 CFU of M. tuberculosis H37Rv. The challenged experiment was performed once. Mice were sacrificed at 3, 6 or 9 weeks after the challenge. Organs were separated two parts for bacterial viable count or histopathology, respectively. CFUs counts were performed on serial dilutions of the homogenate, plated onto Middlebrook 7H11+OADC agar and examined after 3 weeks of growth at 37 °C. Tissues were collected into 10% neutral-buffered formalin and processed routinely for histopathology by paraffin embedding and sectioning into 4–6 µm sections stained with hematoxylin and eosin. Sections were evaluated by a board-certified pathologist who had no prior knowledge of the experimental groups and were evaluated at least twice to verify the reproducibility of the observations. All animal work was approved by institutional animal experimentation committees.

Data analysis

The difference comparison was made using a Student's F-test within the Microsoft excel data analysis program. The difference was considered to be statistically significant when P was <0.05.


Analysis of the rBCG Strains

The coding regions of ag85b, ag85b-esat6 and ag85b-esat-6-ifn-γ were cloned into the shuttle vector pMV261, respectively (Fig. 1), and the three rBCG strains (rBCG-A, rBCG-AE and rBCG-AEI) were obtained by transformation of BCG with the recombinant plasmids. Immunoblot was used to analyze the overexpression of proteins after rBCG strains induced by heat. Because of the signal sequence of the Ag85B, the expression of target proteins could be detected in the culture supernatants with three different antibodies (Fig. 2). There were higher amounts of proteins in the cell lysates (Fig. 2a). rBCG expressed and secreted a relatively higher level of Ag85B than BCG (Fig. 2a).

Figure 2

Western blot analysis. Supernatants were analysed for expression of recombinant antigens by immunoblotting using antibodies recognizing Ag85B (a), ESAT-6 (b) and IFN-γ (c). M, Protein marker; 1, Ag85B protein (including His-tag); 2, BCG; 3, rBCG-A; 4, rBCG-AE; 5, rBCG-AEI; 6, Cell lysates of rBCG-AEI.

Sandwich ELISA was used to measure the IFN-γ activity of rBCG-AEI. The bioactivity of IFN-γ produced by the rBCG-AEI strain was 30.04±10.05 pg mL−1 (about 3 × 106 U mg−1).

Humoral responses

Groups of mice were immunized with the three different rBCG strains (rBCG-A, rBCG-AE, rBCG-AEI) and the control group was immunized with BCG. Figure 3 illustrates the level of antibody response in the sera of mice from different groups, using the recombinant purified Ag85B or ESAT-6 protein as the antigen.

Figure 3

Antibody response against Ag85B and ESAT-6 in mice immunized with BCG and rBCG. C57BL/6 mice (nine per group) were immunized with BCG, rBCG-A, rBCG-AE and rBCG-AEI, and sacrificed after 4, 8 and 12 weeks to prepare serum for examining the Ab response of IgG (a, b) and the ratio of IgG2b/IgG1 (c, d). Results are expressed as the mean (±SE) *, the end point titer was significantly higher than those of the group inoculated with the BCG strain (P<0.05). (c) In contrast to BCG, the ratios of mice immunized with rBCG-A, rBCG-AE and rBCG-AEI were significant. (d) In contrast to BCG, the ratios of mice immunized with rBCG-AE and rBCG-AEI were significant.

Compared with the BCG group, mice vaccinated with the three different rBCG strains could induce higher levels of antibody against Ag85B, although there was not much difference in the anti-85B titers generated among the rBCG groups (Fig. 3a). Figure 3b illustrates that the anti-ESAT-6 titers of the two rBCG strains (rBCG-AE and rBCG-AEI) were significantly higher than that of BCG and rBCG-A (P<0.05).

In C57BL/6 mice, the gene coding for IgG2a is deleted (Martin & Lew, 1998). Therefore, in the absence of a functional IgG2a gene, the IgG2b isotype was used as an indicator of a T-helper type 1(Th1) response. Figure 3c and d illustrates the antibody levels of IgG1 and IgG2b isotype against the protein Ag85B or ESAT-6. The ratios of IgG2b/IgG1 were calculated to determine the induction of Th1/Th2 responses in animals. As a result, in contrast to the other rBCG and BCG, the ratios of mice immunized with rBCG-AEI were higher whatever against the protein Ag85B or ESAT-6 at 12 weeks when the ratios peaked (P<0.05). Overall, the results revealed that the capability of the induction of Th1/Th2 responses increased in the following sequence: BCG, rBCG-A, rBCG-AE, rBCG-AEI.

Cytokine response

CD4+ T-cell responses were evaluated by sensitive IFN-γ ELISPOT assays. Splenocytes isolated from mice immunized with BCG and rBCG proliferated following stimulation with purified Ag85B, ESAT-6 protein or PPD, respectively. The amount of IFN-γ secreted is shown in Fig. 4(a–c). It was noted that in response to purified Ag 85B and ESAT-6, IFN-γ secretion was maximum at the time point of 8 weeks in all groups, and there was a greater, but not significantly different, IFN-γ response after immunization with rBCG-AEI. However, when stimulated with PPD, the amount of IFN-γ secreted showed no significant difference at 8 weeks. The results also demonstrated that mouse splenocytes immunized with rBCG-AE and rBCG-A released more of IFN-γ compared with rBCG-AEI and BCG at 4 weeks in response to PPD. Especially in the group of mice immunized with rBCG-AEI, a correspondingly low secretion of IFN-γ was found at 4 weeks but was highly increased at 8 weeks. Moreover, the amount of IFN-γ secreted exhibited a decline at 12 weeks in all the groups.

Figure 4

Frequency of IFN-γ-secreting cells in the spleens of immunized mice. Splenic lymphocytes were isolated from three individual mice per group at 4, 8 and 12 weeks postimmunization and restimulated with purified Ag 85B (a), ESAT-6 (b) protein or PPD (c), respectively. Scale bars represent mean (±SE) spot-forming units (SFU) *, the end point SFU was significantly higher than those of the group inoculated with the BCG strain (P<0.05).

Protective efficacy of rBCG in C57BL/6 mice

Groups of C57BL/6 mice were inoculated subcutaneously with BCG or rBCG and challenged with M. tuberculosis H37Rv intravenously after 8 weeks. Three mice per group were sacrificed at 3, 6 and 9 weeks to determine the bacterial enumeration (Fig. 5a and b) and histopathology (data not shown). Before being sacrificed, each individual mouse was weighed (Fig. 5c). The results showed that the BCG or rBCG vaccine inhibited the growth of M. tuberculosis H37Rv in the spleen and was of a comparable efficacy at 6 and 9 weeks, although the rBCG-AEI vaccine could inhibit better than others at 3 weeks. However, in the lung immunization with rBCG the bacterial load was significantly reduced compared with mice vaccinated with BCG at 6, 9 weeks, and the numbers of CFU recovered from the mice vaccinated with rBCG-A and rBCG-AEI were reduced 0.1 log10 compared with that fromrBCG-AE mice (P<0.05).

Figure 5

Protective efficacy of BCG- and rBCG-vaccinated mice against Mycobacterium tuberculosis H37 Rv. C57BL/6 mice (three/time point) were inoculated subcutaneously with BCG or rBCG and challenged intravenously 8 weeks later with M. tuberculosis H37Rv. At 3, 6 and 9 weeks postinfection, the bacterial load was assessed in the lung (a) and spleen (b). Individual mice were weighed to calculate the mean weight gain or loss (c). Data are expressed as the mean (±SE) of mice.

Light photomicrographs of lung lesions from sacrificed mice at 9 weeks illustrate the following results. In mice given BCG or rBCG, the majority of the pulmonary parenchyma was replaced by a greater tubercle (about 15–20 alveolar) or middle tubercle (about 10 alveolar), and there was little alveolar macrophage proliferation. This was comparable with saline controls in which there was significant lamellar coalesced focus and alveolar macrophage proliferation. It was also noted that there was less tubercle in the rBCG-AEI vaccinated mice compared with BCG- or other rBCG-vaccinated animals. In addition, the mice vaccinated with rBCG-AEI gained weight and other groups lost weight to different degrees (P<0.05) (Fig. 5c). These data indicated that rBCG-AEI was superior to BCG and the other two rBCG in reducing the severity of the disease in the lung.


rBCG has several advantages over other vaccines because it is inexpensive, readily produced and can be stored conveniently. Therefore, it cannot easily be replaced by other vaccine candidates despite the controversies surrounding its use. Improvement of BCG remains among the best choices for the rational design of a vaccine in the case of TB. It can be envisaged that continuous overexpression of selected candidate epitopes as the case for a live vaccine like BCG may be pivotal for vaccine efficacy (Eddine & Kaufmann, 2005). Although increasing defined antigens have been identified, the majority of vaccines rely on only a small number of immunodominant antigens from M. tuberculosis, and the most popular and important antigens are Ag85B and ESAT-6 (Brandt et al., 2000; Palendira et al., 2005). In addition, it has been proved that vaccines containing multiple epitopes are more effective than a single antigen (Olsen et al., 2001; Langermans Jan et al., 2005). Moreover, cytokines are considered to stimulate more potent immune responses, and the potent IFN-γ producers are crucial for protection against M. tuberculosis (Luo et al., 2001). Therefore, in this work, the rBCG-AEI vaccine has been constructed and compared with BCG, rBCG-A and rBCG-AE in protective efficacy against M. tuberculosis in C57BL/6 mice. The results demonstrated that the novel rBCG strain could confer better protective efficacy than BCG, rBCG-A and rBCG-AE. It also suggests that in order to investigate the protective efficacy difference between rBCG and parental conventional BCG, it is needed to prolong the observed time after challenge.

The current strategies for vaccination against TB are attempts to control the spread of M. tuberculosis by preventing disease reactivation, but this depends on a competent Mycobacteria-specific T-cell being sustained (Kaufmann, 2006). A hypothetical view of the peripheral immune response from T-cell activation to memory and suppression suggests that the reason for the reactivation is that the T-cell response is shifted from the Th1 towards the Th2 pole, resulting in Th2 cells producing IL-4 and IL-10, and Th2 responses generally dominate over Th1 responses (Kaufmann, 2006). In abstracto, rBCG-AEI expressing IFN-γ could increase the Th1 responses, and weaken the Th2 responses when memory is suppressed by regulatory mechanisms. This hypothesis was confirmed to be true in this study, as the positive result of rBCG-AEI could more effectively protect mice against M. tuberculosis than BCG, rBCG-A and rBCG-AE.

The study of Palendira (2005) showed that a rBCG strain secreting Ag85B-ESAT-6 fusion protein displayed a satisfactory safety profile and improved the protective efficacy of the existing vaccine against M. tuberculosis challenge within the lung. In this study, mouse-IFN-γ was coexpressed on this basis, and it was hoped to stimulate more potent immune responses and establish a negative feedback mechanism to trigger the continual production of IFN-γ.

Moreover, it has been proved that the IFN-γ produced by the rBCG-AEI strain was bioactive, which was determined by sandwich ELISA. The bioactivity was probably negatively affected by fusion. However, before the proteins present to major histocompatibility complex (MHC) class II molecules, they will be processed by enzymatic digestion, and will become antigenic peptides. It is believed that the stronger protection induced by BCG-AEI over BCG-AE is due to two possibilities. One is that IFN-γ in the fusion construct may promote the development of protective immunity if IFN-γ is still active; another is that the fusion protein configuration may benefit the processing and presentation of the protective epitopes from the bacterial protein by MHCII to T cells. Previous work indicated that cytokine-secreting BCG recombinants may serve as improved reagents for bladder cancer immunotherapy (Arnold et al., 2004), and BCG strains secreting cytokines, in particular GM-CSF, IL-2 and IFN-γ, can modify and potentiate the immune response to BCG antigens (Murray et al., 1996). Further, the local expression of IFN-γ by the rBCG results in more efficient bacterial clearance that is accompanied by a reduction in tissue pathology (Wangoo et al., 2000). IFN-γ production may act in part by inducing the production of inducible nitric oxide (iNOS) (Chan et al., 1992), and iNOS expression may have a dual role. The present results demonstrated that mouse splenocytes immunized with rBCG-AEI released less of IFN-γ compared with rBCG-A and rBCG-AE at 4 weeks in response to Ag 85B and PPD; however, in the group of immunized with rBCG-AEI, the IFN-γ secretion highly increased at 8 weeks. IFN-γ secreted from immunized with rBCG-AEI mice might upregulate iNOS in the earlier stages of infection than mice immunized with other vaccines and may therefore, induce a more rapid sacrificing effect. But iNOS might also play a role in the down-regulation of IFN-γ and hence may have a negative feedback on iNOS production before any tissue damage is caused (Wangoo et al., 2000). Hence, in the later stage, the secretion of IFN-γ can maintain a balance. Nonetheless, IFN-γ secretion decreased at 12 weeks, and was not produced continually. On the other hand, iNOS production at the sites of granulomas might serve to limit the extent of immunopathology by down-regulating the production of proinflammatory cytokines (Wangoo et al., 2000); this result agrees with the point that there was less lung consolidation in the rBCG-AEI-vaccinated mice compared with animals vaccinated with BCG or other rBCG.

The aim in this study was to enhance qualitatively the antigenicity of BCG, but the safety of the rBCG should be considered. Therefore, further study should be performed such as altering the vaccinated dose and controlling the expressing levers of recombinant protein in rBCG. In addition, the stability of plasmid in the absence of selection should be considered.

In the conclusion, the present study revealed that a novel rBCG strain coexpressing Ag85B, ESAT-6 and mouse-IFN-γ can effectively protect mice against M. tuberculosis H37Rv infection.


The authors are very grateful to Dr Junyan Liu and Dr Shengwu Liu (Wuhan University) for their help with the animal challenge experiment. This work was supported by grants from the Major Programs of Science and Technology Commission Foundation of Shanghai (03DZ19230) and the National High Technology Research and Development Program of China (863 Program) (2004AA212502).


  • Editor: Kai Man Kam


View Abstract