Research ArticleVaccines

A Defined Tuberculosis Vaccine Candidate Boosts BCG and Protects Against Multidrug-Resistant Mycobacterium tuberculosis

See allHide authors and affiliations

Science Translational Medicine  13 Oct 2010:
Vol. 2, Issue 53, pp. 53ra74
DOI: 10.1126/scitranslmed.3001094


Despite the widespread use of the childhood vaccine against tuberculosis (TB), Mycobacterium bovis bacillus Calmette-Guérin (BCG), the disease remains a serious global health problem. A successful vaccine against TB that replaces or boosts BCG would include antigens that induce or recall the appropriate T cell responses. Four Mycobacterium tuberculosis (Mtb) antigens—including members of the virulence factor families PE/PPE and EsX or antigens associated with latency—were produced as a single recombinant fusion protein (ID93). When administered together with the adjuvant GLA-SE, a stable oil-in-water nanoemulsion, the fusion protein was immunogenic in mice, guinea pigs, and cynomolgus monkeys. In mice, this fusion protein–adjuvant combination induced polyfunctional CD4 T helper 1 cell responses characterized by antigen-specific interferon-γ, tumor necrosis factor, and interleukin-2, as well as a reduction in the number of bacteria in the lungs of animals after they were subsequently infected with virulent or multidrug-resistant Mtb strains. Furthermore, boosting BCG-vaccinated guinea pigs with fusion peptide–adjuvant resulted in reduced pathology and fewer bacilli, and prevented the death of animals challenged with virulent Mtb. Finally, the fusion protein elicited polyfunctional effector CD4 and CD8 T cell responses in BCG-vaccinated or Mtb-exposed human peripheral blood mononuclear cells. This study establishes that the protein subunit vaccine consisting of the fusion protein and adjuvant protects against TB and drug-resistant TB in animals and is a candidate for boosting the protective efficacy of the childhood BCG vaccine in humans.


Tuberculosis (TB) is a chronic infectious disease caused by Mycobacterium tuberculosis (Mtb). The World Health Organization has declared TB a global public health emergency and predicts that almost 1 billion people will be infected, with 35 million dying from the disease by 2020. TB persists as a global health concern in part because infected individuals are often noncompliant with the 6-month or longer drug treatment. This is particularly true in the developing world where more than 95% of infections occur. Treatment noncompliance has contributed to the current TB pandemic by increasing the probability of transmission and sustaining the development of multidrug-resistant (MDR) strains of Mtb. MDR strains are resistant to the two most powerful first-line drugs, rifampicin and isoniazid. Since the discovery of MDR-TB in the 1990s, the resistance pattern of TB has continued to evolve, and isolates resistant to both first- and second-line drugs have been identified (1). Therefore, development of a safe, effective, and affordable prophylactic vaccine that also provides long-lasting protection in bacillus Calmette-Guérin (BCG)–immunized people is a critical step toward controlling TB.

High mortality and morbidity continue to be associated with Mtb infections despite the widespread use of BCG, a live attenuated vaccine. BCG is the only TB vaccine currently licensed for use in humans and appears to be effective at preventing severe disseminated disease in newborns and young children, but fails to protect against pulmonary TB in adults (2). Even though variable success is obtained with BCG vaccination in human trials, BCG is unlikely to be replaced in the near future and is the gold standard to which all other experimental vaccines are compared. A few countries with lower incidence of TB, including the United States, have not adopted or have withdrawn from routine BCG vaccination, preferring to screen for and treat TB with antibiotics. Therefore, the development of a TB vaccine that is safe and efficacious in BCG- and non–BCG-primed individuals alike would be advantageous.

Protective immunity to TB is conferred by T helper 1 (TH1) CD4 and CD8 T cells (3), and several groups have shown, in various TB animal models, that an effective vaccine requires the generation of a T cell–mediated immune response. We have recently identified Mtb antigens, recognized by human T cells that elicit dominant TH1 responses, which are associated with reduced bacterial burden in a mouse model of TB (4). On the basis of this knowledge, we designed a recombinant subunit vaccine, called ID93, which combines four antigens belonging to families of Mtb proteins associated with virulence (Rv2608, Rv3619, and Rv3620) or latency (Rv1813). In general, recombinant proteins are poorly immunogenic by themselves and require an adjuvant to elicit adaptive immune responses. We have developed a synthetic monophosphoryl lipid A (MPL) and formulated it in a stable oil-in-water emulsion called GLA-SE (also known as EM005) (5). This adjuvant system has been successfully combined with antigens for the induction of high antibody titers (6, 7) and of TH1 immune responses that have been associated with protection in TB and Leishmania challenge models (8, 9).

Here, we characterized immune responses induced by ID93/GLA-SE in mice, guinea pigs, and cynomolgus monkeys. We report on the prophylactic efficacy of this vaccine (i) in mice against Mtb H37Rv and MDR strain TN5904, and (ii) in guinea pigs against Mtb H37Rv with a BCG prime-boost strategy.


Human peripheral blood mononuclear cells show cytokine recall responses to recombinant fusion protein ID93

The recombinant fusion protein ID93 is composed of the four Mtb antigens Rv3619, Rv1813, Rv3620, and Rv2608 linked in tandem (Fig. 1A). SDS–polyacrylamide gel electrophoresis (SDS-PAGE) analysis of three purified, scaled-up fermentation lots of ID93 revealed a single major band at the expected molecular mass of 93 kD (Fig. 1B). A similar band profile was observed for samples in reducing and nonreducing conditions, demonstrating the absence of protein aggregates. The identity of the major band and relative absence of aggregates and large protein fragments were confirmed by immunoblotting with a mouse polyclonal serum raised against ID93 (Fig. 1C). Furthermore, ID93 was recognized by polyclonal sera raised against each of the single proteins, confirming the presence of the four antigens within the fusion protein (Fig. 1D). Finally, the absence of bacterial by-products in the purified ID93 product was confirmed by immunoblotting with polyclonal antibody to Escherichia coli (Fig. 1C).

Fig. 1

ID93 protein construct and characterization. (A) Schematic of the ID93 fusion protein. (B to D) SDS-PAGE and immunoblot of three lots of ID93. (B) ID93 (2 μg per lane; lanes 1 to 3) was run in reducing and nonreducing conditions on a 4 to 20% tris-glycine gel. (C) Immunoblot of ID93 with mouse antibody to ID93 (α-ID93) and rabbit antisera to E. coli (α-E. coli) (50 ng and 1 μg of ID93, respectively). (D) Immunoblot of ID93 with mouse antibody to Rv3619, Rv1813, Rv3620, and Rv2608 (50 ng of ID93). Lanes 1 to 3, ID93; lane 4, Rv3619; lane 5, Rv1813; lane 6, Rv3620; lane 7, Rv2608. EC, E. coli protein standards; Ctrl, control.

The antigenicity of ID93 was tested by ex vivo stimulation of human leukocytes. Peripheral blood mononuclear cells (PBMCs) from seven subjects reactive to Mtb purified protein derivatives (PPD+), either with (BCG+) or without (BCG) a history of BCG vaccination, were used to characterize interferon-γ (IFN-γ), tumor necrosis factor (TNF), and interleukin-2 (IL-2) cytokine responses to ID93 stimulation (Fig. 2). PBMCs from PPD subjects were used as negative controls. We determined the percent of CD4CD45RO+ and CD8CD45RO+ memory T cells expressing one (1+), any combination of two (2+), or all three cytokines (3+) by intracellular cytokine staining (ICS) and flow cytometry. ID93 stimulation of PPD+ PBMCs resulted in a higher frequency of CD4 and CD8 T cells expressing one or more of these three cytokines compared to PPD controls, indicating that the fusion protein was antigenic and retained an activity similar to that of the single antigens (4). Furthermore, T cells from BCG-vaccinated adults (PPD+BCG+) were as responsive to ID93 stimulation as T cells from Mtb-exposed subjects (PPD+BCG), suggesting that ID93 might be efficacious in a BCG prime/protein boost vaccine regimen in previously vaccinated people.

Fig. 2

Human PPD+ CD4 and CD8 T cells respond to ID93 antigen stimulation. PBMCs from seven healthy PPD+ subjects (three with a history of BCG vaccination) with diverse human leukocyte antigen types were incubated for 8 hours with medium, ID93 (20 μg/ml), PPD (10 μg/ml), or SEB (1 μg/ml). T cells were identified by ICS based on CD3 expression and further gated as CD4/CD45RO+ or CD8/CD45RO+ memory T cells. Data are expressed as percent of CD4 (top) and CD8 (lower) T cells expressing IFN-γ, TNF, IL-2, or combinations of two (red) or three (blue) of the cytokines in response to antigen stimulation.

The frequency hierarchy of 1+ T cells was IFN-γ (46%) > TNF (22%) > IL-2 (7%), and IFN-γ/TNF double-positive T cells were the most frequent among the 2+ T cells (9%). CD4 and CD8 T cells from all donors responded to staphylococcal enterotoxin B (SEB) superantigen stimulation, whereas only PPD+ PBMCs stained positive for cytokines in response to PPD antigen (Fig. 2).

GLA-SE–adjuvanted ID93 vaccine generates TH1 immune responses that protect against TB and MDR-TB

We conducted experiments to characterize the immunogenicity of ID93 adjuvanted with the synthetic MPL adjuvant formulation GLA-SE in C57BL/6 mice. ID93-specific antibody and T cell responses were measured 1 and 3 weeks, respectively, after the last injection. Vaccination with ID93/GLA-SE, but not with GLA-SE alone, induced measurable levels of antigen-specific immunoglobulin G1 (IgG1) and IgG2c (Fig. 3A). The mean of reciprocal dilutions was 6.0 ± 0.1 log10 (IgG1) and 6.3 ± 0.2 log10 (IgG2c). The elevated titers of ID93-specific IgG2c and a positive IgG2c/IgG1 log10 ratio are indicative of an IFN-γ–dependent switch in IgG subclasses. Characterization of cytokine expression after ID93 in vitro recall of vaccine-induced T cells further confirmed that IFN-γ, IL-2, and TNF were produced by CD4 TH1 phenotypic cells (Fig. 3B). Expression of IFN-γ, TNF, and/or IL-2 by splenic CD4/CD44high or CD8/CD44high T cells was determined at the single-cell level by ICS. There was a fourfold increase in the frequency of IFN-γ+CD4/CD44high T cells responding to ID93 stimulation in mice injected with ID93/GLA-SE, whereas cells from the GLA-SE–only control did not respond to ID93 (Fig. 3B). The proportion of CD4/CD44high T cells expressing one, two, or three cytokines was also determined. The small numbers of ID93-responsive CD4/CD44high T cells from mice injected with GLA-SE alone were mainly of the 1+ phenotype (96% of cells); most of these 1+ cells expressed TNF (45%) or IL-2 (41%). In contrast, CD4 T cells from animals that received the ID93/GLA-SE vaccine showed a significant increase in both the 2+ (34% of cells, with 84% being IFN-γ/TNF double positives) and the 3+ T cells (5%). Frequencies of IFN-γ–, TNF-, and IL-2–producing ID93-specific CD8 T cells in the ID93/GLA-SE vaccine group were <0.1%, suggesting that ID93/GLA-SE induced mainly a CD4 TH1 cell immune response.

Fig. 3

Immunogenicity and efficacy of ID93/GLA-SE against Mtb H37Rv and MDR TN5904 strains in mice. (A) ID93-specific IgG1 and IgG2c antibodies were measured on day 49 in sera from animals immunized three times, 3 weeks apart with GLA-SE (20 μg of TLR-4 agonist) or ID93 (8 μg) plus GLA-SE. Data are shown as mean of reciprocal dilution ± SD. (B) Cytokine recall responses to ID93 of T cells in vitro were measured 3 weeks after the last immunization by ICS and flow cytometry. CD4 T cells were identified by CD3 expression and further gated as CD44hi. Data show the percentage of cells expressing IFN-γ in response to ID93 stimulation (mean ± SD, n = 3 mice), and pie charts below show the proportion of cells expressing one, two, or three of the cytokines IFN-γ, TNF, and IL-2. Data are representative of two experiments. (C and D) The number of viable bacteria in the lungs in mice was determined 3 or 4 weeks after challenge with aerosolized TN5904 (D) or H37Rv (C) strains, respectively. One-way ANOVA followed by Dunnett’s multiple comparison test was used for statistical analysis (vaccine groups were compared to saline control group). **P < 0.01. (E and F) Histopathological evaluation of lung tissues after challenge with H37Rv (E) or TN5904 (F). Granuloma (g) formation is shown in H&E-stained sections (scale bar, 20 μm), and AFBs (→) were evaluated (scale bar, 5 μm). Data shown are representative of four mice per group in two independent experiments.

ID93/GLA-SE vaccine efficacy was subsequently tested in mice challenged with H37Rv or MDR TN5904 Mtb strains. Animals were vaccinated with GLA-SE, ID93/GLA-SE, BCG, or saline (control) and challenged by aerosol administration of Mtb. ID93/GLA-SE significantly reduced the number of viable bacteria in the lung [colony-forming unit (CFU)] by 0.48 log10 (H37Rv) and 0.52 log10 (TN5904) compared to saline controls (P < 0.01), whereas immunization with GLA-SE alone was ineffective (−0.06 log10 and 0.05 log10, respectively; P > 0.05) (Fig. 3, C and D, and Table 1). In comparison, one dose of BCG resulted in a decrease in lung CFU of 0.92 log10 (H37Rv) and 1.20 log10 (TN5904). A similar effect of ID93/GLA-SE was observed in the spleens of mice challenged with H37Rv, whereas no significant reduction was seen in this group after exposure to MDR TN5904 (Table 1). Hematoxylin and eosin (H&E) staining of formalin-fixed lung tissues showed few differences in granuloma size and lung cellularity between vaccine groups (Fig. 3, E and F). However, numerous acid-fast bacilli (AFB) could be detected throughout the lung sections of saline- and GLA-SE–immunized mice, whereas few AFBs were seen in the lungs of mice vaccinated with ID93/GLA-SE or BCG. AFB results correlated well with the CFU of bacteria observed in these animals (Table 1). Finally, similar pathology and AFBs were observed in ID93/GLA-SE–vaccinated mice after infection with both H37Rv and MDR TN5904 Mtb strains.

Table 1

ID93/GLA-SE–induced protection against TB and MDR-TB.

View this table:

Analysis of cellular infiltrates to the lung after H37Rv challenge revealed an increase in granulocytes (Gr-1+; Fig. 4A) and T cells (Fig. 4B) in the GLA-SE and ID93/GLA-SE vaccine groups compared to saline or BCG groups. Unprotected animals (saline and GLA-SE groups) showed a higher frequency of CD44high CD4 and CD8 T cells responding to in vitro recall stimulation with PPD and ID93 (Fig. 4C), with most of the cells staining positive for TNF or IFN-γ (Fig. 4D). In contrast, the ID93/GLA-SE and BCG groups had fewer T cells responding to in vitro recall, but most of the responding cells in the lungs stained for both TNF and IFN-γ. The number of CD44high CD8 T cells positive for granzyme B was also higher in saline (734 ± 229) and GLA-SE (1255 ± 277) groups compared to the BCG (372 ± 200) and ID93/GLA-SE (571 ± 225) groups.

Fig. 4

Lung cellular infiltrates after challenge. Mice were immunized three times, 3 weeks apart with saline, 20 μg of GLA-SE, and 8 μg of ID93/GLA-SE, or once with 5 × 104 live BCG. Four weeks after the last boost, mice were challenged with an aerosol dose of 50 to 100 H37Rv bacilli. Lung cell phenotype and in vitro cytokine recall responses to ID93 were measured by ICS and flow cytometry 4 weeks after challenge. (A) Number of Gr-1+ cells in the lungs. SSC, side scatter; FSC, forward scatter. (B) Number of T cells (×103), identified based on CD3 expression and further gated as CD4CD44lo, CD4CD44hi, CD4+CD44lo, or CD4+CD44hi. (C) Number of CD44hiCD4+ and CD44hiCD8+ T cells expressing TNF and/or IFN-γ in response to media, ID93 (10 μg/ml), or PPD (10 μg/ml) stimulation. Mean ± SD (n = 2 pools of three mice each) are shown for a representative experiment. (D) Proportion of cells expressing IFN-γ, TNF, or TNF/IFN-γ effector cytokines in response to ID93 in vitro stimulation.

GLA-SE–adjuvanted ID93 vaccine boosts BCG and confers long-term protection in a guinea pig model of TB

Childhood vaccination with a single dose of the live attenuated BCG vaccine is standard practice in most of the world outside the United States and will not be quickly replaced. Its efficacy, however, wanes with time and additional injections during adolescence or adulthood are ineffective (2). Therefore, the development of a subunit vaccine that can be administered to BCG-immunized individuals and successfully boost their waning protective immune responses is of the utmost importance.

To demonstrate the efficacy of ID93/GLA-SE in a heterologous prime-boost strategy, we gave guinea pigs a single dose of BCG or saline (control). Four months later, half of the animals in the BCG groups received the ID93/GLA-SE vaccine, whereas the other half was injected with saline. ID93-specific IgG1 and IgG2 antibodies were detected in sera of animals from the BCG→ID93/GLA-SE vaccine group (Fig. 5A). The mean of reciprocal dilutions was 5.0 ± 0.5 log10 (IgG1) and 6.9 ± 0.3 log10 (IgG2) 1 week after the last vaccine injection, and 4.6 ± 0.2 log10 (IgG1) and 5.0 ± 0.2 log10 (IgG2) more than 1 year later. The mean of reciprocal dilutions was <2.0 log10 in the BCG→saline or saline groups. Animals were challenged with Mtb and monitored for weight loss thereafter as an indicator of declining health and increased morbidity associated with TB (Fig. 5B). All animals in the saline control group (nine of nine) and two-thirds of those in the BCG→saline vaccine group (six of nine) exhibited >20%weight loss (compared to their preinfection body weight) and had to be euthanized or succumbed to the disease by day 432 after infection (Fig. 5C). On day 432, guinea pig 383-793 in the BCG→saline vaccine group showed signs of illness with 10.4% body weight loss. In contrast, only one of eight guinea pigs died in the BCG→ID93/GLA-SE vaccine group during the same period, and zero of seven exhibited symptoms (weight loss of ≥10%). The median survival was 245 and 389 days for the saline and BCG→saline groups, respectively (P < 0.001). The decreased mortality of the BCG→ID93/GLA-SE vaccine group was statistically different compared to the saline (P < 0.001) and BCG→saline (P < 0.05) groups.

Fig. 5

ID93/GLA-SE vaccine confers long-term protection against Mtb in guinea pigs in a BCG prime/protein boost regimen. Guinea pigs received a single injection of either saline (one of three groups) or BCG (two groups). Four months later, BCG-primed animals received the ID93/GLA-SE vaccine or saline. (A) Mean reciprocal dilution of serum ID93-specific IgG1 and IgG2 antibodies ± SD. (B and C) Weekly weight monitoring after Mtb challenge for overt signs of disease and weight loss (B), resulting in euthanasia of the sick animals (C). Log-rank test was used for statistical comparisons of median animal survival among the experiment groups. *P < 0.01, compared to saline; **P < 0.01, compared to BCG→saline. (D and E) H&E-stained sections of lung and spleen, respectively, of each vaccinated animal on day 432 after challenge with H37Rv. Scale bar, 0.2 mm. Guinea pig 547-834 in the BCG→saline group was euthanized on day 418 because of morbidity and excessive weight loss. (F) Granulomas are shown in H&E-stained sections, and Masson’s trichrome (TC) stain was used to assess the amount of lung fibrosis in blue. Scale bar, 20 μm. Data shown are representative of four animals per group; GP, guinea pig.

Histopathological analyses were performed on formalin-fixed lung tissues (Fig. 5, D and F) and liver (Fig. 5E) of four animals from each group that were still alive at study termination on day 432, with the exception of guinea pig 547-834 in the BCG→saline group, whose organs were preserved at euthanasia on day 418. Three animals in the BCG→ID93/GLA-SE group showed no lesions in the lungs (score = 0), whereas one animal exhibited a mild, multifocal granulomatous pneumonia (score = 2). Minimal to mild cellular infiltrations (score = 1 to 2) were observed in the livers. No AFBs were found in any of the lung or liver sections analyzed in the BCG→ID93/GLA-SE group.

In the BCG→saline group, lung sections of animals 547-834 and 383-793 showed marked granulomatous bronchopneumonia (score = 4). Lung granulomas had no evidence of central caseous necrosis or mineralization. We found scattered lymphocytes throughout the granuloma and a few infiltrating neutrophils. A few AFB in the sections of 547-834 and none in those from 383-793 were observed. Lung sections from 530-262 and 535-834 showed minimal to mild histological changes and no AFB (score = 1 and 2, respectively). Liver sections from the animals in the BCG→saline group showed pathologies ranging from none or one lesion (530-262 and 535-834, score = 0 and 1, respectively) to marked, multiple coalescing foci of TB granulomatous hepatitis (383-793 and 547-834, score = 3 and 4, respectively). No AFBs were detected in three of four animals in the BCG→saline group, and minimal AFBs were present in liver sections from 547-834.

Two of three (67%) guinea pigs in the BCG→saline group had measurable CFU when we plated lung homogenates and tested for bacilli growth, whereas only three of seven (43%) animals were positive in the BCG→ID93/GLA-SE group (Fig. 5D). In spleens, the numbers of animals with detectable CFU were two of three (67%) with 4.65 and 4.30 log10 CFU, and one of seven (14%) with 3.55 log10 CFU, respectively.

GLA-SE–adjuvanted ID93 is safe and immunogenic in cynomolgus macaques

To demonstrate the safety and immunogenicity of ID93/GLA-SE in nonhuman primates, we gave six male cynomolgus macaques three doses of the vaccine. Injection site reactions were minimal, with no more than barely perceptible erythema and edema (Draize scale range, 0 to 1) observed for up to 3 days after immunization, and there were no significant changes in body weight and temperature observed (fig. S1).

To characterize the cell-mediated immune response to the vaccine antigen by peripheral leukocytes, we performed a cytokine profiling assay on blood collected from animals before immunization and 2 weeks after the third dose. We determined the concentration of 16 cytokines in whole-blood supernatants in response to restimulation with ID93: IFN-γ, TNF, GM-CSF (granulocyte-macrophage colony-stimulating factor), MCP-1 (monocyte chemoattractant protein-1), IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p40, IL-13, IL-15, IL-17, and IL-18. The results showed that in response to ID93, a subset of these cytokines representing proinflammatory, as well as TH1 and TH2 functional groups, was significantly up-regulated (Fig. 6). TNF, a soluble mediator of Mtb-specific immunity in infected individuals, was significantly up-regulated after vaccination with ID93/GLA-SE. The median concentration of TNF in ID93-stimulated blood from animals that had been vaccinated was ~60 times higher when compared to blood from the same animals before vaccination (133 versus 2 pg/ml, P = 0.01). In addition, ID93-specific GM-CSF and IFN-γ responses were detected, but these differences were more modest and the latter were not statistically significantly different (11 versus <1.5 pg/ml, P = 0.02, and 19 versus <1.5 pg/ml, P = 0.06). No significant difference in the concentration of IL-2 was detected in this assay, and although significant ID93-specific IL-5 responses were measured, there were no changes detected in the concentrations of IL-4 and IL-13, the other two TH2-type cytokines assayed. ID93-specific responses for all the other T cell–associated cytokines assayed were below the detection limit (smaller than 1.4 pg/ml). The greatest differences measured were in IL-6 and IL-8, which had median concentrations from two and four logs higher in blood collected from ID93/GLA-SE–vaccinated animals than in blood collected from the same animals at study initiation [median, 2776 versus 24 pg/ml (P = 0.05) and 10,671 versus <1.5 pg/ml (limit of detection) (P = 0.02), respectively]. Two additional proinflammatory mediators, MCP-1 and IL-1β, were also significantly up-regulated in blood from vaccinated animals after restimulation with ID93 when compared to blood from the same animals before vaccination (522 versus <1.5 pg/ml, P = 0.05, and 302 versus <1.5 pg/ml, P = 0.01, respectively). Overall, these results demonstrated that an ID93/GLA-SE formulation was well tolerated and immunogenic in cynomolgus macaques.

Fig. 6

Cytokine responses to vaccine antigen in immunized macaques. Peripheral blood collected from ID93/GLA-SE–immunized animals was stimulated with ID93 or saline, after which plasma was collected and assayed for cytokines by multiplexed bead array assay. Background-subtracted antigen-specific cytokine responses are shown for blood collected before immunization (day 0, ○) and after third immunization (day 71, •). *P ≤ 0.05; **P ≤ 0.01 by Wilcoxon rank-sum test (n = 6).


Controlling the spread of TB and MDR-TB has become an urgent public health priority. This goal could be achieved by developing new TB vaccines or drugs, or by improving the long-term protection provided by BCG vaccine with prime-boost strategies. We designed the subunit vaccine ID93/GLA-SE to be used in both BCG-vaccinated and nonvaccinated individuals, and performed comprehensive preclinical analyses of its immunogenicity, safety, and efficacy against challenge with live virulent Mtb. ID93 consists of the fusion of four Mtb antigens that had been previously shown to confer partial protection in a mouse model of TB (4). ID93 was further shown to induce memory T cell cytokine recall responses in BCG-vaccinated and Mtb-exposed subjects. ID93/GLA-SE vaccination induced antigen-specific antibody and/or T cell immune responses in mice, guinea pigs, and cynomolgus monkeys. Induction of predominantly TH1-type CD4 T cell responses was associated with reduced TB and MDR-TB bacterial burdens in the lungs of vaccinated mice. Furthermore, we demonstrated that boosting with the ID93/GLA-SE vaccine enhanced BCG-induced immunity to Mtb in guinea pigs by significantly reducing TB-associated pathology and mortality when compared to BCG alone.

Combining multiple antigens in recombinant fusion proteins such as Mtb72f, ID83, Ag85B-ESAT6, CSU-F36, and Ag85B-TB10 (8, 1013) leads to increased vaccine efficacy against Mtb. The ID93 vaccine candidate is composed of antigens from three different families of Mtb proteins: the EsX family of virulence factors (Rv3619 and Rv3620), the PE/PPE (Rv2608), and the latency antigens (Rv1813). Because of their low intrinsic immunogenicity, protein-based vaccines need suitable adjuvant systems for the induction of strong in vivo immune responses. Three of the most advanced subunit vaccines against nonviral diseases currently in human clinical trials—RTS,S [malaria vaccine (14)], Mtb72F [TB vaccine (15)], and LEISH-F1 [leishmaniasis vaccine (16)]—are based on liposomal (AS01) or oil-in-water (AS02 and MPL-SE) formulations containing the Toll-like receptor–4 (TLR-4) agonist MPL. Our institute developed a proprietary synthetic MPL molecule in an oil-in-water emulsion called GLA-SE (5). Our selection of GLA-SE was based on its ability to activate dendritic cells in a TLR-4–dependent fashion and potentiate TH1-type CD4 T cell responses (8, 9), which are critical for controlling Mtb infection.

Three doses of the ID93/GLA-SE vaccine induced antigen-specific TH1 responses in mice, guinea pigs, and nonhuman primates. IgG2/IgG1 antibody ratios >1 and/or increased production of IFN-γ and/or TNF were observed. In mice, the ability of ID93/GLA-SE to induce polyfunctional T cell responses was associated with a partial reduction in lung and spleen bacterial burden after aerosol exposure to Mtb. Only one study to date has documented the efficacy of a TB vaccine against MDR-TB (17). In that study, Okada et al. used a DNA vaccine expressing Mtb heat shock protein HSP65 and human IL-12 delivered by the hemagglutinating virus of Japan encapsulated in liposomes. However, it remains unclear whether prophylactic DNA-based vaccines are a viable technology for human use because of safety concerns about the possibility of integration of foreign genetic material.

Most people in the world have been vaccinated with BCG; however, its efficacy wanes over time. Therefore, boosting the protective immunity provided by BCG might be the most realistic path toward a new TB vaccine. Homologous boosting with BCG was ineffective in humans (2) and caused severe disseminated lesions in guinea pigs (18). As a result, heterologous prime-boost strategies such as priming with BCG and boosting with adjuvanted (19, 20) or vectored (21) antigens found in BCG and/or Mtb have been pursued with promising results in mice. Given that all four Mtb antigens that make up ID93 are also found in BCG, ID93 may prove to be a superior vaccine candidate for boosting BCG-induced immunity. We demonstrated in the more susceptible TB guinea pig model that boosting BCG with ID93/GLA-SE reduced pathology and afforded long-term protection against TB. Protection with ID93/GLA-SE in guinea pigs was similar to that observed with Mtb72f/AS02A (22) and appeared greater than with the HyVac4/IC31 (23) or AdAg85A (24) vaccines, all three currently in clinical trials. The ID93-specific cytokine response in the blood of vaccinated cynomolgus macaques was a complex mixture of TH1- and TH2-type responses, a phenotype consistent with observed T cell memory responses in successful vaccine regimens, such as for yellow fever (25). Our findings are also consistent with previous work demonstrating that another subunit vaccine, Mtb72f/AS02A, was immunogenic in macaques and increased protection against Mtb in a prime-boost regimen (10). A recent publication by van Dissel et al. suggests that humans may also react slowly to adjuvanted protein vaccines, but eventually generate a substantial response (26). Our most important finding is the boosting potential of ID93 demonstrated with BCG-vaccinated guinea pigs and humans. When tested on PBMCs isolated from healthy PPD+ BCG-vaccinated subjects, ID93 induced preexisting CD4 and CD8 T cell IFN-γ and TNF cytokine responses, suggesting that the ID93/GLA-SE vaccine may also boost BCG-induced protective responses. In conclusion, these results support the advancement of the ID93/GLA-SE TB vaccine to a phase I clinical trial.

Materials and Methods

Cloning and purification of ID93

ID93 was generated through a tandem fusion of the individual cloned and amplified genes of Rv3619, Rv1813, Rv3620, and Rv2608 with restriction site linkers. The recombinant pET28a plasmids (Novagen) containing the individual Rv3619, Rv1813, Rv3620, and Rv2608 genes were previously described (4). ID93 polymerase chain reaction (PCR) primers were designed to incorporate specific restriction enzyme sites 5′ and 3′ of the gene of interest with primer sequences as follows: Rv1813–5′Nde I, CAATTACATATGGGTACCCATCTCGCCAACGGTTCGATG; Rv1813–3′Sac I, CAATTAGAGCTCG- TTGCACGCCCAGTTGACGAT; Rv3620–5′Sac I, CAATTAGAGCTCATGACCTCGCGTTTTATGACG; Rv3620–3′Sal I, CAATTAGTCGACGCTGCTGAGGATCTGCTGGGA; Rv2608–5′Sal I, CAATTAGTCGACATGAATTTCGCCG- TTTTGCCG; Rv2608–3′Hind III, CAATTAAAGCTTTTAAGTACTGAAAAGTCGGGGTAGCGCCG; Rv3619–5′Nde I, CAATTACATATGACCATCAACTATCAATTC; and Rv3619–3′Kpn I, CAATTAGGTACCGGCCCAGCTGGAGCCGACGGC. The DNA sequences were amplified from plasmid DNA templates with Pfx DNA polymerase (Invitrogen) with 30 cycles at 94°C for 15 s, 58°C for 30 s, and 68°C for 1.5 hours. The Rv1813 PCR product was digested with Nde I–Sac I restriction enzymes and then cloned into the pET28a vector. Rv3620 and the Rv1813pET construct were digested with Sac I–Sal I and ligated. Rv2608 was digested with Sal I–Hind III and ligated into the Sal I–Hind III–cut pET28a-Rv1813-3620 vector. Rv3619 was digested with Nde I–Kpn I and then ligated into the above vector. The resulting plasmid containing the fusion gene construct Rv3619-Rv1813-Rv3620-Rv2608 was DNA-sequenced and named ID93 because it encodes a 93-kD protein. ID93 was expressed in E. coli host BL-21plysS, purified under denaturing conditions by chromatography on DEAE and Q Sepharose columns, and analyzed by SDS-PAGE on a 4 to 20% tris-glycine gel (Invitrogen). Protein identity was confirmed by immunoblotting with mouse polyclonal sera raised against ID93 (1:1000), followed by goat antibody to mouse IgG1 conjugated to horseradish peroxidase (HRP) (1:1000; Southern Biotechnologies Inc.). The presence of each of the four antigens within ID93 was confirmed by immunoblotting with mouse polyclonal sera raised against Rv3619, Rv1813, Rv3620, and Rv2608 (1:1000), followed by goat antibody to mouse IgG1 conjugated to HRP (1:1000; Southern Biotechnologies). The absence of E. coli by-products was confirmed by immunoblotting with HRP-conjugated rabbit polyclonal antibody to E. coli (1:1000; ViroStat Inc.). Residual lipopolysaccharide contamination was evaluated by the Limulus amoebocyte lysate assay (Cambrex Corp.) and determined to be <25 endotoxin units per milligram of protein.

Immunization, challenge, and CFU

All mice and guinea pigs were maintained in the Infectious Disease Research Institute (IDRI) animal care facility under specific pathogen-free conditions and were treated in accordance with the regulations and guidelines of IDRI Animal Care and Use Committee. The nonhuman primate study protocol and all experimental procedures were reviewed and approved by the SNBL USA Institutional Animal Care and Use Committee before study initiation.

Female C57BL/6 mice, 4 to 6 weeks old (Charles River), were immunized subcutaneously three times, 3 weeks apart with ID93 (8 μg) + GLA-SE (20 μg). The BCG group received a single intradermal dose of 5 × 104 CFU (Pasteur strain, Sanofi Pasteur). Three to 4 weeks after the last immunization, groups of seven mice were challenged by low-dose aerosol exposure to Mtb strain H37Rv (American Type Culture Collection) or MDR strain TN5904 (provided by B. N. Kreiswirth, Public Health Research Institute TB Center, Newark, NJ) using a University of Wisconsin-Madison aerosol exposure chamber calibrated to deliver 50 to 100 bacteria into the lungs. TN5904 is a non–W Beijing family MDR clinical isolate with resistance to streptomycin, isoniazid, rifampicin, and sodium p-aminosalicylate (27). Vaccine efficacy was determined 4 weeks after challenge by harvesting lungs and spleen from the infected mice and counting CFU as described (4).

Female outbred Hartley guinea pigs, 400 to 450 g (Charles River), were injected intradermally with a single dose of 5 × 104 live BCG (prime) and, 4 months later, immunized three times (boosts), 3 weeks apart with ID93 (20 μg) + GLA-SE (20 μg). Four weeks after the last immunization, groups of eight to nine guinea pigs were challenged by low-dose aerosol exposure to Mtb H37Rv (20 to 50 bacteria in the lungs). Monitoring of weight loss and survival was done through day 432 after challenge, at which time the experiment was terminated.

Male cynomolgus macaques were selected from purpose-bred colonies housed at SNBL USA Ltd. and were maintained in accordance with guidelines set forth by the Association for Assessment and Accreditation of Laboratory Animal Care guidelines. All animals were between 4 and 6.5 kg and 3 and 8 years old at study initiation. Six animals received three intramuscular injections of ID93 (10 μg) + GLA-SE-H (a GLA-SE formulation suitable for use in humans; 20 μg) spaced 4 weeks apart. Blood was obtained before the first injection (day 0) and 2 weeks after the third dose (day 71) on restrained, conscious animals, and collected into Vacutainer tubes containing heparin as an anticoagulant (BD Biosciences).

Determination of antibody titers

Sera from immunized mice and guinea pigs were tested for ID93-reactive IgG1 and IgG2c, respectively, as previously described (4). HRP-conjugated antibodies to IgG1, IgG2c, or IgG2 were used at 1:2000 dilution in phosphate-buffered saline (PBS)–0.05% Tween 20–0.1% bovine serum albumin. Reciprocal dilutions corresponding to endpoint titers were determined with GraphPad Prism 4 (GraphPad Software Inc.) with a cutoff of 0.1.

Flow cytometry

Splenocytes and lungs from immunized animals (1 × 106 to 2 × 106 cells) were stimulated with antibody to CD28/CD49d (1 μg/ml each; eBioscience), PPD (10 μg/ml), or ID93 (20 μg/ml) for 8 hours at 37°C in the presence of GolgiStop (eBioscience). The cells were then fixed and stained with fluorochrome-conjugated antibodies to CD3, CD8, CD44, IFN-γ, TNF, IL-2, granzyme B (eBioscience), and CD4 (Invitrogen) as described elsewhere (4).

PBMCs were purified from heparinized blood obtained from PPD and PPD+ healthy subjects and stored frozen in liquid nitrogen. Informed consent was obtained from all the subjects, and the study was approved by Western Institutional Review Board (Seattle, WA). PBMCs (1 × 106 to 2 × 106 cells) were stimulated with antibody to CD28/CD49d (1 μg/ml each), PPD (10 μg/ml), ID93 (20 μg/ml), or SEB (1 μg/ml) for 8 hours at 37°C in the presence of GolgiStop. After stimulation, live cells were stained with Aqua Live/Dead fixable dye (Invitrogen) and then prepared for polychromatic flow cytometric analysis of CD3, CD45RO, CD4, CD8, IFN-γ, and TNF (4). Viable lymphocytes were gated by forward and side scatter, and 100,000 CD3+ events for each sample were acquired on an LSRII and analyzed with BD FACSDiva software v5.0.1 (BD Biosciences).

Peripheral blood cytokine profiling assay

Blood obtained from monkeys was assayed within 4 hours of venipuncture. Under sterile conditions, 900 μl of blood was mixed with 100 μl of either PBS, ID93 (20 μg/ml), or a mitogen-positive control (phytohemagglutinin, 400 μg/ml) and allowed to incubate for 22 to 24 hours at 37°C and 5% CO2. Plasma was then collected by centrifugation, filtered, and stored at −80°C for subsequent cytokine analysis. Quantitation of cytokine concentrations was performed with a commercial multiplex bead array assay (Millipore) specific for nonhuman primate cytokines per the manufacturer’s instructions.


A single lobe of the lung and part of the liver were fixed for at least 7 days in 10% normal buffered formalin. After fixation, tissues were embedded in paraffin, cut, and stained with H&E, Fite’s acid-fast stain, or trichrome stain (lung only) by the Benaroya Research Institute Histology Core facility for mouse tissues and BioGenetics Research Laboratories Inc. for guinea pig tissues. Between 3 and 16 nonoverlapping fields per tissue section were captured at a magnification of ×1 to ×600. Qualitative analyses were performed by a veterinary pathologist and included comparing all fields within a group (four animals per group, two to three sections examined per animal) for organization of cellular infiltrate, number of granulomas, visual quantification of AFB, and pulmonary fibrosis.

Lesion scoring was as follows: 0, normal tissue morphology with no lesion present; 1, lesions involving <10% of the tissue and minimal infiltration of fibroblast, mononuclear, or polymorphonuclear inflammatory cells; 2, lesions affecting 10 to 20% of the tissue and mild cellular infiltration; 3, lesions covering 21 to 40% of the tissue and moderate cellular infiltration; 4, lesions covering 41 to 100% of the tissue and marked cellular infiltration.

Statistical analysis

Standard one-way analysis of variance (ANOVA), followed by Dunnett’s multiple comparison test, was used for statistical analysis of CFU. Log-rank test was used for statistical comparisons of median guinea pig survival among the experimental groups. Nonhuman primate cytokine data analysis was performed with MasterPlex QT (MiraiBio) and JMP software (SAS Institute). Test groups were compared with a nonparametric Wilcoxon rank-sum test. P values ≤0.05 were considered significant.

Supplementary Material

Materials and Methods

Fig. S1. Body weight and temperature of cynomolgus monkeys after ID93/GLA-SE vaccination.


  • * Present address: Novartis Vaccines and Diagnostics, via Fiorentina 1, Siena 53100, Italy.

  • These authors contributed equally to this work.

  • Citation: S. Bertholet, G. C. Ireton, D. J. Ordway, H. P. Windish, S. O. Pine, M. Kahn, T. Phan, I. M. Orme, T. S. Vedvick, S. L. Baldwin, R. N. Coler, S. G. Reed, A defined tuberculosis vaccine candidate boosts BCG and protects against multidrug-resistant Mycobacterium tuberculosis. Sci. Transl. Med. 2, 53ra74 (2010).

References and Notes

  1. Acknowledgments: We thank A. Bhatia and R. Howard for reviewing this manuscript; D. Argilla, A. Bernard, K. Carper, T. Dutill, T. Evers, M. Henao Tamayo, E. Kristalinskaia, E. Laughlin, J. Laurance, G. Poshusta, V. Reese, C. Shanley, I. Tukacovic, W. Wicomb, I. Zharkikh, and J. Zheng for their technical expertise; and L. Kunz for guinea pig tissue histology. We thank B. N. Kreiswirth (Public Health Research Institute TB Center, Newark, NJ) for providing MDR strain TN5904. Funding: Supported in part by NIH–National Institute of Allergy and Infectious Diseases (NIAID) grants AI-044373 and AI-067251 (S.G.R.) and NIH-NIAID contracts AI-25479 and HHSN272200800045C and grant U01-AI078054 (R.N.C.). Author contributions: S.B., G.C.I., S.L.B., R.N.C., and S.G.R. designed the study, conducted the experiments, and analyzed the data; D.J.O. and I.M.O. conducted the MDR-TB experiments; T.S.V. provided GLA-SE–based adjuvant formulations; T.P. produced and analyzed ID93 lots; M.K. conducted the guinea pig experiment; H.P.W. analyzed the mouse histology; S.O.P. conducted the nonhuman primate study; S.B. drafted the paper; G.C.I., H.P.W., S.O.P., S.L.B., R.N.C., and S.G.R. contributed to the writing. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article