The ESAT-6 (early secretory antigenic target) molecule is a very important target for T cell recognition during infection with Mycobacterium tuberculosis. Although ESAT-6 contains numerous potential T cell epitopes, the immune response during infection is often focused toward a few immunodominant epitopes. By immunization with individual overlapping synthetic peptides in cationic liposomes (cationic adjuvant formulation, CAF01) we demonstrate that the ESAT-6 molecule contains several subdominant epitopes that are not recognized in H-2d/b mice either during tuberculosis infection or after immunization with ESAT-6/CAF01. Immunization with a truncated ESAT-6 molecule (Δ15ESAT-6) that lacks the immunodominant ESAT-61–15 epitope refocuses the response to include T cells directed to these subdominant epitopes. After aerosol infection of immunized mice, T cells directed to both dominant (ESAT-6-immunized) and subdominant epitopes (Δ15ESAT-6-immunized) proliferate and are recruited to the lung. The vaccine-promoted response consists mainly of double- (TNF-α and IL-2) or triple-positive (IFN-γ, TNF-α, and IL-2) polyfunctional T cells. This polyfunctional quality of the CD4+ T cell response is maintained unchanged even during the later stages of infection, whereas the naturally occurring infection stimulates a response to the ESAT-61–15 epitope that consist almost exclusively of CD4+ effector T cells. ESAT-6 and Δ15ESAT-6 both give significant protection against aerosol challenge with tuberculosis, but the most efficient protection against pulmonary infection is mediated by the subdominant T cell repertoire primed by Δ15ESAT-6.
Mycobacterium tuberculosis (MTB)3 establishes a life-long chronic infection in the face of a very strong immune response, and the background for this paradox is not well understood. The existing vaccine bacillus Calmette-Guérin (BCG) has, despite being widely used, a very limited if any effect against the establishment of latent TB and reactivation. All new vaccine candidates currently in clinical trials are classical preventive vaccines based on Ags that are recognized in infected individuals, such as the components from the ESAT-6 (early secretory antigenic target) or Ag 85 family (1, 2). However, as a treated and cured primary infection does not always protect against a secondary infection (3, 4), and a significant proportion of latently infected individuals reactivate their disease, future vaccines must address this problem to be successful. Immune responses generated during a TB infection are often very strong, and the low-molecular mass-secreted Ag ESAT-6 is the single Ag that attracts most attention from the immune system (5, 6). Compared with the numerous potential epitopes that ESAT-6 contains (7), the response in infected individuals (8) and in experimental animals (9) is often focused on a very limited number of T cell epitopes. A similar phenomenon is observed in chronic viral diseases and is referred to as immune focusing. That such a bias of the immune response is inappropriate for the hosts’ attempts to the control infection were demonstrated HIV-1, for example, where a highly focused Ab response against immunodominant epitopes in hypervariable regions on the genome were reported to dysregulate the immune system’s ability to focus on protective subdominant epitopes in less variable regions (10, 11, 12). Also in hepatitis C virus and HIV, focusing of the immune response has been associated with a poor prognosis, whereas donors characterized by a broader T cell repertoire controlled the disease (13, 14). One way of broadening the repertoire is through priming of immune responses to subdominant epitopes that are normally “silent” and not involved in the response during infection or primed by conventional immunization. Such a refocusing of the Ab response to subdominant epitopes has for viral pathogens been shown to result in a more broadly protective response (15, 16). Thus, we hypothesized that a potential problem associated with current immunization strategies against a chronic disease such as TB is that they prime an immune response to epitopes that are already dominant T cell targets during the natural infection. In the present study we have therefore focused on a strategy that aims at filling the holes in the T cell repertoire with new T cell specificities that are not recognized during the natural infection. The fact that they remain silent and are not primed during the natural infection can be due to a number of reasons: (1) that they are not produced in any significant quantity by the bacteria and therefore are not available for the immune system, (2) that they are not processed and presented in the context of the host MHC, and (3) that they lose the competition to other epitopes that may have higher affinity. However, when primed through immunization, we have previously observed that T cells directed at least to some of these subdominant epitopes can recognize and engage the infected macrophage and contain or eliminate MTB infection (17). Also, a number of immunization studies against viral diseases have demonstrated that CTL responses to subdominant epitopes can compensate for the loss of a dominant epitope and provide efficient protection against subsequent challenge (18, 19), and focusing on subdominant responses have, for example, in HIV been suggested as a potential strategy to avoid escape mutants that under selective pressure delete their dominant epitopes (20).
In the present study we have engineered an ESAT-6 vaccine molecule (Δ15ESAT-6) that lacks the immunodominant epitope localized within the first 15 aa of the ESAT-6 sequence (9) and analyzed the immune response raised to this molecule in a mouse model that has a natural immune response during infection highly targeted to the ESAT-61–15 epitope (9). We observed a refocusing of the response in the animals immunized with Δ15ESAT-6 to subdominant epitopes that were not promoted by immunization with the full-size molecule or during the natural infection. After aerosol infection, the vaccine-promoted T cells were expanded and recruited to the lung where they maintained a stable high quality phenotype with coexpression of several cytokines. Interestingly, the Δ15ESAT-6-promoted response toward subdominant epitopes provided a more efficient protection against an MTB infection than did T cells specific for the immunodominant ESAT-61–15.
Materials and Methods
Inbred female CB6F1 (BALB/C x C57BL/6) mice were obtained from Harlan Scandinavia. Infected mice were housed in cages with a laminar flow safety enclosure and provided with irradiated food and filtered drinking water. All animal experiments were done according to the Danish Ministry of Justice and animal protection committees and in compliance with European Community Directive 86/609.
Δ15ESAT-6 was amplified from M. tuberculosis H37Rv chromosomal DNA by PCR using the oligonucleotides 5′-ggggacaagtttgtacaaaaaagcaggcttaAGCGCAATCCAGGGAAATGT-3′ and 5′-ggggaccactttgtacaagaaagctgggtcCTATGCGAACATCCCAGTGAC-3′, and inserted into the expression vector pDest17 in frame with the encoded 6-His tag downstream from the T7 promoter by recombination, as recommended by the supplier (Invitrogen), and sequenced. Protein purification was done from 6-liter Luria-Bertani broth cultures that were induced for 4 h by 1 mM isopropyl-β-d-thiogalactopyranoside at an OD600 of ∼0.5. First, we exploited that both proteins formed inclusion bodies in Escherichia coli. After centrifugation, the bacteria pellet was resuspended in a mild detergent solution (bacterial protein extraction reagents, B-PER; Pierce) to disrupt the outer membrane, and insoluble material was pelleted and washed twice with 100× diluted B-PER to remove soluble proteins. Pellets containing recombinant proteins were dissolved in 100 ml of 8 M urea, 50 mM Na2HPO4 (pH 8) (buffer A) and metal affinity purified under denaturing conditions by exploiting the His tag fused to the proteins. The dissolved proteins were bound to 20 ml of Streamline chelating column material (Amersham Biosciences) already chelated with Ni2+ and preequilibrated in buffer A and washed by alternating between 10 mM Tris-HCl (pH 8.0), 60% 2-propanol (buffer B) and buffer A. Each preparation was washed three times with 2 column vol of buffers A and B before being eluted in buffer A supplemented with 200 mM imidazole. Eluted fractions were analyzed by Coomassie blue staining of SDS-polyacrylamide gels, and the purest fractions were pooled. For anion exchange chromatography, the buffer was changed by dialysis against 25 mM Tris-HCl (pH 8.0), 6 M urea (buffer C) before the proteins were applied to Mono Q columns (GE Healthcare), washed with 5 column vol of buffer C, and eluted with a 0–1 M linear NaCl gradient made up in buffer C. All fractions were again analyzed for recombinant proteins, and the purest fractions were pooled and refolded by dialysis against 25 mM HEPES (pH 7.5), 10% glycerol, 0.15 M NaCl. Finally, the protein concentration was measured by a bicinchoninic acid test (Pierce) using BSA as a protein standard, and the LPS content was determined by a kinetic turbidimetric Limulus amebocyte lysate assay (Charles River Laboratories). No toxicity or nonspecific stimulation was found for any of the Ags on naive murine PBMC when they were tested at concentrations up to 10 μg/ml.
Overlapping peptides containing 15 aa covering the entire ESAT-6 molecule were synthesized (see Fig. 1 A), and the purity was verified by HPLC and the identity by mass spectrometry (JPT Peptide Technologies). Each peptide was purified to >80% purity before use. All impurities consist of shorter versions of the peptides caused by <100% coupling efficiency in each round of synthesis.
Immunizations and animal infections
CB6F1 F1 hybrid mice were immunized s.c. at the base of the tail three times, 2 wk apart with 5 μg of ESAT-6, 5 μg of Δ15-ESAT, or 25 μg of ESAT-6 single peptides. Ags were formulated in cationic adjuvant formulation, CAF01, and a total of 200 μl was injected per animal per immunization round. Negative control mice received three equivalent doses of CAF01, and positive control mice received a single dose of 5 × 104 CFU BCG Danish 1331 (Statens Serum Institut) in the first round of immunization. Ten weeks after the first immunization the animals were challenged with M. tuberculosis Erdman. Using a Glas-Col inhalation exposure system, virulent mycobacteria suspended in PBS Tween 20 (0.05%) were aerosolized and delivered via the respiratory route at ∼25–50 CFU per mouse. Depending on the experiment, six mice per group were sacrificed 4, 6, or 10 wk after infection, and lung homogenates were prepared in PBS Tween 80 (0.05%) and plated at 3-fold serial dilutions on Middlebrook 7H11 Bacto agar. After incubation at 37°C, CFU were enumerated 3 wk later.
Blood cells were obtained from immunized mice and pooled within the immunization groups (five to six animals per group) before red cells were removed by centrifugation over Lympholyte-Mammal (Cedarlane Laboratories). Splenocytes were isolated from individual animals by homogenizing spleens through a fine metal mesh. PBMCs and splenocytes were cultured in triplicate in microtiter wells (96-well Maxisorb; Nunc) or ELISPOT plates (BD Biosciences) containing 2 × 105 cells in 200 μl of RPMI 1640 supplemented with 0.5 μM 2-ME, 1% penicillin-streptomycin, 1 mM glutamine, 1% HEPES, 10% (w/v) inactivated FCS (all from Invitrogen), and 1% nonessential amino acids (MP Biochemicals). Culturing was done at 37°C/5% CO2 in the presence of medium alone or medium supplemented with Ag. Protein and peptide Ags were used at 1 and 2 μg/ml, respectively. The mitogen Con A (1 μg/ml) was used in all experiments as a positive control for cell viability. For IFN-γ ELISA, cell cultures were incubated for 72 h before the supernatants were harvested and analyzed for IFN-γ by a double-sandwich ELISA using specific mAbs (BD Pharmingen). For ELISPOT assay, the microtiter plates (ELISPOT plates; BD Biosciences) were coated with 2.5 μg/ml monoclonal hamster anti-murine IFN-γ (Genzyme) before culturing for 48 h. Cells were removed by washing, and the sites of cytokine secretion were detected by biotin-labeled rat anti-murine IFN-γ mAb (BD Pharmingen) and phosphatase-conjugated streptavidin (Jackson ImmunoResearch Laboratories) using 5-bromo-4-chloro-3-indolylphosphate (Sigma-Aldrich) as enzyme substrate. Spots were counted using a Synergy HT ELISPOT reader (Bio-Tek Instruments).
Cells isolated 6 wk after immunization were stimulated with 2 μg/ml ESAT-61–15 peptide for 1 h in the presence of 1 μg/ml anti-CD28 (clone 37.51) and anti-CD49d (clone 9C10(MFR4.B); both BD Pharmingen) and subsequently incubated for 5–6 h at 37°C after addition of 10 μg/ml brefeldin A (Sigma-Aldrich). Following overnight storage at 4°C, cells were washed in FACS buffer (PBS containing 0.1% sodium azide and 1% FCS) and subsequently stained 30 min at 4°C for surface markers with mAbs as indicated using 1/100 dilutions of anti-CD4-allophycocyanin-Cy7 (clone GK1.5), anti-CD8-PerCp-Cy5.5 (clone 53-6.7), and anti-CD44-FITC (clone IM7) (all BD Pharmingen). Cells were then washed in FACS buffer, permeabilized using the Cytofix/Cytoperm kit (BD Pharmingen) according to the manufacturer’s instructions, and stained intracellularly for 30 min at 4°C in dilutions of 1/100 using anti-IFN-γ-PE-Cy7 (clone XMG1.2; eBioscience), anti-TNF-α-PE (MP6-XT22; BD Pharmingen), or anti-IL-2-allophycocyanin (clone JES6-5h4; BD Pharmingen) mAbs. Cells were subsequently washed in FACS buffer, resuspended in formalin, and analyzed using a BD FACSCanto flow cytometer (BD Pharmingen) and FlowJo software V.7.2.5 (Tree Star).
The efficacies of the vaccines and immune responses to individual peptides were compared by one-way ANOVA of the log10 CFU followed by Tukey’s multiple comparison test of the log10 CFU means. Values of p of <0.05 were considered significant.
Infection with MTB in H-2b/d mice results in a stable T cell recognition pattern focused toward the immunodominant ESAT-61–15 CD4 epitope
It has previously been shown that the ESAT-6 epitope that is recognized by CD4+ T cells during the early phase of infection in C57BL/6 (H-2b) or BALB/C (H-2d) is found within aa 1–15 of the ESAT-6 sequence. We started by evaluating if F1 mice (C57BL/6 × BALB/C; referred to as CB6F1) would also exclusively recognize the epitope in the N terminus or if they in addition would recognize subdominant epitopes in other parts of the sequence. A series of 13 synthetic 15-aa-long overlapping peptides was synthesized based on the sequence of ESAT-6 (Fig. 1,A), and the recognition pattern throughout the course of a TB infection was investigated. CB6F1 mice were infected by the aerosol route and sampled five times during a period of 168 days. Splenocytes from three mice per time point were isolated and stimulated in vitro with the 13 ESAT-6 peptides individually, and T cell recognition was monitored by measuring IFN-γ in the supernatant (Fig. 1 B). At all four sampling points after infection only the peptide covering aa 1–15 in the ESAT-6 sequence (ESAT-61–15) was recognized, and this peptide triggered a significant release of IFN-γ by the splenocyte cultures. No recognition of this peptide was found preinfection. There was therefore a preferential targeting of the T cell repertoire toward the immunodominant epitope present in ESAT-61–15, and this pattern remained stable over the course of the infection.
Identification of subdominant epitopes in ESAT-6
We continued by analyzing the sequence of ESAT-6 for the potential presence of subdominant epitopes in addition to the immunodominant ESAT-61–15. Thirteen groups of CB6F1 mice were each immunized with one of the single peptides derived from ESAT-6 (Fig. 2) formulated in CAF01 adjuvant. Each mouse received three immunizations, and 3 wk after the last immunization splenocytes were harvested and restimulated in vitro with ESAT-6, the homologous ESAT-6-derived peptide, or a control peptide derived from another part of the ESAT-6 sequence. IFN-γ cytokine release assay revealed, as expected, that after immunization with P1 (ESAT-61–15) there was a clear recognition and a dose-dependent IFN-γ response to the homologous peptide, as well as a response to the full-length ESAT-6 protein (Fig. 2). No IFN-γ release was found in response to the control peptide (ESAT-621–35). In addition to the immunodominant P1 (ESAT-61–15) epitope, the analysis of peptides covering the full sequence of ESAT-6 revealed two additional subdominant epitopes to which T cells are not primed during the natural infection. Immunization with both P3 (ESAT-615–29) and P13 (ESAT-681–95) resulted in dose-dependent recall responses, demonstrating that it is possible to direct a specific response against other epitopes in the ESAT-6 molecule and that the ESAT-6 protein contains additional epitopes that can be processed and recognized by T cells. Interestingly, T cells directed to P3 (ESAT-615–29) did not recognize full-length ESAT-6 protein in vitro, indicating that this epitope is not naturally processed from the full-size ESAT-6 protein in vitro in the CB6F1 mouse strain. In contrast, the T cells specific for the epitope in the C-terminal P13 (ESAT-681–95) peptide responded as strongly as did the P1- (ESAT-61–15) specific T cells to the in vitro restimulation with full-length ESAT-6 protein (Fig. 2).
Exposing subdominant T cell epitopes through molecular redesign of ESAT-6
We next hypothesized that it would be possible to expose and redirect the T cell response to subdominant epitopes by engineering an ESAT-6-derived molecule in which the dominant ESAT-61–15 epitope was removed (Δ15ESAT-6). Groups of CB6F1 mice were immunized three times with 2-wk intervals with either full-length ESAT-6 or Δ15ESAT-6 formulated in adjuvant. One and 3 wk after the third immunization three mice per group were sacrificed and splenocytes were isolated. In vitro restimulation of these cells with ESAT-6 and Δ15ESAT-6 demonstrated a stronger recognition of the homologous ESAT-6 protein preparation but still a substantial recognition of the heterologous preparation, suggesting that the two proteins promote distinct but cross-reactive T cell repertoires. ESAT-6-immunized mice recognized primarily full-size ESAT-6, and we observed a 4-fold higher frequency of IFN-γ-producing T cells in response to the ESAT-6 protein (206 ± 36 SFU/106 splenocytes) compared with Δ15ESAT-6 (48 ± 3 SFU/106 splenocytes) (Fig. 3,A, top panel). In contrast, when animals were immunized with Δ15ESAT-6, the frequency of cells responding against full-length ESAT-6 was ∼3-fold lower (115 ± 14 SFU/106) than against Δ15ESAT-6 (320 ± 27 SFU/106) (Fig. 3,B, top panel). This indicates that even though full-size ESAT-6 contains all possible epitopes that are theoretically available in the Δ15ESAT-6-truncated molecule, the presence of the ESAT-61–15 immunodominant epitope may negatively influence the in vitro processing of potential remaining epitopes in the molecule. We therefore continued by comparing the T cell epitope pattern promoted by the two immunogens. Splenocytes isolated from immunized mice were stimulated in vitro with the complete set of single peptides spanning the sequence of ESAT-6 (Fig. 1,A). After immunization with the full-length ESAT-6 protein, the recognition pattern was focused almost exclusively toward the N-terminal ESAT-61–15 peptide (Fig. 3,A, lower panel), identical to the response found after an aerosol MTB infection (Fig. 1,B). In contrast, immunization with Δ15ESAT-6 refocused the response toward the subdominant epitope P13 (ESAT-681–95) identified above, but also exposed two subdominant epitopes, P6 (ESAT-633–47) and P9 (ESAT-655–69), that were not seen when immunizing with the individual peptides (Fig. 3 B, lower panel). The responses to the subdominant epitopes were weaker (800–2000 pg/ml released IFN-γ) than found for the T cell response directed to the dominant ESAT-61–15 epitope after ESAT-6 immunization (∼12,000 pg/ml IFN-γ) but were clearly significant, as these peptides gave <50 pg/ml in nonimmunized animals (data not shown). Therefore, deletion of the dominant epitope exposed three subdominant epitopes in ESAT-6.
Involvement of T cells directed to dominant and subdominant epitopes from ESAT-6 during infection
Whereas ESAT-6 immunization promotes a response where the epitope recognition pattern is indistinguishable from the natural infection, immunization with Δ15ESAT-6 stimulated a response that is clearly different. We continued by investigating the development of the fine specificity of the ESAT-6-specific T cell repertoire during MTB infection in animals primed by ESAT-6, Δ15ESAT-6, and in saline control animals. The epitope recognition pattern was measured by IFN-γ release in splenocyte cultures restimulated in vitro with the panel of overlapping peptides (Fig. 4,A) and by a detailed analysis of cells recruited to the lung at different time points after infection (Fig. 4, B and C). The recognition pattern was analyzed on the day of challenge and 2 and 12 wk postinfection. Saline control animals did not recognize any of the ESAT-6 peptides before infection. In agreement with what we had observed earlier (Fig. 1,B), infection with MTB promoted a strong ESAT-6 response focused to P1 (ESAT-61–15), with significantly increased IFN-γ release (up to 4000 pg/ml) from week 2 and onward (p < 0.05) (Fig. 4,A). ESAT-6-immunized animals recognized P1 throughout infection, and no significant increase was found for the P1 response postinfection. Compared with the infected control animals, the P1 epitope was recognized already at the time of challenge in the ESAT-6-immunized mice (Fig. 4 A). The Δ15ESAT-6 immunized group recognized as expected the three subdominant epitopes P6, P9, and P13 after immunization, and this broad epitope recognition pattern remained constant at weeks 2 and 12 postinfection with a significant increase in responses only detectable to P9 (2 wk postinfection; p < 0.05). Interestingly, the strong infection-driven P1 response that was seen in the saline control animals was not seen in the Δ15ESAT-6 immunized group (p < 0.05 for 0 and 2 wk postinfection), indicating efficient control of bacterial replication as has been reported before (21). The use of P1 as an infection marker could obviously not be extended to the ESAT-6-immunized animals where P1 responses are promoted by the vaccine itself. Only at the late time point was the immunization-promoted response to the subdominant epitopes (which at this time point had waned to 500–800 pg/ml) supplemented by an infection-driven P1 epitope response at a level 5-fold below the P1 level in the saline control group. Thus, at this time point Δ15ESAT-6-immunized animals recognized all four epitopes at almost identical levels.
To further investigate the direct involvement of T cells specific for the subdominant epitopes at the site of infection in the lung, we investigated the T cells recruited to the lung from 0 to 12 wk postinfection by multiparameter flow cytometry. We perfused lungs to remove residual blood and thereby restrict our analysis to T cells that were actively infiltrating the lung tissue during TB infection. The animals received BrdU in their drinking water for 3 days before necropsy to allow the distinction between recruitment of nonproliferating effector T cells and proliferating memory T cells that had expanded after contact with MTB. The specificity of the T cells recruited was analyzed by peptide restimulation and intracellular-FACS for IFN-γ, TNF-α, and IL-2. To calculate the frequency of CD4+ T cells specific for each of the epitopes within the dividing or nondividing T cell population, CD4+ T cells were first divided into BrdU-positive and BrdU-negative and then analyzed for the expression of the three cytokines (Fig. 4,B). Cells positive for any of the cytokines singly or in combination were counted as positive (Fig. 4,C). At the day of infection almost no ESAT-6-specific cytokine-producing T cells could be detected in the lungs, and no clear recognition of the peptides was found. This was clearly different at week 2, where 1.8% of the CD4+ T cells in the lungs of ESAT-6-immunized animals recognized the ESAT-6 molecule and the epitope contained in the P1 (ESAT-61–15) peptide and produced at least one of the cytokines measured (Fig. 4,C). In the Δ15ESAT-6-immunized animals the CD4+ T cell response was focused toward the three subdominant epitopes contained in P6, P9, and P13, and, in particular, P6 and P9 were recognized at a level (1.2%) that was comparable to the P1 response in the ESAT-6-immunized animals (1.8%). Overall, about a third of the peptide-specific CD4+ T cells isolated from the lung at week 2 had incorporated BrdU, indicating that these cells had proliferated at the site of infection in vivo. The remaining two-thirds of the ESAT-6-specific CD4+ T cells were BrdU-negative, suggesting that these cells either recently had been recruited to the lung as fully differentiated effector cells and had not proliferated following contact with the pathogen or they had been recruited and proliferated before BrdU incorporation. At the late time point investigated (week 12) the epitope recognition pattern for the ESAT-6-immunized animals was unaffected and the focus was still on the P1 peptide, but only a very low fraction of the specific CD4+ T cells had divided and incorporated BrdU. In the Δ15ESAT-6-immunized animals we found the same tendency as we had observed at this time point in the spleen with low but detectable responses to ESAT-6 and minimal responses to the peptides (Fig. 4 C).
Vaccine-promoted polyfunctional ESAT-6-specific CD4+ T cells
Given that P1 (ESAT-61–15) was the immunodominant epitope during infection in both nonimmunized and ESAT-6-immunized animals, it enabled a direct comparison of a vaccine-promoted and an infection-driven response to a single epitope. Using multiparameter flow cytometry, we compared the quality of the immune response to this epitope as it was recruited and developed during a long-term infection in the lung (2–24 wk). CD4+ T cells were analyzed for their expression of IFN-γ, TNF-α, and IL-2, allowing us to follow the frequency of cells that produced various combinations of IFN-γ, TNF-α, and IL-2 being positive for either a single or up to three cytokines simultaneously (Fig. 5,A). As expected, we detected P1 (ESAT-61–15) responses in ESAT-6-immunized animals 2 wk postchallenge, whereas this response was delayed in the mock-immunized animals (Fig. 5,A). The P1-specific cells at week 2 were equally divided into IL-2-expressing cells with a memory potential (either IFN-γ, TNF-α, and IL-2 triple-positive or TNF-α and IL-2-positive) and effector cells (mainly TNF-α single-positive or IFN-γ and TNF-α double-positive). At week 6 we found a substantial population of cytokine-positive CD4+ T cells specific for P1 in both groups, but the numbers of P1-specific cells were higher in the immunized group (4.4 vs 1.7%). At 24 wk postchallenge the frequency of P1-responding T cells was almost identical in the two groups (2.45% vs 2.4%). In the ESAT-6-immunized group the cytokine expression pattern of the P1-specific cells was stably maintained throughout infection, with the main subsets being either triple-positive or TNF-α and IL-2 double-positive. Even at the late time point (week 24) the response was characterized by >60% of responding cells belonging to these subsets. This was in contrast to the nonimmunized group that was characterized by an effector population where >80% of the CD4+ T cells that recognized P1 produced IFN-γ and/or TNF-α, but not IL-2 (Fig. 5 B). The IFN-γ mean fluorescence intensity (MFI) of the IFN-γ, TNF-α, and IL-2 triple-positive CD4+ T cells in the ESAT-6-immunized group was 4.5-fold higher (MFI, 8955 ± 687) than the single-positive effector cells (MFI, 2115 ± 61), in agreement with earlier demonstrations (22, 23) of the superior cytokine production potential of this subset. A similar quality was found for CD4+ T cells directed to the subdominant epitope responses promoted by Δ15ESAT-6, although the frequency was much lower than for the immunodominant epitope (results not shown). Thus, immunization not only accelerated the development of an immune response but also increased the response quality by inducing polyfunctional CD4+ T cells that were maintained throughout infection.
Subdominant T cell responses promoted by Δ15ESAT-6 efficiently protect against TB
Given the vaccine-induced increase in immune response quality and accelerated response in the lung after infection, we compared the vaccine efficacy of ESAT-6 and Δ15ESAT-6 against an aerosol challenge with virulent MTB. Mice were immunized with ESAT-6/CAF01, Δ15ESAT-6/CAF01, or BCG. The BCG vaccine (5 × 104 CFU) was injected s.c. at the base of the tail at the time of the first subunit immunization. All groups of mice received an aerosol challenge delivering ∼50 CFU into the lung of each animal 10 wk after the first immunization. First, we performed three individual experiments that were terminated 6 wk postinfection, the normal necropsy time point in the aerosol model. In these experiments the ESAT-6-based vaccine reduced the lung CFU between 0.37 and 0.50 log10, whereas Δ15ESAT-6 reduced the CFU between 0.66 and 0.84 log10. In comparison, BCG reduced the CFU between 0.81 and 1.03 log10 (Fig. 6,A). In the experiments where a saline control group was included, the bacterial load in this group was identical to the level found in the adjuvant control group (data not shown), in agreement with earlier observations (24). Statistical analysis on the pooled data from the three experiments shows that Δ15ESAT-6 gave significantly better protection than did ESAT-6 (p < 0.01). To further compare the activity of the two ESAT-6 vaccine constructs at an early time point and late time points after infection, we compared bacterial numbers in the lung isolated at weeks 4 and 10 postinfection (Fig. 6, B and C). Whereas both vaccines reduced bacterial load at the early time point (4 wk postinfection), only the Δ15ESAT-6-based vaccine gave significant levels of protection (0.9 log10 reduction of CFU) 10 wk postinfection (p < 0.05).
The present study demonstrate that it is possible to expose subdominant epitopes in a vaccine Ag such as ESAT-6 and that the T cells directed to these epitopes protect against an infection with MTB. Based on three individual experiments, the consistent finding was improved vaccine efficacy of the subdominant T cell repertoire promoted by Δ15ESAT-6 immunization, relative to full-size ESAT-6. Immunization with the intact ESAT-6 molecule focuses the immune recognition on the N-terminal epitope ESAT-61–15, and it therefore promotes a response with the same fine specificity as the response that evolves during infection. Our data therefore suggest that for a chronic bacterial disease like TB, it is possible through vaccine design to induce a more protective immune response than the response naturally occurring during infection. This has great implications for our understanding of two important enigmas in TB immunology: (1) the high rate of reinfection in patients previously cured for TB where the strong memory immunity from the primary infection would be expected to prevent a reinfection with MTB (3); and (2) the reactivation of latent TB in millions of people every year, which makes it a major source of new TB cases globally, often in the face of a very strong cell-mediated immunity response that is continuously maintained by the latent infection. In this regard, it is very interesting that we observed the same level of vaccine efficacy in the first phase of infection by ESAT-6 and the truncated Δ15ESAT-6 molecule, whereas Δ15ESAT-6 protected more efficiently during the late stage of infection where animals immunized with this molecule had T cells to a broad epitope repertoire that recognized both the subdominant epitopes promoted by the vaccine and the immunodominant epitope promoted by the infection. This was clearly different from the ESAT-6-immunized animals that had a response directed only to the immunodominant ESAT-61–15 epitope. Our study suggests that subdominant imprinting results in an overall change and expansion of the T cell repertoire and prevents the immunofocusing normally seen during TB infection in animal models.
The vaccine potential of subdominant epitopes has so far primarily attracted interest in viral infections. For viral pathogens such as HIV with highly variable genomes, the focus on subdominant epitopes has been driven primarily by the need to avoid escape mutants by focusing on epitopes with no selective pressure instead of immunodominant decoy epitopes (15, 25). Supporting the relevance of this approach, there are several studies demonstrating that SIV-infected rhesus monkeys with a strong natural ability to control infection are characterized by a very broad T cell repertoire that also encompasses subdominant epitopes (14, 26). However, MTB has a relatively slow replication rate and limited variability within clinical strains (27), and although recent information from using modern molecular biology techniques have identified numerous single nucleotide polymorphisms, their impact on variation within the sequence of immunodominant Ags such as ESAT-6 family members has been very limited and these Ags are highly conserved among clinical isolates (28). Despite a strong host immune pressure against ESAT-6, escape mutants are unlikely as a survival strategy for MTB, not only because the general mutation rate is much lower in mycobacteria than in virus, but because ESAT-6 is an essential gene for MTB survival and replication in the host, and even single amino acid substitutions have been shown to be detrimental for the survival of MTB in vivo (29). Therefore, in TB the advantage of the subdominant repertoire may relate primarily to the quantitative advantage with more T cell specificities and a more diverse T cell repertoire with the potential to recognize the infected macrophage.
In the present study we exposed the subdominant repertoire by removing the immunodominant epitope ESAT-61–15 from the ESAT-6 molecule. Engineering of vaccine Ags by removing immunodominant epitopes has also been evaluated in other infection models with variable outcomes. In agreement with our observations some studies report that deletion of an epitope results in the exposure of subdominant epitopes in, for example, Staphylococcus aureus nuclease (25, 30). In contrast, even single amino acid substitutions in human IFN-β were enough to eliminate in vivo responses to both the immunodominant and subdominant epitope within the IFN-β molecule (31). The explanation for this latter observation is not clear, but it was suggested that for some subdominant epitopes the efficient priming of responses depends on the establishment of a productive proliferative response to the immunodominant epitope (31).
Immunodominance is a functional description, and the molecular factors that control dominance and subdominance are still not fully understood. In mice it has been shown that the ESAT-61–15 epitope was recognized by as much as 25–35% of the total mycobacteria-reactive T cell repertoire recruited to the site of infection in the first phase of a recall response in C57BL/6J mice (9). The immunodominance of this epitope may be a consequence of the high expression of ESAT-6 during infection (32), as well as easy proteolytic access to the protruding part of the ESAT-6 molecule containing this epitope (33). However, it is known from viral studies that epitope dominance cannot only be explained in terms of peptide loading and affinity for MHC (34), and there are data to suggest that T cells responding to one Ag can actively interfere with T cells responding to another (35, 36). T cell competition for epitopes has therefore been suggested as another way to explain why subdominant epitopes appear to be inhibited by responses to other, more powerful Ags (37, 38). Importantly, it has been demonstrated that repeated immunizations or pathogen exposure increases the average affinity of the TCR for the peptide/MHC complex (39, 40, 41). This process appears to be due to a preferential outgrowth of the higher affinity T cells present within the pool of primary responders, suggesting that the higher affinity cells have a competitive advantage as they more efficiently maintain an interaction with Ag-bearing APCs, and the consequence is a narrowing of the TCR repertoire during a recall response (39). However, as demonstrated by our data, if primed by immunization, T cell populations specific for the subdominant epitopes are recruited to the lung during infection and are not diluted out by the preferential outgrowth of T cells directed to the dominant epitopes. This results in a balanced T cell repertoire that is stably maintained in the organs over time without decreasing diversity. This resembled the observations in SIV-infected monkeys where vectored vaccines generated responses to subdominant epitopes that were expanded following exposure to virus and not overtaken by responses to the dominant epitopes (42).
Recently, a number of reports dealing with various disease targets in different animal models have demonstrated a correlation between protection and high-quality T cells that coexpress multiple cytokines (22, 43). ESAT-6 immunization provides an ∼10-fold reduction of bacterial numbers, but other than an accelerated response from ESAT-61–15-specific CD4+ cells, the ESAT-6-immunized and nonimmunized animals have identical fine specificity of their response to ESAT-6, and 12 wk into the infection they also exhibit similar levels of IFN-γ recall responses to the peptide. However, the infection-driven ESAT-61–15-specific CD4+ T cell response in nonimmunized animals consists predominantly of effector T cells producing TNF-α and IFN-γ, whereas the vaccine-promoted ESAT-61–15-specific response includes both double- (TNF-α+, IL-2+) and triple-positive T cells expressing IL-2 at various time points throughout infection, including the late time point 24 wk postimmunization. Vaccine-promoted maintenance of polyfunctional CD4+ T cells was associated with a 4- to 5-fold more efficient production of cytokines on a per cell basis than the single-positive CD4+ T cells, in agreement with a recent characterization of this subset (22, 44). The exceedingly efficient maintenance of a high-quality T cell response even during an ongoing infection is most likely the reason for the immunity expressed in these animals and may relate to the CAF01 delivery system used to adjuvant the vaccines, as recently demonstrated after immunization with the Ag85B-ESAT-6 fusion molecule in the same liposome formulation (44).
Although detailed studies of the epitope potential of ESAT-6 have demonstrated numerous epitopes for human T cell recognition (7), the human T cell response during TB infection often ends up being focused on a very limited number of T cell epitopes (8). Our data suggest that by employing immunization strategies that refocus the immune response to include subdominant epitopes, it will be possible to achieve ESAT-6-based vaccines with higher protective efficacies. Most likely such a concept can be expanded to other Ags and chronic diseases. For human use where the T cell epitopes may vary significantly depending on the host HLA haplotype, it would be ideal to expose all epitopes, subdominant and dominant, without the need for identifying and modifying the dominant epitope. Currently, we are trying to accomplish this by synthetic peptide-based immunizations.
We gratefully acknowledge the assistance of Vivi Andersen, Charlotte Fjordager, Lene Rasmussen, and Kristine Persson for their excellent technical help in this study.
The authors have no financial conflicts of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Tuberculosis Vaccine Cluster-European Commission (TBVac-EC) Grant CT2003-503367.
Abbreviations used in this paper: MTB, Mycobacterium tuberculosis; BCG, bacillus Calmette-Guérin; CAF, cationic adjuvant formulation; ESAT, early secretory antigenic target; MFI, mean fluorescence intensity.