Abstract
In the present study we investigated the shaping and evolution of the immunodominance of the T cell response during a chronic mycobacterial infection. Using a recombinant bacille Calmette-Guérin expressing a reporter Ag, the Escherichia coli MalE protein, we analyzed the peptide specificity and the cytokine profile of the T cell response to the reporter Ag by ELISPOT. During the early steps of infection, the T cell response was focused on two dominant MalE epitopes and was characterized by a pure IFN-γ response. Then, in the course of infection the initial IFN-γ response to these two epitopes shifted to a mixed IFN-γ/IL-4 response. At the same time, the peptide specificity of the T cell response was broadened to two additional MalE epitopes characterized by a unique IL-4 response resulting in the establishment of a dominant IL-4 response to the MalE protein at 16 wk postinfection. However, this phenomenon did not impair the outcome of a predominant IFN-γ response upon subsequent MalE recall in vivo performed in the presence of CFA, a Th1-driving adjuvant. These results indicate that the Th2 nature of the immune response established during a chronic infection, which most likely reflects regulatory mechanisms to allow the return to T cell homeostasis, does not shape the Th1/Th2 nature of the memory response.
Interleukin-12 plays a major role in the control of mycobacterial infection (1, 2) by inducing the production of IFN-γ, which regulates both innate and acquired immune responses by activating NK cells and promoting Th1 differentiation (3). In this regard, Th1 CD4+ T cells, by producing IFN-γ, play an important role in mycobacterial immunity, since this cytokine stimulates macrophages to produce reactive nitrogen (4). Mycobacterial products can have opposite effects, since lipoproteins can stimulate the production of IL-12 by macrophages via Toll-like receptor 2 (5), whereas mannosylated lipoarabinomannans have been shown to inhibit IL-12 production by human dendritic cells (6). However, mycobacterial infections induce IL-12 production by dendritic cells in vitro (7, 8) and in vivo (9).
Presumably due to their capacity to stimulate IL-12 production by dendritic cells, mycobacteria are potent activator of Th1 CD4+ T cell responses. However, in recent years, several studies reported the occurrence of Th2 cytokine responses in human tuberculosis (10, 11, 12, 13) even though these observations remain controversial (14, 15). In mice it has been established that the mycobacterial dose defines the Th1/Th2 nature of the immune response independently of the route of immunization (16). Indeed, relatively low doses of bacille Calmette-Guérin (BCG)3 lead to an almost exclusive Th1 response, whereas higher doses induced mixed Th1/Th2 responses (16). By analyzing CD4+ T cell responses at various times during the course of a virulent Mycobacterium tuberculosis infection, Orme et al. (17) demonstrated the existence of two separate waves of cytokines. These responses were characterized by an early Th1 response, specific for secreted proteins of the bacilli, followed by a Th2 response, in which the mycobacterial 60-kDa heat shock protein appeared to be a strong target (17). In a recent study Rhodes et al. (18) also demonstrated distinct kinetics of IFN-γ and IL-4 responses in bovine tuberculosis. Indeed, an early IFN-γ production was observed in cattle infected with Mycobacterium bovis that was maintained throughout the period studied. In contrast, the IL-4 response was delayed and confined to a peak of activity lasting 6–8 wk (18).
From these studies it is difficult to conclude whether these different kinetics of Th1 and Th2 responses are related to the fine specificity of antibacterial immune responses or to the histopathology of the infection. Indeed, a correlation was previously established between the kinetics of Th1/Th2 cells and pathology in a murine model of experimental pulmonary tuberculosis (19).
Therefore, in the present study we analyzed the kinetics of Th1 and Th2 responses directed against a defined mycobacterial Ag as well as against its epitopic determinants during the course of BCG infection using a recombinant BCG expressing a reporter Ag, the Escherichia coli MalE protein (9, 20). The cytokine response specific for MalE was analyzed as well as the specific T cell responses to defined MHC class II-restricted MalE epitopes. These responses were monitored using the sensitive and quantitative ELISPOT technique to precisely follow the fate of the Th1/Th2 immunodominance of the T cell response and its fine specificity.
Materials and Methods
Mice
Six- to 10-wk-old female BALB/c (H-2d) mice from Janvier (Le Genest Saint-Isle, France) were used.
Peptides and Ags
Based on the sequence of MalE (21) a complete set of 385 overlapping 15-mer peptides was synthesized on polyethylene pins according to standard PEPSCAN procedures (22). In addition to this milligram scale synthesis, selected regions of the MalE sequence were synthesized at 20 mg scale according to standard synthesis procedures for peptides using Wang resin (p-alkoxybenzylalcohol resin; Bachem, Bubendorf, Switzerland). All peptides were provided by Dr. J. Langeveld (Institute for Animal Science and Health, Lelystadt, The Netherlands). The MalE protein was donated by Dr. J. M. Clement (Institut Pasteur, Paris, France). Purified protein derivative (PPD) was supplied by Seruminstitut (Copenhagen, Denmark).
Bacterial strains and culture conditions
The M. bovis BCG Pasteur vaccine strain 1173 P2 was grown in Middlebrook 7H9 (Difco, Paris, France) medium supplemented with 0.1% Tween 80 and albumin-dextrose-catalase (Difco) or in Middlebrook 7H10 (Difco) solid medium supplemented with oleic acid albumin-dextrose-catalase (Difco). The construction and characterization of rBCG.MalE expressing the E. coli MalE gene (previously described as rBCG(pBlaF*-SSBlaF-SSmalE-malE) have been reported previously (20). For immunization, BCG strains were first cultivated on Loewenstein-Jensen medium containing 20 μg/ml kanamycin, and then transferred onto Sauton medium. After immunization, the growth of BCG strains was followed by the numeration of BCG or rBCG CFU in the spleen at various intervals. Suitable dilutions of spleen cells were plated on Sauton medium with or without kanamycin (10 μg/ml) to determine the numbers of CFU of BCG and rBCG.MalE, respectively.
ELISPOT assay
Mice were immunized i.v. with 106 rBCG.MalE or BCG cells. At different time points, spleen cells from individual mice were recovered and serially diluted in RPMI 1640 medium (Life Technologies, Paisley, Scotland) supplemented with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, and 5 × 10−5 M 2-ME). They were then plated onto ester-cellulose-bottomed plates (Millipore, Moishem, France) coated with a capture mAb specific for mouse IFN-γ (R4-6A2), IL-2 (JES6-1A12), or IL-4 (BVD4-1D11; BD PharMingen, San Diego, CA) at 4 μg/ml in PBS. Cells were incubated at 37°C for 24–36 h with Ag in duplicate. Then cells were removed by washing twice with H2O-0.05% Tween 20 and five times with PBS-0.05% Tween 20 (PBST), and then biotinylated anti-mouse IFN-γ (XMG1.2), IL-2 (JES6-5H4), or IL-4 (BVD6-24G2; BD PharMingen) was added at 4 μg/ml in PBST plus BSA 1% and incubated for 3 h at room temperature. After extensive washing in PBST, streptavidin-alkaline phosphatase (BD PharMingen) was added for 3 h at room temperature. After washing with PBST and PBS without Tween, 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium substrate (Sigma-Aldrich, St. Louis, MO) was added. The number of spots for each cell dilution was counted, and the results for each antigenic stimulation correspond to the mean of duplicates, except for Fig. 1⇓A, which corresponds to single determinations. Results are expressed as the number of spots per 106 spleen cells or as the number of cytokine-secreting cells per spleen.
Epitope mapping of the IFN-γ response against the MalE protein following immunization with rBCG. MalE. A, BALB/c mice were i.v. immunized with 106 CFU of rBCG.MalE or BCG.wt. Four weeks later, spleen cells were harvested, and the number of IFN-γ-secreting T cells was determined after in vitro stimulation of splenocytes with MalE protein (2.5 μg/ml), PPD (10 μg/ml), or MalE peptides (2 μM) using ELISPOT assay. Results are expressed as the number of IFN-γ-secreting cells per 106 splenocytes obtained with an individual mouse and are representative of two experiments. B, Two BALB/c mice were i.v. immunized with 106 CFU of rBCG. MalE. Four weeks later, spleen cells were harvested, and the number of IFN-γ-secreting T cells was determined after in vitro stimulation of splenocytes with p68–82 or p277–291 peptides or MalE protein. Splenocytes from individual mice were either left untreated or were depleted of CD4+ or CD8+ cells before stimulation. The results are the means of the number of specific IFN-γ-secreting cells per million splenocytes obtained for two individual mice. C, Mice were immunized with MalE (five mice) in CFA, and 10 days later the proliferative responses of pooled draining lymph node cells were measured by [3H]thymidine incorporation on day 4 after in vitro stimulation with p68–82, p100–114, p151–165, p277–291, or p40–54 MalE peptides. Results are expressed as mean counts per minute of duplicate determinations and are representative of four experiments.
For the depletion of CD4+ or CD8+ cells, anti-CD4 mAb (GK1-5) or anti-CD8 mAb (H35-17-2) was used together with Dynabeads coated with sheep anti-rat Ig (Dynal, Oslo, Norway). The depletion was verified by staining cells with anti-CD4 and anti-CD8 mAbs (BD PharMingen). FACS analysis was performed using a FACScan (BD Biosciences, Mountain View, CA) and CellQuest software (BD Biosciences).
Generation of MalE-specific T cell hybridomas
T cell hybridomas were produced as previously described (22). BALB/c mice were immunized s.c. with 10 μg of MalE in CFA (Sigma-Aldrich). Seven days later, lymph nodes were removed, and a single-cell suspension was prepared and cultured in complete medium with 10 μg/ml of MalE. Three days later, viable lymphocytes were isolated and fused with the hypoxanthine, aminopterin, thymidine (HAT)-sensitive BW5147 (α−β−) thymoma (provided by Dr. J.-G. Guillet, Institut Cochin de Gènètique Molèculaire, Paris, France) using 50% polyethylene glycol 1500 (Roche, Mannheim, Germany). Then, the cell suspension in CM supplemented with 20% FCS was plated in the presence of feeder HAT-sensitive A20 cells. Sixteen hours later, HAT (Roche) was added to each well. Hybridomas were screened for MalE and peptide reactivity. All hybridomas were I-Ad restricted, among which FBU.B11 is specific for immunodominant p68–82 peptide, whereas FBC.D1, FBH.B7, and FBQ.E3 are specific for subdominant p100–114, p151–165, and p277–291 peptides, respectively.
Ex vivo Ag presentation assay
BALB/c mice were i.v. injected with 108 live rBCG.MalE or BCG cells or 100 μg of purified MalE protein in PBS. Spleens were removed and perfused with 1 ml of collagenase type IV (400 U/ml in RPMI 1640; Roche) containing 50 μg/ml of DNase I (Roche). Spleens were cut into small pieces, digested in 5 ml of collagenase type IV containing DNase I for 20 min at 37°C, and further dissociated in Ca2+-free PBS in the presence of 2 mM EDTA. Low buoyant density APC were prepared by centrifugation over a dense BSA (Sigma-Aldrich) gradient of spleen cell suspensions depleted of T cells with anti-Thy1.2 mAb (TIB 99; American Type Culture Collection, Manassas, VA) and complement. The low density fraction, enriched in dendritic cells (∼10% of the cell fraction), was directly used as APC. T cell hybridomas (105) were added to those APC for 24 h, then supernatants were tested for IL-2 content using the IL-2-dependent CTLL cell line. Results are expressed as counts per minute.
Results
Epitope mapping of MalE-specific T cell responses induced by rBCG.MalE
We first analyzed the MalE-specific T cell repertoire stimulated in vivo following immunization with rBCG.MalE using epitope mapping. Since BCG is known to induce a vigorous Th1 response, the T cell response was analyzed using an ELISPOT assay for IFN-γ. BALB/c mice were i.v. immunized with rBCG.MalE or BCG.wt, then 4 wk later, immune spleen cells were restimulated in vitro with a set of 385 overlapping peptides covering the entire MalE amino acid sequence (tested as pools of 12 overlapping peptides). As shown in Fig. 1⇑A, IFN-γ-secreting cells were only detected against peptides corresponding to aa sequences 59–84, 71–96, and 275–300 of the MalE protein. The fine mapping with individual peptides revealed that these positive responses correspond to epitopes p68–82 and p277–291 (data not shown). When mice were immunized with BCG.wt, a PPD-specific IFN-γ response was observed, whereas, as expected, MalE peptides and MalE protein did not stimulate any IFN-γ production by spleen cells.
To further determine whether IFN-γ produced by splenocytes from rBCG.MalE-immunized mice was due to CD4+ or CD8+ T cell stimulation, a selective depletion of CD4+ or CD8+ cells was performed before the ELISPOT assay. As shown in Fig. 1⇑B, CD4+ T cell depletion fully abrogated the IFN-γ response to MalE or p68–82 and p277–291 peptides, whereas the number of IFN-γ-secreting cells remained unaffected after removing CD8+ cells. Identical results were obtained when measuring MalE-specific IL-2 spots (data not shown). Together these results clearly indicate that the IFN-γ-producing CD4+ T cells primed by i.v. administration of rBCG.MalE to BALB/c mice were specific for two MHC class II-restricted MalE peptides. In contrast, when the proliferative T cell responses were measured following immunization of BALB/c mice with the purified MalE protein in CFA, we identified four responding peptides: p68–82, p100–114, p151–165, and p277–291 (Fig. 1⇑C), which are all restricted by MHC class II I-Ad molecules (our unpublished observations).
Ex vivo detection of MHC class II-MalE peptides complexes following rBCG.MalE immunization
As no IFN-γ response against p100–114 and p151–165 epitopes was detected following rBCG.MalE immunization, we next assessed whether Ag delivery by M. bovis BCG was able to qualitatively modulate MHC class II-restricted presentation of MalE epitopes. To study Ag presentation of the different MalE epitopes, T cell hybridomas specific for p68–82, p100–114, p151–165, and p277–291 peptides, named FBU.B11, FBC.D1, FBH.B7, and FBQ.E3, respectively, were produced as described in Materials and Methods. To determine whether the full pattern of Ag presentation of MalE epitopes is displayed in vivo, we analyzed ex vivo the Ag-presenting activity of splenocytes from mice immunized with rBCG.MalE. Spleens from mice i.v. injected with rBCG.MalE, BCG, or purified MalE were recovered 4 h after injection to prepare APCs enriched in dendritic cells. Then the formation of MalE epitope-class II complexes on these APC was monitored ex vivo by the stimulation of the four MalE-specific CD4+ T hybridomas. Fig. 2⇓ shows that T cell hybridomas specific for the p68–82 and p277–291 peptides were strongly stimulated by APC recovered from rBCG.MalE-immunized mice. In contrast, in these conditions the T cell hybridoma specific for p100–114 was only poorly stimulated, and no stimulation of FBH-B7, specific for the p151–165 peptide, was observed. The same pattern of T cell hybridoma stimulation was observed when using spleen APC from mice injected with purified MalE, although the responses were usually higher. APC prepared from BCG-injected mice were unable to stimulate any of the T hybridomas, confirming that the stimulation was due to the specific recognition of MalE epitope-class II complexes. We previously reported for the p68–82 peptide that ex vivo presentation by rBCG.MalE-infected APC is not detectable after 48 h (9). Therefore, it was not possible to follow Ag presentation ex vivo during the chronic phase of infection.
Ex vivo detection of MalE peptides presented by spleen APC in vivo following rBCG.MalE infection. Dendritic cell-enriched fractions were prepared from BALB/c mice (three or four mice per group) i.v. injected 4 h before with 108 CFU of rBCG.MalE or BCG.wt or with 100 μg of MalE. They were then cocultured with the T cell hybridomas FBU.B11, FBC.D1, FBH.B7, or FBQ.E3 specific for the MalE p68–82, p100–114, p151–165, and p277–291 peptides, respectively. IL-2 release by these hybridomas was measured in a CTLL proliferation assay. Results are expressed as counts per minute (mean of duplicates) and are representative of two experiments.
Together, these results indicate that the lack of detectable activation of IFN-γ T cells specific for p151–165 and p100–114 in vivo following rBCG.MalE might be due to a lack of appropriate threshold of activation for these peptides due to their poor presentation.
Kinetics of T cell responses to MalE protein and PPD after rBCG.MalE immunization
We next performed a kinetic study of T cell responses to the MalE protein following i.v. infection by rBCG.MalE to monitor the Th1/Th2 phenotype during the course of infection. At 1, 2, 3, 4, 8, and 16 wk postinfection, the IFN-γ, IL-2, and IL-4 responses to MalE protein were measured by ELISPOT. As shown in Fig. 3⇓A, the number of IFN-γ-secreting cells specific for the MalE protein rapidly increased, reached a peak at 3 and 4 wk, and then slightly decreased over time. During the course of infection, the number of cells secreting IL-2 in response to MalE remained quite stable during the infection, with a peak at 8 wk. Interestingly, MalE-specific, IL-4-secreting cells were detected only at 8 wk postinfection, and their number was comparable to the number of IFN-γ-secreting cells at 16 wk postinfection. The specificity of these responses was assessed by the lack of MalE-specific secreting cells in splenocytes of mice that received BCG.wt (Fig. 3⇓B). The occurrence of a mixed Th1/Th2 cytokine profile during the course of infection was further confirmed by analyzing the response to PPD, which was also characterized by an IFN-γ response at 4 wk and then by a mixed IFN-γ and IL-4 response at 10 wk (Fig. 3⇓C). The same cytokine profile was observed in response to PPD with spleen cells from mice immunized with BCG.wt (data not shown). At 4 and 10 wk, a similar number of bacteria was detectable in the spleen of mice infected with rBCG.MalE (∼5 × 104 CFU/spleen). However, only 10% of mycobacteria were still expressing the plasmid at 4 wk, and 1–2% were still expressing it at 10 wk following rBCG.MalE injection (Fig. 3⇓D). Together, these data show that the initial MalE-specific Th1 response induced by BCG shifted to a mixed Th1 and Th2 response during the course of infection. Interestingly, this phenomenon does not seem to be directly related to the amount of Ag available during the infection, as it was observed for the mycobacterial PPD Ag as well as for the transiently expressed MalE reporter Ag.
Analysis of the cytokines produced by Ag-specific T cells during the course of BCG infection. BALB/c mice (two per group) were i.v. immunized with 106 CFU of rBCG.MalE (A) or BCG.wt (B). At different time points, the MalE-specific T cells secreting IFN-γ, IL-2, or IL-4 were measured from individual spleen cells using an ELISPOT assay. C and D, BALB/c mice (two per group) were i.v. injected with 106 CFU of rBCG.MalE. At 4 and 10 wk after immunization, the PPD-specific T cells secreting IFN-γ or IL-4 were detected using an ELISPOT assay (C). At the same time, rBCG or BCG CFU were numerated from these spleen cells using Sauton medium with or without kanamycin (D). For ELISPOT (A–C) the results are expressed as the mean number of Ag-specific secreting cells per spleen. B, All results correspond to the limit of detection of the ELISPOT assay and vary depending on the number of total spleen cells.
Analysis of Th1/Th2 cytokines produced by splenocytes from rBCG.MalE-immunized mice in response to MalE peptides
The level of peptide presentation by MHC class II molecules, which stems in part from the intrinsic binding properties of peptides for a given MHC, was previously shown to influence the outcome of CD4+ Th1 or Th2 subset differentiation (23, 24). Therefore, it could be suggested that in response to BCG infection, Th1 and Th2 cytokines are produced by T cells specific for different epitopes. Thus, following infection of BALB/c mice by rBCG.MalE, we analyzed the cytokines produced in response to the four MalE epitopes (p68–82, p100–114, p151–165, and p277–291). The development of MalE-specific T cell responses first started at 2 wk postinfection by a predominant IFN-γ response specific for p68–82 (Fig. 4⇓A) and a subdominant response (because of a much lower number of spots) to p277–291 (Fig. 4⇓D). The number of cells secreting IFN-γ in response to these two peptides reached a peak at 4 wk and then slowly declined over the time remaining after 16 wk to the level observed at 2 wk. During the same period of time, between 2 and 16 wk postinfection, no IFN-γ response was observed for p151–165 (Fig. 3⇑C), and a poor, very variable, and transient response was detected for p100–114 (Fig. 4⇓B; only two mice responded to p100–114 of six tested at 4 wk postinfection). As observed for the response to the MalE protein (Fig. 3⇑A), MalE-peptide-specific, IL-4-secreting cells were detected from 8 wk postinfection (Fig. 4⇓). Strikingly, this IL-4 response was markedly different from the IFN-γ response in terms of fine specificity. Indeed, high and comparable numbers of IL-4-secreting cells were found in response to the four MalE peptides, p68–82, p100–114, p151–165, and p277–291. This latter result clearly shows that during the late phase of rBCG.MalE infection, responses specific for p68–82 and p277–291 correspond to a mixed Th1/Th2 (or to a Th0) phenotype, whereas responses specific for p100–114 and p151–165 MalE peptides were exclusively Th2-like. The number of IL-4-secreting cells was predominant at 16 wk postinfection, with 80% of MalE peptide-specific, IL-4-producing cells as opposed to the 100% of MalE peptide-specific IFN-γ secreting cells observed during the first 4 wk of infection (Fig. 4⇓E). The relative percentage of IL-4-secreting cells after restimulation with MalE protein (Fig. 3⇑A) was lower than that after peptide restimulation, most likely due to the suboptimal in vitro presentation of p100–114 and p151–165 following MalE processing.
Cytokines produced by T cells in response to the different MalE epitopes following rBCG.MalE infection. BALB/c mice (two per group) were i.v. immunized with 106 CFU of rBCG.MalE. At different time points, the numbers of MalE peptide-specific T cells secreting IFN-γ (□) or IL-4 (•) were measured in individual immune spleen after in vitro stimulation with p68–82 (A), p100–114 (B), p151–165 (C), or p277–291 (D) MalE peptides using an ELISPOT assay. The results are expressed as the mean number of peptide-specific secreting cells per spleen. Results are representative of two independent experiments. E, The cumulative numbers of cells secreting IFN-γ or IL-4 for each mouse after stimulation with the four peptides were used to calculate the percentage of MalE peptide-specific CD4+ T cells secreting IFN-γ or IL-4 at 4 and 16 wk after rBCG.MalE immunization. The results of five individual mice are shown.
Since IL-4-producing cells specific for p100–114 and p151–165 peptides were only detected from 8 wk postinfection, we wondered whether these cells could produce cytokines other than IL-4 or IFN-γ at earlier times of infection. We thus analyzed the IL-2-secreting cells in response to the different MalE peptides at 4 and 8 wk postinfection (Fig. 5⇓). In these experiments we consistently detected IL-2-secreting cells specific for p68–82, p100–114, and p277–291 peptides even in the absence of IL-4- or IFN-γ-secreting cells. For p151–165, IL-2-producing cells were only rarely evidenced, and in such a case, a low number of specific cells were detected.
Specificity of IL-2-secreting cells following rBCG.MalE infection. BALB/c mice (two per group) were i.v. immunized with 106 CFU of rBCG.MalE or BCG.wt. At 4 (A) and 8 (B) wk postinfection, the MalE peptide-specific T cells secreting IL-2 were measured in individual immune spleens after in vitro stimulation with p68–82, p100–114, p151–165, or p277–291 peptides using an ELISPOT assay. The results are expressed as the mean number of peptide-specific CD4+ T cells secreting IL-2 per spleen. Results are representative of two independent experiments.
The Th2 shift during BCG infection does not prevent the elicitation of IFN-γ effector functions upon in vivo recall
Eight to 10 wk after rBCG.MalE infection, the reactivity to the four MalE peptides was characterized by mixed IFN-γ and IL-4 responses to p68–82 and p277–291 and by an IL-4 response to p100–114 and p151–165. Therefore, we next analyzed the influence of these polarized immune responses on the cytokine profile induced by a subsequent injection of purified MalE protein. At 12 wk postinfection with rBCG.MalE, mice were boosted with MalE in CFA, and peptide-specific IFN-γ and IL-4 responses were monitored and compared with those elicited following immunization of naive mice with MalE in CFA (Fig. 6⇓, A and B). As expected, MalE peptide-specific responses in mice previously primed with the rBCG.MalE were stronger after the MalE boosting injection than those in mice that received only one MalE injection. This increase in MalE-specific T cell responses was observed for all peptides, but only for IFN-γ production and not for the IL-4 response. The production of IFN-γ and the lack of IL-4 response in rBCG.MalE-primed mice correspond to the pattern of response induced by a single immunization with MalE in CFA.
Modulation of the Th1/Th2 profile during BCG infection does not preclude the elicitation of Th1 effector functions upon in vivo recall with the MalE protein. BALB/c mice (eight per group) infected 12–14 wk before with 106 CFU of rBCG.MalE or left untreated were s.c. immunized with 20 μg of MalE protein emulsified in CFA (A and B) or IFA (C and D). Ten days later, draining lymph node cells pooled from immunized mice (groups were divided into two subgroups of four mice and tested independently) were restimulated in vitro with 10 μg/ml of p68–82, p100–114, p151–165, or p277–291 MalE peptides or with medium only in triplicate. Culture supernatants were harvested and tested for IL-4 (A and C) and IFN-γ (B and D) contents by ELISA. Results are expressed as picograms per milliliter and correspond to the mean value obtained for two subgroups of mice that received the same treatment. The results are representative of three independent experiments.
We then compared the recalled response of rBCG.MalE primed mice following boosting with MalE emulsified in IFA (Fig. 6⇑, C and D). Following immunization of naive mice with MalE in IFA, the MalE peptide-specific response was characterized by a mixed production of IL-4 and IFN-γ. When rBCG.MalE-primed mice were boosted with MalE in IFA, an enhanced IL-4 response as well as a strong IFN-γ production to all MalE peptides were observed. Together, these results show that the phenotype of the recalled response is driven by the properties of the adjuvant used for the boost (either CFA or IFA). Therefore, the Th2 shift of the response observed during the course of the BCG infection does not interfere with the phenotype of the T cell response induced by Ag re-exposure in vivo.
Discussion
In the present study we have analyzed the CD4+ T cell responses induced against a reporter Ag delivered by recombinant BCG. We demonstrated that during the early steps of infection, the T cell response induced against this Ag is focused on two dominant epitopes and is characterized by IFN-γ production. Then, in the course of infection, the initial IFN-γ response to these two epitopes shifted to a mixed IFN-γ/IL-4 response. At the same time, the peptide specificity of the T cell response was broadened to two additional epitopes characterized by a unique IL-4 response resulting in the establishment of a dominant IL-4 response to the reporter Ag at 16 wk postinfection. However, this phenomenon did not impair the outcome of a predominant IFN-γ response upon subsequent recall in vivo with the purified reporter Ag performed in the presence of CFA, a Th1-driving adjuvant.
Orme et al. (17) previously analyzed the CD4+ T cell responses of mice at several times during the course of a virulent M. tuberculosis infection and provided evidence of two separate waves of cytokine-producing CD4 T cells. They showed that early after infection, the immune response induced by mycobacteria is characterized by IFN-γ production, in which secreted proteins of the bacilli are primary targets. A few weeks later, a Th2 response is observed in which the mycobacteria 60-kDa heat shock protein is a strong target. In the present study a similar sequential switch from Th1 to Th2 responses was observed. However, our data clearly demonstrated that these Th1 and Th2 responses did not differ in terms of Ag specificity, since they were both directed against the same reporter Ag. Indeed, two MalE epitopes (p68–82 and p277–291) induced, first, a pure Th1 response that subsequently shifted to a mixed Th1 and Th2 pattern, whereas the same two epitopes induced a Th1 response in CFA.
Using the same recombinant BCG carrying the reporter MalE Ag, we have recently demonstrated that dendritic cells are the only cells that after mycobacterial infection in vivo can present MalE peptide-MHC II complexes to specific T cell hybridoma. This event is transient, and no MalE peptide-MHC II complexes were detectable a few days after infection, although dendritic cells remained infected for weeks after BCG administration (9). These results suggest that the IL-4-producing T cells are derived from effector or memory Τ cells recognizing the same epitopes and producing IFN-γ, which could switch to a Th2 profile when stimulated by infected dendritic cells presenting low amounts of MalE peptides-MHC II complexes. Indeed, after recall of these mice with MalE in CFA, these mice developed memory Th1 responses against these epitopes, showing that memory T cells are present and that the Th1/2 phenotype of the expressed memory response seems to be governed by Ag re-exposure conditions.
Th1/2 polarization was shown to be dependent on the strength of the signal delivered by the APC, which is controlled in the first instance by ligand density (amount of Ag and affinity of the peptide for a given MHC once processing steps are successfully achieved) and then modulated by accessory and costimulatory molecules (23, 24, 25). Therefore, the Th1 to Th2 switch observed a few weeks after infection could be due to the decrease in the amount of MHC II complexes at the surface of infected dendritic cells. Alternatively, the loss of IFN-γ-producing cells observed in this study could be explained by the negative feedback loop recently described in BCG-infected mice (26). Indeed, in these mice the large expansion of an activated splenic CD4+ T cell population was followed by a rapid contraction of this population to normal numbers. The contraction of the activated CD4+ T cells was associated with increased apoptosis and was shown to depend upon IFN-γ. Thus, following this depletion of activated Th1 cells, new T cells could be stimulated under conditions promoting the activation of Th2-polarized responses. In this respect, it should be noted that upon rBCG infection, we detected T cell responses characterized by dominant IL-4 production against two additional MalE epitopes (p100–114 and p151–165), but only after 8 wk of infection.
A third explanation could be related to the recent demonstration that dendritic cells produce large amounts of IL-12 only during a brief interval after activation, which coincides with their capacity to stimulate naive T cells to become Th1 effector memory cells (27). At later time points dendritic cells might only be able to stimulate Th2 responses or central memory cells (28). This exhaustion of cytokine production could indeed play a major role in the regulation of immune responses during chronic infection such as BCG and could explain the temporal Th1 to Th2 switch observed in the present study.
Moreover, upon Ag challenge with CFA or IFA, the secondary response elicited displayed a Th1 or Th1/Th2 phenotype, indicating that a nonpolarized T cell response was restimulated that may correspond to central memory T cells developing into effector memory responses (29).
The consequences of this Th1/2 shift to the fate of mycobacterial infections are still unclear. However, since IL-4 deficiency does not modify the course of mycobacterial infections in mice (30, 31), this could suggest that the Th2 response occurring in the late phase of infection does not affect the course of this infection. The loss of Th1 responses associated with the activation of Th2 responses may represent a mechanism to avoid a prolonged inflammatory response and its damaging effects.
In conclusion, this study has documented the Th1 to Th2 shift of the adaptive T cell response during a mycobacterial infection. Following a slight decrease of the IFN-γ response, an IL-4 response emerges during the late phase of BCG infection that becomes predominant in cell number. By analyzing the T cell response to a reporter Ag, we showed here that the immunodominance of the T cell response in terms of peptide specificity and cytokine secretion is variable across the time of infection, but does not predetermine the Th1/2 phenotype of a memory response.
Footnotes
-
↵1 This work was supported by Grant PRA B01-04 (Association Franco-Chinoise pour la Recherche Scientifique et Technique) and by the Fok Ying Tung Education Foundation and Pasteur Weizmann grants (to X.J.).
-
↵2 Address correspondence and reprint requests to Dr. Claude Leclerc. Unìté de Biologie des Rêgulations Immunitaires, Institut Pasteur, 75015 Paris, France, E-mail address: cleclerc{at}pasteur.fr
-
↵3 Abbreviations used in this paper: BCG, bacille Calmette-Guérin; HAT, hypoxanthine, aminopterin, thymidine; PPD, purified protein derivative.
- Received March 6, 2002.
- Accepted November 15, 2002.
- Copyright © 2003 by The American Association of Immunologists
References
In this issue
