Interference or competition between CD8+ T cells restricted by distinct MHC-I molecules can be a powerful means to establish an immunodominant response. However, its importance during infections is still questionable. In this study, we describe that following infection of mice with the human pathogen Trypanosoma cruzi, an immunodominant CD8+ T cell immune response is developed directed to an H-2Kb-restricted epitope expressed by members of the trans-sialidase family of surface proteins. To determine whether this immunodominance was exerted over other non-H-2Kb-restricted epitopes, we measured during infection of heterozygote mice, immune responses to three distinct epitopes, all expressed by members of the trans-sialidase family, recognized by H-2Kb-, H-2Kk-, or H-2Kd-restricted CD8+ T cells. Infected heterozygote or homozygote mice displayed comparably strong immune responses to the H-2Kb-restricted immunodominant epitope. In contrast, H-2Kk- or H-2Kd-restricted immune responses were significantly impaired in heterozygote infected mice when compared with homozygote ones. This interference was not dependent on the dose of parasite or the timing of infection. Also, it was not seen in heterozygote mice immunized with recombinant adenoviruses expressing T. cruzi Ags. Finally, we observed that the immunodominance was circumvented by concomitant infection with two T. cruzi strains containing distinct immunodominant epitopes, suggesting that the operating mechanism most likely involves competition of T cells for limiting APCs. This type of interference never described during infection with a human parasite may represent a sophisticated strategy to restrict priming of CD8+ T cells of distinct specificities, avoiding complete pathogen elimination by host effector cells, and thus favoring host parasitism.
Major histocompatibility complex class Ia-restricted CD8+ T cells are important mediators of the adaptive immune response against infections caused by intracellular microorganisms. Pathogens have a number of potential amino acid sequences that can bind to MHC-I molecules and provide targets for specific CD8+ T cells. In contrast, hosts have a vast number of TCR clonally distributed on these CD8+ T cells to recognize a variety of MHC class I (MHC-I)3-peptide complexes. Despite this large possible combination, pathogen-specific CD8+ T cells seem to recognize preferentially a small number of epitopes. This preference, known as immunodominance, may occur due to different mechanisms including the formation of stable MHC-I peptide complexes on the surface of APC, higher amounts or higher affinity specific T cell precursors or competition of T cells for APC. The full biological implications of immunodominance during effector and memory immune responses are unknown and are being thoroughly studied during infection with viruses or bacteria (reviewed in Refs. 1, 2, 3, 4, 5, 6).
Infection of humans or mice with the digenetic intracellular protozoan parasite Trypanosoma cruzi also induces MHC class Ia-restricted CD8+ T cells specific for parasite epitopes (7, 8). This T cell subpopulation is critical for host survival following infection even when small challenge doses of parasites are used to initiate infection (Refs. 9, 10, 11 and reviewed in Ref. 12). Despite the fact that CD8+ T cell mediated immune response is critical for host survival during acute infection, T. cruzi manages to survive within the host and establishes a lifelong chronic infection. Parasite persistence is considered a key element in the development of symptoms that occur many years or decades after initial infection in approximately one-third of these chronically infected individuals (13, 14, 15). Thus, understanding the specificity, magnitude, and longevity of CD8+ T cell immune responses may greatly help us to explain the complex immunopathology caused by T. cruzi and to propose new means for intervention for Chagas’ disease, which affects ∼20 million people in Latin America.
Studies on the immunodominance of CD8+ T cell immune responses during T. cruzi infection were thought not to be simple to accomplish due to the large parasite genome which contains >12,000 genes (16). Nevertheless, the description of epitopes targets of powerful CD8+ T cell immune responses elicited during T. cruzi infection in distinct inbred mouse strains allowed us to initiate studies on the immunodominance phenomenon (7, 8). Surprisingly, these studies suggest that CD8 response is highly confined to epitopes expressed by different members of a large family of T. cruzi surface Ags named trans-sialidases (TS). Thus, studies performed in our laboratory and elsewhere, show that C57BL/6, BALB/c, and more recently B10.A mice infected with T. cruzi develop a strong and long-lasting CD8+ T cells specific for, respectively, H-2Kb-, H-2Kd-, and H-2Kk-restricted epitopes expressed by members of the TS family of surface Ags.
The purpose of the present study was to determine the mechanisms that established the immunodominance during naturally acquired immunity following experimental infection with T. cruzi. Toward that goal, we compared the magnitude of CD8+ T cell immune responses to a number of different epitopes from members of the TS family, which are recognized by H-2Kb-restricted T cells. We found that the epitope VNHRFTLV was immunodominant following infection of C57BL/6 mice with parasites of two different strains. Subsequently, we investigated whether this immunodominance could be exerted on the immune response to CD8 T cell epitopes restricted by different MHC-I molecules. We accomplished that by comparing specific immune responses to three distinct H-2Kb-, H-2Kd-, or H-2Kk-restricted epitopes following T. cruzi infection of homozygote (C57BL/6, BALB/c, and B10.A) or heterozygote (F1) mice. Our results provide strong evidence that homozygote (C57BL/6) and heterozygote mice develop similar immune responses to immunodominant H-2Kb-restricted epitope VNHRFTLV. However, immune responses to H-2Kd- or H-2Kk-restricted epitopes were severely diminished in heterozygote mice. This interference/competition was restricted to T. cruzi infection because following immunization with recombinant adenovirus, heterozygote mice responded to these same T. cruzi epitopes, as well as or better than homozygote strains. Finally, the possible mechanism mediating this strong immunodominance was evaluated by simultaneous infection with two T. cruzi strains containing distinct immunodominant epitopes. These mice responded equally well to the “dominant” and “subdominant” epitopes suggesting that the operating mechanism most likely involves competition of T cells for limiting APCs. We believe that this competition between T cells of different specificities is a sophisticated strategy that T. cruzi developed to restrict CD8+ T cell responses and escape complete elimination by host effector cells, and thus favoring host parasitism.
Materials and Methods
Mice and parasites
Female 8- to 10-wk-old wild-type C57BL/6, BALB/c, B10.A, F1 C57BL/6 × BALB/c, F1 C57BL/6 × B10.A, F1 B10.A × BALB/c mice were obtained from Federal University of São Paulo. Experimental procedures were approved by the Committee of Ethics of Federal University of São Paulo.
Parasites of Y, G, or CL-Brener strains of T. cruzi were used in this study (17, 18). Bloodstream trypomastigotes of Y strain were obtained from plasma of A/Sn mice infected 7 days earlier. Tissue culture trypomastigotes of the G or CL-Brener strain were used to infect mice. The concentration of parasites was adjusted and each mouse was inoculated i.p. with 0.2 ml containing the indicated amount of trypomastigotes. C57BL/6, B10.A, F1 C57BL/6 × BALB/c, F1 C57BL/6 × B10.A, F1 B10.A × BALB/c were challenged, in most experiments, with 104 bloodstream trypomastigotes i.p. BALB/c mice were challenged with only 2.5 × 103 parasites because they are more susceptible to infection and succumb to challenge doses higher than 5 × 103 parasites.
Peptides VNHRFTLV, TEWETGQI, TsKb-18 (ANYDFTLV), TsKb-20 (ANYKFTLV) were prepared by standard Nα[9-fluorenylmethyloxycarbonyl] on a PSSM8 multispecific peptide synthesizer (Shimadzu) by solid-phase synthesis with a scale of 30 μM. Peptide was purified by HPLC in a Shimadzu system. Peptides were analyzed in a C18 Vydac column (10 × 250 mm, 5-μm particle diameter). Different peptide batches were obtained in a range of 80–90% purity. Their identities were confirmed by Q-TOF Micro equipped with an electrospray ionization source (Micromass). Peptide IYNVGQVSI was purchased from Neosystem. As estimated by HPLC analysis, peptide IYNVGQVSI was >90% pure.
For the in vivo cytotoxicity assays, splenocytes of the different mouse strains were divided into two populations and labeled with the fluorogenic dye CFSE (Molecular Probes) at a final concentration of 5 μM (CFSEhigh) or 0.5 μM (CFSElow). CFSEhigh cells were pulsed for 40 min at 37°C with 1–2.5 μM of H-2Kb ASP-2 peptide (VNHRFTLV), or H-2Kb TsKb-18 peptide (ANYDFTLV), H-2Kb TsKb-20 peptide (ANYKFTLV), or H-2Kd TS peptide (IYNVGQVSI) peptide or H-2Kk ASP-2 peptide (TEWETGQI). CFSElow cells remained unpulsed. Subsequently, CFSEhigh cells were washed and mixed with equal numbers of CFSElow cells before injecting i.v. 15–20 × 106 total cells per mouse. Recipient animals were mice that had been infected or not with T. cruzi. Spleen cells of recipient mice were collected 20 h after transfer, fixed with 3.7% paraformaldehyde and analyzed by FACS, using a FACSCalibur Cytometer (BD Biosciences). Percentage of specific lysis was determined using the formula: 1 − ((% CFSEhigh infected/% CFSElow infected)/(%CFSEhigh naive/% CFSElow naive)) × 100%. In experiments using adenovirus, we estimated the percentage of CFSEhigh or CFSElow cells in immunized, not infected mice. ELISPOT assay for enumeration of IFN-γ-producing cells was performed essentially as described earlier (19).
MHC-I tetramer IYNVGQVSI/Kd or pentamer VNHRFTLV/Kb were synthesized at the Tetramer Core Facility (Emory University, Atlanta, GA) or were purchased from ProImmune, respectively. Mouse splenocytes were analyzed. Single-cell suspensions of splenocytes were washed in PBS, stained for 15 min at 37 °C with tetramers, then stained 30 min at 4°C with labeled anti-CD8 Abs as well as anti-CD4, -CD11b, and -B220. At least 1,000,000 cells were acquired on a FACSCalibur flow cytometer (BD Pharmingen) then analyzed with FlowJo (Tree Star) using a biexponential transform.
Recombinant adenoviruses and immunization
pAdCMV-TS is an adenoviral transfer plasmid that contains an eukaryotic expression cassette formed by the CMV immediate-early promoter and the SV40 RNA polyadenylation sequences. Inside this cassette, we cloned the DNA sequences encoding T. cruzi TS protein signal peptide and catalytic domain obtained by restriction enzyme digestion of plasmid p154/13 (adeno-TS). Equivalently, pAdCMV-amastigote surface protein-2 (ASP2) encodes ASP2 sequences obtained by restriction digestion of plasmid pIgSP clone 9 (adeno-ASP-2). Viruses were generated and purified as described earlier (20).
Mice were inoculated i.m. in each tibialis anterioris muscle with 50 μl of viral suspension containing 5 × 107 PFU of each adenovirus. Immunological assays were performed 15 days after viral inoculation.
Values were compared by one-way ANOVA followed by Tukey honestly significant difference tests available at http://faculty.vassar.edu/lowry/VassarStats.html. Differences were considered significant when the p value was <0.05.
Immunodominance of the CD8+ T cell immune response to the H-2Kb-restricted epitope VNHRFTLV in C57BL/6 mice infected with T. cruzi of Y or G strains
Three epitopes recognized by H-2Kb-restricted CD8+ T cells were described as targets of strong and long-lasting immune responses following infection of C57BL/6 mice with T. cruzi (Table I, Refs. 7 and Ref. 8). These epitopes are expressed by members of the TS family of surface proteins/Ags. The characterization of the immune response to the epitope VNHRFTLV was performed by our group following infection of C57BL/6 mice with parasites of the Y strain of T. cruzi (7). The immune responses to the other two epitopes (TsKb-18 and TsKb-20) were described in C57BL/6 mice infected with parasites of the Brazil, CL, and Y strain (8). TsKb-18 (ANYDFTLV) and TsKb-20 (ANYKFTLV) epitopes display a single amino acid substitution (D→K) but are recognized by CD8+ T cells of distinct specificities (Table I and 8).
|Peptide/Epitope .||Sequence .||H-2K Restriction .||Ref. .|
|Peptide/Epitope .||Sequence .||H-2K Restriction .||Ref. .|
To evaluate the immunodominance pattern among these three epitopes after infection of C57BL/6 mice with T. cruzi, we estimated in vivo peptide-specific cytotoxicity and ex vivo the number of peptide-specific IFN-γ-producing cells following infection with parasites of the Y strain. In this mouse model, specific immune responses were first detected 9 days following infection. The maximum in vivo cytotoxicity occurred from days 15 to 90 following infection (7). Fig. 1,A shows the kinetics of the in vivo cytotoxicity specific for target cells coated with peptide VNHRFTLV. More than 96% specific lyses were detected at day 15, 30, or 60 after challenge. These values were significantly higher than cytotoxicity levels specific for target cells coated with peptides TsKb-18 or TsKb-20. In these cases, the maximum specific cytotoxicities were 52.0 ± 2.3 or 63.8 ± 4.0, respectively. The numbers of splenic IFN-γ-producing cells specific for peptides VNHRFTLV or TsKb-20 detected ex vivo by ELISPOT assay were quite similar at day 15 postinfection. However, at days 30 and 60 postinfection, the number of IFN-γ-producing cells specific for peptide VNHRFTLV was more than three times higher than cells specific for TsKb-20 (Fig. 1,B). Despite the fact that TsKb-18 peptide has a single amino acid substitution when compared with TsKb-20, very few IFN-γ-producing cells specific for this epitope were detected ex vivo at day 15, 30, or 60 postinfection (Fig. 1,B). Analysis of the immune T cell affinities for the peptides VNHRFTLV or TsKb-20 failed to provide any significant degree of difference (Fig. 1 C).
We also estimated the number of IFN-γ-producing cells detected ex vivo following infection with a second strain of T. cruzi (G strain). We found that the number of IFN-γ-producing cells specific for peptide VNHRFTLV was significantly higher than the number of cells specific for peptides TsKb-18 or TsKb-20 (Fig. 1,D). In mice infected with the G strain, the values of the in vivo cytotoxicity specific for target cells coated with peptide VNHRFTLV were also significantly higher than cytotoxicity levels specific for target cells coated with peptide TsKb-18 or TsKb-20 (p < 0.001, Fig. 1 E).
Similar experiments performed with C57BL/6 mice infected with a third T. cruzi strain (CL-Brener) provided a distinct result. In these animals, we found that the number of IFN-γ-producing cells specific for peptide VNHRFTLV was significantly lower than the number of cells specific for peptides TsKb-18 or TsKb-20 (Fig. 1,D). In mice infected with the CL-Brener strain, the values of the in vivo cytotoxicity specific for target cells coated with peptide VNHRFTLV were also significantly lower than cytotoxicity levels specific for target cells coated with peptide TsKb-18 or TsKb-20 (Fig. 1 E).
In addition, a number of other peptides from T. cruzi previously described as capable of binding to H-2Kb were also tested ex vivo by ELISPOT using splenic cells from C57BL/6 infected 15 or 30 days earlier with parasites of the Y strain (21, 22, 23). None of them stimulated IFN-γ-producing cells (data not shown). Based on these observations, it seems that the CD8+ T cell immune response to VNHRFTLV is immunodominant during infection of C57BL/6 mice with parasites of the Y or G strain of T. cruzi. In contrast, during infection of C57BL/6 mice with parasites of the CL-Brener strain, CD8+ T cell immune response to TsKb-18 and TsKb-20 are immunodominants.
T. cruzi- infected heterozygote mice displays significantly reduced CD8+ T cell immune responses to certain parasites epitopes
This immunodominance could be explained by several nonmutually exclusive reasons. The first possibility is that during T. cruzi infection the number of MHC-I-VNHRFTLV complexes formed on the surface of APC is higher than the other two epitopes. Alternatively, this complex could be exposed for a longer period. A third possibility is that C57BL/6 mice have a greater number of, or higher affinity, precursor T cells specific for MHC I-VNHRFTLV complexes. Also, it is possible that a competition occurs between specific T cells for APC. This last possibility could be caused, or increased, by the first three and may provide a much stronger mechanism of immunodominance. In this case, the immunodominant response would “interfere” with priming of immune responses specific for epitopes restricted by distinct MHC-I molecules (24).
To investigate some of these possibilities, we took advantage of the fact that we had previously identified two other epitopes recognized by H-2Kk- (TEWETGQI) or H-2Kd (IYNVGQVSI)-restricted CD8+ T cells from T. cruzi infected mice (Table I and Ref. 7). The presence of strong immune responses to epitopes restricted by distinct MHC-I alleles allowed us to investigate a possible interference of one CD8+ T cell-mediated immune response on the others, or simultaneously on one another. To evaluate that possibility, we infected heterozygote and homozygote mice and compared the immune responses to three distinct epitopes. Fifteen days after challenge, homozygote mice (C57BL/6 or BALB/c) displayed >90% in vivo cytotoxicity specific for target cells coated with peptide VNHRFTLV (C57BL/6) or IYNVGQVSI (BALB/c). Heterozygote mice (F1 C57BL/6 × BALB/c) also presented an in vivo cytotoxicity against target cells coated with peptide VNHRFTLV above 95%. In contrast, only 12.92 ± 5.91% of the target cells coated with peptide IYNVGQVSI were eliminated in vivo (Fig. 2,A). Analysis of the number of splenic peptide-specific IFN-γ-producing cells revealed a similar picture. Although homozygote mice (C57BL/6 or BALB/c) had a significant number of T cells specific for the respective MHC-I-restricted epitope, heterozygote mice responded well only to H-2Kb-restricted epitope VNHRFTLV (Fig. 2 A). In F1 C57BL/6 × BALB/c mice, the number of T cells specific for H-2Kd-restricted epitope IYNVGQVSI was only 18.51% of the number detected in BALB/c mice.
Similar experiments were performed by comparing CD8+ T cell immune responses of homozygote mice (C57BL/6 or B10.A) with heterozygote F1 C57BL/6 × B10.A. Homozygote mice presented an immune response capable of eliminating in vivo >90% of peptide-coated target cells. Heterozygote animals also displayed an in vivo cytotoxicity against target cells coated with peptide VNHRFTLV of 86.39 ± 6.90%. In contrast, in these animals only 29.5 ± 10.6% of target cells coated with peptide TEWETGQI were eliminated in vivo (Fig. 2 C).
Analysis of the number of splenic peptide-specific IFN-γ-producing cells revealed that homozygote mice had a significant number of T cells specific for the respective MHC-I-restricted epitope. Heterozygote mice responded to the H-2Kb-restricted epitope VNHRFTLV similarly to homozygote animals (Fig. 2 D). In contrast, in F1 C57BL/6 × B10.A mice, the number of TEWETGQI-specific cells was 8.59% of the number detected in homozygote B10.A mice.
The last group of heterozygote mice we studied (F1 B10.A × BALB/c) presented an interesting picture. When measured in vivo by cytotoxicity, the elimination of target cells coated with peptide TEWETGQI or IYNVGQVSI was lower in heterozygote mice than in homozygote ones, suggesting that immune responses of both specificities interfered with one another (Fig. 2,F). Analysis of the number of splenic peptide-specific IFN-γ-producing cells revealed that homozygote mice had a significantly higher number of T cells specific for the respective MHC-I-restricted epitope. Heterozygote mice had <40% inhibition of the immune response to the H-2Kk-restricted epitope TEWETGQI when compared with homozygote animals. However, a drastic reduction of 92.75% was seen in the immune response to H-2Kd-restricted epitope IYNVGQVSI (Fig. 2 F).
The impaired immune responses observed against these parasite epitopes in heterozygote mice could be dependent on the dose of T. cruzi used for challenge. To test this hypothesis, we challenged each mouse with 105 parasites, a dose 10 times higher than used in the previous experiments. An increase in the dose of parasite challenge did not improve the limited immune response observed in F1 C57BL/6 × BALB/c mice to the epitope IYNVGQVSI. Both, the in vivo cytotoxicity or the number of IFN-γ-producing cells were similar in mice challenged with 104 or 105 parasites (Fig. 3, A–C). In contrast, the number of IFN-γ-producing cells specific for the immunodominant epitope VNHRFTLV increased in mice challenged with 105 parasites (Fig. 3,C). Also, the reduced immune responses observed against the epitope IYNVGQVSI in F1 C57BL/6 × BALB/c mice was not due to a delay in the priming and/or expansion of specific CD8+ T cells. The kinetics of IFN-γ-producing T cells specific for VNHRFTLV showed a vigorous expansion during the first 30 days following infection while cells specific for IYNVGQVSI expanded very little in that period (Fig. 3 C). Specific immune response measured by in vivo cytotoxicity to the epitope IYNVGQVSI or VNHRFTLV was performed in mice infected 60 days earlier. Although the immune response to VNHRFTLV was still high (91.1 ± 0.5%, n = 3) at that point in time, the immune response to the epitope IYNVGQVSI was still negative (0.0 ± 0.0, n = 3, data not shown).
Finally, we used the staining of specific T cells to visualize whether in fact there was a significantly lower number of CD8+ T cells specific for the epitope IYNVGQVSI in F1 C57BL/6 × BALB/c mice. As shown in Fig. 4,A, the total number of CD8+ spleen cells from infected F1 C57BL/6 × BALB/c mice stained with the pentamer VNHRFTLV/Kb was only 18.35% lower than the number of cells in C57BL/6. When similar analysis was performed to estimate the number CD8+ spleen cells stained with the tetramer IYNVGQVSI/Kd, we found that infected F1 C57BL/6 × BALB/c mice had 91.5% less cells than BALB/c mice (Fig. 4 B).
Analysis of the immune T cell affinities for the peptides VNHRFTLV or IYNVGQVSI were performed using immune T cells from C57BL/6 or BALB/c mice, respectively. This analysis provided evidence that the affinity of T cells specific for the H-2Kd-restricted epitope IYNVGQVSI was slightly higher than the affinity of the T cells specific for the H-2Kb-restricted epitope VNHRFTLV, a fact that cannot explain the immunodominance of the later (Fig. 5).
Analysis of the expression of MHC-I molecules H-2Kb and H-2Kd were performed using spleen cells of C57BL/6, BALB/c, and F1 C57BL/6 × BALB/c mice (Fig. 6). The expression of either H-2Kd or H-2Kb molecules on the surface of F1 C57BL/6 × BALB/c mice was lower than the homozygote animals. Nevertheless, in normal F1 C57BL/6 × BALB/c mice, the expression of H-2Kd molecules (nonresponder allele) in B220+ or CD11c+ or other cell types were higher than the expression of H-2Kb molecules (responder allele). The same picture was observed when we analyzed spleen cells from infected F1 C57BL/6 × BALB/c. Therefore the absence of immune response in the F1 C57BL/6 × BALB/c to the H-2Kd-restricted epitope IYNVGQVSI cannot be attributed to a significantly lower expression of H-2Kd molecules on APCs. Other experiments provided below also contribute to this interpretation.
Immune responses of heterozygote mice to T. cruzi epitopes following immunization with recombinant adenovirus are not impaired
A possible explanation for the impaired immune response to the H-2Kd-restricted epitope IYNVGQVSI in heterozygote mice could be the development of a severely skewed T cell repertoire in the presence of MHC-I H-2Kb or H-2Kk molecules. Also, the number of H-2Kd molecules on the surface of cells from heterozygote mice could be responsible for the reduced response to the epitope IYNVGQVSI. To test these hypotheses in vivo, we immunized homozygote (BALB/c) and heterozygote (F1 C57BL/6 × BALB/c or F1 B10.A × BALB/c) mice with a recombinant adenovirus expressing the TS protein, the Ag which contains the epitope IYNVGQVSI (20). Following immunization with recombinant adenovirus, the peak CD8+ T cell immune response is at 15 days (25). The immune responses of heterozygote mice measured by the in vivo cytotoxicity against target cells coated with the peptide IYNVGQVSI was higher than homozygote BALB/c (Fig. 7 A). Therefore, after immunization with a recombinant virus, the heterozygote mice immune response to the peptide IYNVGQVSI was not impaired.
The limited immune response to epitope TEWETGQI in heterozygote (F1 C57BL/6 × B10.A) could be explained by a reduced Ag presentation in vivo to H-2Kk-restricted T cells specific for the epitope TEWETGQI caused by some intrinsic property of the H-2Kb-restricted epitope VNHRFTLV. Alternatively, T cells specific for VNHRFTLV could inhibit the activation of T cells specific for TEWETGQI by a direct interaction between CD8+ T cells. To approach these questions, we used a second recombinant adenovirus expressing the ASP-2 of T. cruzi. This Ag contains both of these epitopes (20). Homozygote (C57BL/6 or B10.A) or heterozygote (F1 C57BL/6 × B10.A) mice immunized with the recombinant adenovirus adeno-ASP-2 developed immune response specific for H-2Kb-restricted epitope VNHRFTLV significantly higher than the response to H-2Kk-restricted epitope TEWETGQI. Nevertheless, the immune response to the epitope TEWETGQI was identical in homozygote or heterozygote mice (Fig. 7 B). These results indicated that although the immune response to the epitope VNHRFTLV seems to be stronger following immunization with adeno-ASP-2 virus, it did not interfere with the smaller response to the epitope TEWETGQI.
Comparison of the CD4 T cell dependence on the immune response following infection with T. cruzi or immunization with recombinant adeno-ASP-2 was performed using CD4 knockout mice. We observed that these mice had a severe reduction on the in vivo cytotoxic immune response to the epitope VNHRFTLV following infection or immunization proving that both cytotoxic immune responses were dependent on CD4+ T cell activation (data not shown).
Immune responses of C57BL/6 mice to T. cruzi epitopes following infection with two parasite strains containing distinct immunodominant epitopes
Precisely how this immunodominant response to VNHRFTLV epitope interferes with T cells of other specificities during T. cruzi infection was not clear. It was possible that T cells specific for the epitope VNHRFTLV prevented the presentation of the other epitopes when both were present in the same APCs. In that case, we predicted that strong responses to both epitopes (dominant and subdominant) would be induced when the epitopes were expressed in different APCs. To test this hypothesis, we infected C57BL/6 mice with two different parasite strains expressing distinct immunodominant epitopes. As described above, infection with the Y strain of T. cruzi induced a strong response to the epitope VNHRFTLV, but not to TsKb-18, as measured by in vivo cytotoxicity or ELISPOT assay (Fig. 8). In contrast, infection with parasites of the CL-Brener strain elicited strong immune responses to the epitope TsKb-18, but not to VNHRFTLV (Fig. 8). As we predicted, mice infected with both parasite strains at the same time displayed strong immune response to both epitopes (Fig. 8). This result suggests that separation of the “dominant” and “subdominant” epitopes can circumvent the immunodominance.
Interference or competition between CD8+ T cells restricted by distinct MHC-I molecules have been described following immunization with different Ags (24, 26, 27, 28). It is a powerful means used by CD8+ T cells specific for a particular MHC-I-peptide complex to restrict expansion of T cells of all other specificities establishing an immunodominant response. Whether this type of interference/competition between CD8+ T cells restricted for different MHC-I molecules could in fact be critical during infections is still a matter of debate (29, 30, 31). Here, we demonstrate that CD8+ T cell immune response specific for H-2Kb-restricted epitope VNHRFTLV is immunodominant during infection of C57BL/6 mice with T. cruzi of Y and G strains. In addition, we have demonstrated that CD8+ T cells specific for the immunodominant epitope VNHRFTLV greatly interferes with the generation of T cells restricted for other MHC-I molecules. Thus, following infection of heterozygote mice with T. cruzi, a significant reduction in the priming of CD8+ T cells restricted for H-2Kd or H-2Kk molecules was observed providing evidence that interference/competition is an important factor down-modulating immune response during experimental infection with this human pathogen. In the case of infection of F1 B10.A × BALB/c mice, a mutual interference takes place reducing both specific responses to the H-2Kd- or H-2Kk-restricted T cells.
Precisely how this immunodominant response to VNHRFTLV epitope interferes with T cells of other specificities is not clear. The interference observed following infection is not related to the epitopes themselves or caused by a skewed T cell repertoire in heterozygote mice as F1 C57BL/6 × BALB/c mice immunized with recombinant adenovirus developed immune responses higher than the homozygote BALB/c to the H-2Kd-restricted epitope IYNVGQVSI.
A few nonmutually excluding hypotheses can be proposed. Interference could be due to the presence of CD4+CD25+ regulatory T cells. Thus, Haeryfar et al. (32) found that CD4+ CD25+ regulatory T cells selectively suppressed responses to the most immunodominant CD8+ T cell epitopes in three distinct systems. However, in a recent study, Kotner and Tarleton (33) described that depletion of these cells had little impact on the immune response of CD8+ T cells specific for T. cruzi immunodominant or subdominant epitopes. Alternatively, it could be possible that CD8+ T cells specific for VNHRFTLV inhibit expansion/proliferation of T cells of other specificities by acting directly on responder T cells. This possibility also seems unlikely because after immunization with adeno-ASP-2, heterozygote and homozygote mice presented a strong response to VNHRFTLV and also generated immune responses of a similar degree to TEWETGQI epitope. Also, during concomitant infection with two parasites strains, the immune response to VNHRFTLV did not interfere with the expansion of TsKb-18 specific cells.
The explanation that we favor is that CD8+ T cells specific for VNHRFTLV compete with T cells of other specificities in limiting APC resources during priming with T. cruzi (23, 26, 27). This hypothesis is supported by the experiment where we infected C57BL/6 mice with two parasite strains containing distinct immunodominant epitopes. In this scenario, distinct parasite strains are most likely processed by different APCs, allowing strong immune response to be developed against both epitopes.
To generate an immunodominant response, H-2Kb-VNHRFTLV complexes should be formed in a larger number, or first, leading to a rapid expansion of specific T cells. These T cells would use the available APC resources precluding or reducing activation of T cells of other specificities. The possibility that more H-2Kb-VNHRFTLV complexes are formed has no experimental support. In fact, the observation that homozygote mice develop powerful immune responses to H-2Kd- IYNVGQVSI or H-2Kk-TEWETGQI complexes suggests that the number of complexes formed on the surface of APCs from heterozygote mice should be sufficient to promote a strong T cell response. Alternatively, the number of T cell precursors and/or their affinities could be factors that accelerate the immune response causing an immunodominance to the epitope VNHRFTLV. In experiments designed to evaluate whether CD8+ T cells from infected BALB/c had a lower affinity for H-2Kd-IYNVGQVSI complexes than cells from infected C57BL/6 mice for H-2Kb-VNHRFLTV, we have found that this is not the case (Fig. 5). Therefore, it seems that the TCR affinity cannot explain the discrepant activation in heterozygote mice.
After contact with APCs, the expansion of T cells with higher affinity or more precursors for the MHC-I-peptide complex would overcome T cells specific for other parasite peptides. However, following immunization with recombinant adenovirus, we were not able to detect any interference in heterozygote mice when compared with homozygote ones, indicating that the affinity/more precursor hypotheses are not a likely explanation for the studied competition phenomenon.
Hence, in our opinion, the most plausible explanation is that this immunodominance is regulated by the timing and duration of H-2Kb-VNHRFTLV display on the surface of APCs. Earlier studies have shown that prolonged Ag exposure influences the generation of immune responses improving priming of T cells specific for immunodominant epitopes (34, 35, 36). Indeed, VNHRFTLV epitope is present in proteins expressed by trypomastigotes (infective forms) and amastigotes (intracellular stages) of the Y strain of T. cruzi (37, 38). This fact may accelerate and improve immune response by providing an advantage for T cells specific for H-2Kb-VNHRFTLV complexes on the surface of APC. In contrast, the epitopes IYNVGQVSI or TEWETGQI are expressed only by trypomastigotes (infective forms) or amastigotes (intracellular stages), respectively (38, 39). This last hypothesis would explain the differences observed following infection or immunization of heterozygote mice with adenovirus-expressing parasite Ag. Importantly, evidence exists that immunodominance may be dictated by the infective strain of T. cruzi. Thus, while expressing the VNHRFTLV epitope (21), the Brazil strain of T. cruzi induced an immunodominant response specific for TsKb-20 and TsKb-18 epitopes (8). This strain-dependent switch has no immediate explanation, but it indicates the importance of parasites in controlling the immunodominant CD8+ T cell response. In the case of the CL-Brener strain, we found during sequence analysis of the cDNA of ASP-2 that the deduced AA sequence was VNYDFTIV (amino acid substitutions are underlined). A synthetic peptide representing this epitope was not recognized by CD8+ T cells specific for the epitope VNHRFTLV (data not shown).
Although previous studies provided compelling evidence of competition between CD8+ T cells for APCs expressing simultaneously two distinct MHC-I molecules, the substrate they compete for is not yet known. This competition occurs early in the activation process of T cells (24), and involves membrane proteins required for full T cell activation and/or soluble factor(s) such as cytokines (1, 24, 40). Despite its potential as a mechanism of immunodominance, during mouse infection with lymphocytic choriomeningitis virus, this interference/competition was not apparent (29, 30). Also, ablation of immunodominant epitopes from Listeria monocytogenes did not cause an increase in the immune responses to subdominant epitopes during infection, suggesting that there was no apparent interference/competition among CD8+ T cells (31). Very recent reports suggest the opposite, that there is interference between CD8+ T cells of different MHC-I specificities during viral infection with HIV, SIV. Severe acute respiratory syndrome and vaccinia (41, 42, 43, 44). Thus, we believe that our observation reinforces the idea that a competitive mechanism for CD8+ T cell immunodominance occurs during infectious diseases with different pathogens.
Although we do not yet know precisely the mechanism used to generate this strong non-MHC specific immunodominance, we would like to suggest that it has an important biological implication for a chronic parasitic disease. In the case of self-resolving common acute viral or bacterial infections, immunodominance focuses the response on few determinants. It was suggested that this strategy successfully maximizes the power of effector immune responses and, at the same time, minimizes the risk of autoimmunity (6). In the case of T. cruzi infection (Chagas’ disease), immunodominance may have a similar function and may guarantee host survival for a long period of time. However, on the negative side, the immunodominance described here seems to hamper T. cruzi-infected mice (homozygotes or heterozygotes) in their development of stronger, and/or broader, specific immune responses. In addition, an immune response relying mainly on few immunodominant CD8+ T cell epitopes may favor the escape of immune response associated with parasite mutation or any particular antigenic variation process. An example supporting this hypothesis was described in a chronic viral infection with hepatitis C virus. A strong immunodominant human CD8+ T cells response was developed causing the selection of an escape mutant (45). Regardless, we favor the hypothesis that either restriction or avoidance of maximal host effector functions is a critical step to prevent parasite elimination, allowing parasite persistence and leading to a chronic lifelong infection.
One important immunological aspect to be considered in the rational design of a vaccine for Chagas’ disease is the pathogenesis and the intense myocarditis composed mainly of CD8+ T cells elicited during T. cruzi infection. Thus, following vaccination with plasmid DNA or recombinant proteins, we generated memory CD8+ T cells that, after a challenge with T. cruzi, displayed a very fast anamnestic immune response (7). These cells were specific for immunodominant epitopes VNHRFTLV or TEWETGQI (7). Such a strong immune response did not cause any discernible immunopathology. It in fact led to a reduced parasite development and prevented the establishment of chronic phase tissue pathologies (46, 47). Also, some of these vaccinated mice completely cleared the parasite and established sterile immunity (46, 47). These findings are also supported by our vaccination studies when we obtained reduced parasitism and limited disease development by increasing the immune responses to these same immunodominant epitopes (20, 46, 47). Thus, the development of a restricted T cell response confined to a limited number of epitopes may be important to avoid elicitation of potentially self-reactive or pathogenic T cell clones.
In summary, our results provide evidence of a competitive expansion of CD8+ T cells restricted by different MHC-I molecules. T. cruzi infection may be an interesting model to evaluate the molecular basis for CD8+ T cells competition and the biological role of immunodominance during infection with intracellular protozoan parasites, which are major causes of life-threatening chronic infection in humans. Among the protozoan parasites, CD8+ T cells have also been shown to greatly influence immunity against Plasmodium, Toxoplasma gondii, and Leishmania sp. (48, 49, 50, 51, 52, 53). Whether this competitive priming and expansion of T cells restricted by different MHC-I molecules also occurs during infection with these parasites also needs to be addressed.
We thanks Dr. Dan Hoft (St. Louis University, St. Louis, MO) for providing the H-2Kd-IYNVGQVSI tetramer and Karina Inácio Carvalho (Centro Interdisciplinar de Terapia Gênic (CINTERGEN), São Paulo, Brazil) for helping with the FACS softwares. FACS analyses were possible due to the use of equipment provided by BD Biosciences (São Paulo, Brazil). We are in debt with Dr. Victor Nussenzweig (New York University, New York, NY), Dr. Moriya Tsuji (Rockefeller University, New York, NY), Dr. George dos Reis (Federal University of Rio de Janeiro, Rio de Janeiro, Brazil), Dr. Milena Soares (Centro de Pesquisas Gonçalo Muniz/Fundação Oswaldo Cruz, Belo Horizonte, Brazil), and Dr. Esper G. Kallas (CINTERGEN-Universidade Federal de São Paulo, São Paulo, Brazil) for helpful discussion and reviewing the manuscript.
The authors have no financial conflict 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 grants from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), and The Millennium Institute for Vaccine Development and Technology (Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)-420067/2005-1), Rede Mineira de Estrutura e Função de Biomoléculas (Fundação de Amparo à Pesquisa do Estado de Minas Gerais-24000). F.T., B.C.G.d.A., and C.C. are recipients of fellowships from FAPESP. A.V.M., O.B.-R., R.T.G., and M.M.R. are recipients of fellowships from CNPq.
Abbreviations used in this paper: MHC-I, MHC class I; TS, trans-sialidase; SFC, spot forming cell; ASP-2, amastigote surface protein 2.