Selection of Similar Naive T Cell Repertoires but Induction of Distinct T Cell Responses by Native and Modified Antigen¹

Upon cognate interaction with APCs, Ag-specific naive CD4⁺ T cells differentiate into two subsets, characterized by distinct patterns of cytokine secretion. Th1 cells secrete IFN-γ, TNF-α, and IL-2, and induce, in the mouse, B cell switching to IgG2a and IgG2b isotype production. Th2 cells secrete, among other cytokines, IL-4 and IL-5, and promote a B cell switch to IgG1 and IgE (1). Several factors contribute to determining the Th1 or Th2 dominance. The type of Ag and the adjuvant used (2, 3, 4), the dose and route of administration (5, 6, 7, 8), and the type of APC involved in T cell priming (9) all play a role in the polarization of T cell responses. Other key factors are the genetic background (10), the affinity of antigenic determinants for the MHC and the TCR (11), the previous immunological history (12), and the cytokine and chemokine milieu in which T cell priming occurs (13).

The factors influencing the strength of the TCR-ligand interaction and their relationship to the intracellular signals that drive the differentiation to Th1 or Th2 cells are now becoming better understood (14). A high density of Ag-MHC complexes on the surface of the APC, owing to high affinity of the peptide for the MHC molecule (15), or to a TCR with high affinity for a given Ag-MHC complex would favor the generation of a Th1 response. Conversely, TCR-Ag/MHC interactions characterized by a lower affinity or a low ligand density promote Th2 differentiation (11), possibly by limiting the duration of peptide/MHC-TCR interactions (16). Altered peptide ligands, peptides modified in their TCR-contacting residues, have been used both in human and model studies for the therapy of autoimmune diseases, and have been shown to modulate the Th phenotype in vivo by acting as TCR antagonists, as partial agonists, or by inducing regulatory T cells (17). Modification of side chains in antigenic proteins has also been shown to alter the Th cell phenotype of the immune response (18).

The polarization of pathogen-specific CD4 T cell responses strongly influences the resistance and susceptibility to many infectious diseases, and inappropriate vaccine strategies may exacerbate rather than protect against subsequent infection, as shown in individuals vaccinated with formalin-inactivated respiratory syncytial virus (19). Thus, understanding how the form of Ag may influence the polarization of T cell responses could be useful in vaccine development, in which the ability to selectively induce Ag-specific memory/effector T cells of the appropriate phenotype would be advantageous. Indeed, previous work has shown that chemically modified protein Ag can be used to selectively elicit Th1-dominated responses (20). Because B cells have been implicated in the induction and maintenance of Th2 responses (21, 22), designing recombinant protein vaccines that lack B cell cross-reactivity with native Ag could favor the induction of effector/memory Th1 cells. We have tested this possibility in the hen egg-white lysozyme (HEL)³ model, in which the reduced and carboxymethylated form of this protein (RCM-HEL) induces a cross-reactive T cell, but a noncross-reactive B cell response to the native HEL (23).

In the present study, we show that HEL and RCM-HEL prime the same population of specific T cells, characterized by the TCR rearrangement Vβ8.2-Jβ1.5, but induce distinct Th1 responses. The enhanced Th1 response induced by RCM-HEL compared with HEL may be relevant for vaccine design.

Two-month-old female BALB/c, C3H, and CBA mice (Charles River Breeding Laboratories, Calco, Italy) were used. HEL was purchased from Sigma-Aldrich (St. Louis, MO) and recrystallized three times. RCM-HEL was prepared by reduction of HEL in a 8 M urea with 100-fold molar excess of 2-ME, followed by treatment with iodoacetic acid (23). Synthetic peptides were prepared, as previously described (24).

Mice were immunized s.c. into the hind footpads with different amounts (from 0.01 to 10 nmol/mouse) of native HEL or RCM-HEL, emulsified in CFA or IFA (Difco, Detroit, MI), or in water. Popliteal lymph nodes were harvested 8–12 days after immunization, and lymph node cells (LNCs) were seeded in 96-well plates (Costar, Cambridge, MA) at 5 × 10⁵ cells/well in the presence of graded concentration of Ag. In all experiments, the culture medium was RPMI 1640 (Life Technologies, Basel, Switzerland), supplemented with 2 mM l-glutamine, 50 μM 2-ME, 50 μg/ml gentamicin (Sigma-Aldrich), and 10% FCS (Life Technologies). Forty-eight hours later, Ag-specific T cell proliferation was assessed by [³H]thymidine incorporation.

Cytokine concentration was quantified in pooled supernatants from triplicate cultures, as previously described (7, 25). All mAbs, except anti-IFN-γ, were purchased from BD PharMingen (San Diego, CA). The anti-IFN-γ mAb used for capture was AN.18.17.24, and the mAb used for detection was peroxidase-conjugated XMG1.2, as previously described (25). In some experiments, IL-2 was also measured by measuring the proliferation of IL-2-sensitive CTLL-2 cells, seeded at 5 × l0⁴ cells/well in 96-well plates for 24 h. Cytokines were quantified in duplicate from three to four titration points using standard curves generated by purified recombinant mouse cytokines (IFN-γ and IL-4 from Hoffmann-LaRoche (Nutley, NJ), and IL-2 from BD PharMingen).

IFN-γ- and IL-4-producing cells were enumerated by a cellular ELISPOT assay, as described (26). Briefly, LNCs (5 × 10⁶/ml) were cultured for 48 h in 24-well plates with the indicated Ag. Millititer hemagglutinin nitrocellulose plates (Millipore, Bedford, MA) were coated overnight at 4°C with anti-IFN-γ or anti-IL-4 Ab. Plates were blocked, and Ag-stimulated cells were added at different concentrations for 24 h at 37°C. The wells were then incubated with biotin-conjugated anti-IFN-γ or anti-IL-4 Ab, followed by incubation with avidin-peroxidase (Vector Laboratories, Burlingame, CA). Spots were developed by the addition of 400 μg/ml 3-amino-9-ethylcarbazole substrate (Sigma-Aldrich) and enumerated by a computerized image analysis system (Lightools Research, Encinitas, CA) using the image analyzer program National Institutes of Health Image 1.61 (National Institutes of Health, Bethesda, MD).

Abs specific for HEL and RCM-HEL were quantified in mouse serum, as described (7). Briefly, mice were immunized, as described above, and bled 15 days later. HEL- or RCM-HEL-coated polyvinyl microtiter plates (Falcon 3012) were incubated with serially diluted sera in PBS-Tween for 90 min at 37°C. Plates were then washed and incubated with a mixture or individual biotin-conjugated, goat anti-mouse Abs specific for IgM, IgG1, IgG2a, IgG2b, and IgG3 isotypes (100 ng/ml each). The bound Abs were revealed with alkaline phosphatase-conjugated streptavidin (Jackson ImmunoResearch Laboratories, Avondale, PA). Standard curves were generated using pooled anti-HEL antisera, and the results are expressed as arbitrary units per milliliter (U/ml, 1 U corresponding to 50% of maximum OD).

Repertoire analysis was performed using a modification of a previously described protocol (27). LNCs were cultured for 3 days in HL-1 medium (BioWhittaker, Walkersville, MD) in the presence or absence of HEL or RCM-HEL. Total RNA was isolated from cell suspensions using RNeasy Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer’s instructions. cDNA was synthesized using an oligo(dT) primer (dT15) (Life Technologies). From each cDNA, PCR were then performed using a Vβ8.2 primer (CATTATTCATATGGTGCTGGC) and a common Cβ primer (CACTGATGTTCTGTGTGACA). Using 2 μl of this product as a template, runoff reactions were performed with a single internal fluorescent primer for each Jβ tested. These products were then denatured in formamide and analyzed on an Applied Biosystems (Foster City, CA) 310 Prism using Gene-scan 2.0 software.

Public Vβ8.2-Jβ1.5 rearrangement quantification was performed by real-time quantitative RT-PCR using an ABI Prism 5700 sequence BioDetector (Applied Biosystems), as previously described (28, 29). Briefly, total RNA was extracted and reverse transcribed. Primers and probes included: Vβ8.2 sense, 5′-ATCCATTATTCATATGGTGCTGGC-3′; Jβ1.5 antisense, 5′-AGTCCCCTCTCCAAAAAGCG-3′; Vβ8.2-Jβ1.5 complementarity-determining region 3 (CDR3)-specific probe FAM-5′-GCGGTACAGGGAACAACCAGGCT-3′-TAMRA; TCR β-chain antisense, 5′-GTGAGCCCTCTGGCCACTT-3′; TCR β-chain probe, FAM-5′-TCCTCCTGTGAAAGCCCATGGAACTG-3′-TAMRA. Transcripts for TCR β-chain CDR3 motifs and for TCR β-chain C regions were determined in the same samples for normalization. Data were expressed as the ratio of public CDR3 over TCR β-chain mRNA copies.

The Ag-presenting activity of lymph node APCs was measured using as readout the T cell hybridomas 1H11.3 (I-E^d, HEL108–116) (30) and 2G12.1 (I-A^d, β₂-microglobulin 26–39) (24). APC populations were enriched from popliteal LNCs, as previously described (12, 25). Briefly, T cell-depleted LNCs were separated into low and high buoyancy density populations by centrifugation over a discontinuous Percoll (Pharmacia Biotech, Uppsala, Sweden) gradient containing a 55% layer. High buoyancy cells were shown to contain small B cells. Low buoyancy cells were separated by MiniMACS columns (Miltenyi Biotec GmBH, Bergisch Gladbach, Germany) using CD11c-specific microbeads, coated with the Ab N418. The negatively separated fraction was enriched in large B cells, while the positive fraction was enriched in N418⁺ dendritic cells (DCs). Alternatively, the low buoyancy cells were first sorted with MiniMACS columns using CD8α-specific microbeads to enrich for lymphoid DCs. The negative fraction was then purified with CD11c-specific microbeads to obtain a population enriched in myeloid DCs. The APC population enrichment was checked by cytofluorimetry using a FACScan flow cytometer (BD Biosciences, Mountain View, CA) equipped with Lysis II software. CD11c positively sorted cells were between 85 and 95% N418⁺ cells, in agreement with previous data (12, 25). Low buoyancy, CD8α-selected populations usually contained 65% N418⁺ cells.

Irrespective of the form of Ag used for immunization, HEL and RCM-HEL are completely cross-reactive in the restimulation of T cell proliferation in vitro (Fig. 1,A), in agreement with previous data (23). In the BALB/c mouse, the T cell response of mice immunized with HEL is dominated by the recognition of the immunodominant determinant located in the region 108–116, with a subdominant epitope spanning residues 9–25 (30). The T cell response of LNCs from BALB/c mice immunized with RCM-HEL shows a fine specificity superimposable to that induced by HEL (Fig. 1,B), suggesting the engagement of overlapping T cell populations by both Ag forms. Notably, no additional HEL epitopes are recognized following priming of H-2^d mice with RCM-HEL, and the fine specificity of the response to HEL is maintained irrespective of the carboxymethylation at Cys¹¹⁵ of the dominant epitope in the priming Ag, whereas the subdominant epitope located in the region 9–25 does not contain any Cys residue, and therefore is not modified by reduction and carboxymethylation. Conversely, Abs against HEL fail to recognize RCM-HEL and vice versa (Fig. 1 C), as expected (23).

LNCs from mice immunized with HEL secrete lower IFN-γ and higher IL-4 levels compared with cells from RCM-HEL-primed mice following in vitro restimulation with either HEL or RCM-HEL (Fig. 2,A). This is paralleled by a significantly lower number of Ag-activated IFN-γ-producing cells and a significantly higher number of IL-4-producing T cells in mice immunized with HEL compared with RCM-HEL, as determined by ELISPOT assays (Fig. 2,B). The distinct pattern of cytokine secretion is mirrored by the isotype of Ag-specific serum Abs (Fig. 2 C). Anti-HEL Abs are predominantly of IgG1 isotype, while this isotype is much less abundant among RCM-HEL-specific Abs. Abs of IgG2a isotype are relatively low in both cases. The difference in IgG1 isotype distribution further indicates that denaturation of HEL skews the in vivo response toward the Th1 pathway.

The enhanced Th1 response induced by RCM-HEL compared with HEL is independent from the dose of Ag and the adjuvant used for priming. In the experiments reported in Fig. 3, we compared the responses to HEL and RCM-HEL obtained by immunizing BALB/c mice with different doses of Ag either emulsified in different adjuvants (CFA or IFA) or administered without adjuvant, in soluble form. In all cases, enhanced IFN-γ and reduced IL-4 production, indicative of a skewing to the Th1 pattern, could be observed in LNCs from mice immunized with RCM-HEL compared with HEL and restimulated in vitro with HEL, irrespective of the adjuvant used and even in the absence of adjuvant. This pattern of cytokine secretion was consistent with the drastically reduced (>90%) levels of Ag-specific IgG1 Abs (data not shown). It could be suggested that the chemical modification of HEL results in the generation of additional epitopes that preferentially drive Th1 responses. However, HEL and RCM-HEL are completely cross-reactive in the restimulation of T cell proliferation in vitro (Fig. 1,A). In addition, the epitope mapping reported in Fig. 1 B shows that the epitopes generated by HEL and RCM-HEL are superimposable, without new HEL epitopes generated by priming with RCM-HEL, supporting the view that the modification of the Th1/Th2 response does not reflect an epitope shift. An enhanced Th1 response to RCM-HEL compared with HEL was also observed in C3H and CBA mice, both of H-2^k haplotype, as demonstrated by the increased IFN-γ secretion as well as by the selective production of RCM-HEL-specific Abs of the IgG2a isotype (data not shown).

The results described above may suggest that RCM-HEL is presented by an APC type able to drive naive T cells preferentially toward the Th1 phenotype. Reduction and carboxymethylation of HEL profoundly alter its three-dimensional structure, as shown by the inability of HEL-specific Abs to recognize RCM-HEL. Importantly, it adds eight new carboxymethyl groups that strongly modify its solubility. These modifications might favor presentation of RCM-HEL by APCs different from those involved in the presentation of HEL, as hypothesized to explain the shift toward Th1 obtained by maleylation of protein Ags (18).

By using the HEL108–116/I-E^d-specific T cell hybridoma 1H11.3 as a readout, both HEL and RCM-HEL injected s.c. into BALB/c mice were found to be presented in the draining lymph nodes predominantly by DCs rather than B cells (Fig. 4, A and B), consistent with our previous observations on HEL presentation (12, 25). Mouse CD8α⁺ and CD8α⁻ DCs have been reported to selectively induce Th1 and Th2 cells, respectively (31). Therefore, we examined whether RCM-HEL was preferentially presented in vivo by CD8α⁺ DCs. Mice were injected s.c. with 10 nmol of either HEL or RCM-HEL emulsified in IFA, and 5 days later the presentation of antigenic complexes by enriched populations of CD8α⁺ or CD8α⁻ DCs from the draining lymph nodes was compared. By using the procedures described in Materials and Methods, CD8α⁻ DCs were enriched to 85 or 95% N418⁺ cells, in cells from RCM-HEL- and HEL-primed mice, respectively, whereas the CD8α⁺ fraction contained 65% N418⁺ cells, both in RCM-HEL- and HEL-primed mice. Both CD8α⁺ and CD8α⁻ DCs, from either HEL- or RCM-HEL-primed mice, activated the 1H11.3 T cell hybridoma. The lower T cell activation by APCs bearing antigenic complexes derived from RCM-HEL compared with HEL reflects the lower efficiency of RCM-HEL to activate the readout hybridoma (data not shown), rather than the absolute number of peptide-MHC complexes displayed by APCs. CD8α⁻, compared with CD8α⁺ DCs, were ∼3-fold more efficient APCs, irrespective of the Ag form administered (Fig. 4, C and D). In contrast, the intrinsic Ag-presenting capacity of the two DC subsets was similar, as shown by their ability to present equally well the endogenous, naturally processed β₂-microglobulin peptide to the T cell hybridoma 2G12.1 (Fig. 4, E and F).

The T cell response to HEL in the BALB/c mouse is consistently associated with the expansion of a public TCR repertoire characterized by the Vβ8.2-Jβ1.5 rearrangement, specific for HEL107–116/I-E^d (32). This repertoire has been shown to be specifically associated with Th1 cells in mice immunized with HEL in CFA (27, 33). Therefore, if HEL and RCM-HEL activate overlapping populations of T cells, the same public repertoire should be expanded by immunization with both Ags. To answer this question, the expansion of the public Vβ8.2-Jβ1.5 clonotype was determined in these different Ag-stimulated CD4⁺ T cell populations by CDR3-length spectroscopy, and quantified by fluorogenic PCR (28, 29) directly ex vivo or after in vitro stimulation.

Using CDR3 fragment length analysis, a similar expansion of the public repertoire was observed in mice primed with HEL or RCM-HEL. Fig. 5,A shows the BV8S2-J1S5 CDR3 length distribution obtained in one representative mouse (of six mice/group tested) ex vivo or after in vitro restimulation with HEL, RCM-HEL, or the peptide encompassing the immunodominant epitope 103–117. Irrespective of the form of Ag used for immunization, the BV8S2-J1S5 public clonotypic cells are expanded equally well by every form of Ag tested. To quantify more precisely the number of clonotypic cells, we used a quantitative fluorogenic PCR. As shown in Fig. 5 B, HEL restimulation of LNCs from mice immunized with RCM-HEL or HEL resulted in the expansion of the public clonotype to a similar frequency. Notably, RCM-HEL supported to the same extent the expansion in vitro of clonotypic cells obtained from mice immunized with HEL or RCM-HEL. Surprisingly, native HEL and the peptide 103–117 stimulated the proliferation of T cells characterized by usage of the public Vβ8.2-Jβ1.5 clonotype more effectively when the T cells were obtained from mice immunized with RCM-HEL. Therefore, T cells obtained from mice immunized with RCM-HEL appear to have a peculiar ability to cross-recognize heteroclitically native HEL and the derived peptide 103–117, indicating that immunization with a structurally altered Ag can generate an immune response that is more effective in the recognition of the native Ag than priming with the native Ag itself. Alternatively, the increased proliferation could be explained by the enhanced Th1 differentiation induced by RCM-HEL. Although naive T cells bearing the HEL-specific Vβ8.2-Jβ1.5 public rearrangement can be efficiently primed in vivo by RCM-HEL, we cannot exclude that LNCs from RCM-HEL-immunized mice may also contain distinct population(s) of CD4⁺ T cells that specifically recognize the modified protein.

The lack of cross-reactivity between Abs against HEL and RCM-HEL offers the possibility to follow in vivo, in mice simultaneously immunized with both Ag forms, the Ab isotypes specific for each of them. Mice were immunized with a fixed amount of HEL (0.1 nmol/mouse) together with RCM-HEL at ratios varying from 1:10 (1 nmol) to 10:1 (0.01 nmol), or with the same amounts of RCM-HEL alone. One group of HEL-primed mice was simultaneously treated with IL-12. Fifteen days later, mice were bled, and IgG1 and IgG2a Abs, specific for either HEL or RCM-HEL, were measured (Fig. 6,A). Mice that had received 0.1 nmol of HEL and 0.01 nmol of RCM-HEL did not respond to RCM-HEL, as mice immunized with 0.01 nmol of RCM-HEL alone, and the expected IgG1-IgG2a ratio of 12 was observed in response to HEL. Mice primed with HEL plus IL-12 had a response dominated by IgG2a (IgG1-IgG2a ratio 0.39), as expected (34). When the amounts of HEL and RCM-HEL administered were sufficient, the generation of two independent responses was observed. The response to HEL was clearly dominated by IgG1 (IgG1-IgG2a ratio >11), independently from the amount of RCM-HEL coinjected. Similarly, the response to RCM-HEL retained a Th1-dominated phenotype, without a significant increase in the RCM-HEL-specific, Th2-dependent IgG1 Ab response (Fig. 6 A).

In contrast with the two independent Ab responses induced in naive mice, the presence of an already established Th1 response specific for RCM-HEL is able to divert the primary Ab response to HEL from a Th2 to a Th1 isotype pattern. As shown in Fig. 6 B, the B cell response to HEL in mice previously immunized with RCM-HEL is significantly skewed to the Th1 phenotype, reflecting the pre-existing RCM-HEL-specific response. Therefore, primed T cells specific for RCM-HEL are restimulated by HEL-derived peptide/MHC complexes, presented mostly by B cells (35), secrete Th1-type cytokines, and promote an Ig switch to Th1-dependent isotypes.

We have analyzed the effects of a structural modification in a protein Ag on TCR repertoire selection and developmental pathways of CD4⁺ cells. RCM-HEL, despite a pronounced alteration of its three-dimensional structure, was presented in vivo by the same APC populations as HEL, induced in vivo the expansion of the public T cell clonotype characterized by the BV8S2-J1S5 rearrangement, typical of HEL107–116-specific CD4⁺ T cells, but elicited a response further polarized to the Th1 pathway compared with HEL.

The ability of some Ags to preferentially induce a Th1 or Th2 response has been attributed either to epitope density on the APCs or to their ability to engage different APC types. Among professional APCs, DCs have been shown to be critical for T cell priming (36, 37) and, in the mouse, CD8α⁻ DCs selectively prime Th2 and CD8α⁺ DCs Th1 cells (31, 38). We could not observe any difference in the presentation by DC populations by the two Ag forms, nor could we detect an increase of IL-12 secretion by DCs from mice immunized with RCM-HEL (F.R., unpublished results). Thus, HEL and RCM-HEL appear to be presented by the same APCs in vivo, even though they differ dramatically in solubility. However, we cannot exclude that low numbers of B cells (below our detection limit) may present native HEL in naive mice challenged with HEL. Although it has been shown that B cells are not necessary to induce Th2 cell development (37); a low number of B cells might selectively promote the development of Th2 response (20) owing to the lack of IL-12 secretion (39), the failure to sustain CD154 expression (40), or the promotion of Th2 cell survival via OX40-OX40 ligand interaction (22).

Once activated, RCM-HEL-specific T cells are effectively stimulated in vivo by HEL-derived peptides. This in vivo observation confirms the in vitro data, showing that T cell populations obtained from mice primed with HEL or RCM-HEL display a complete cross-recognition of the two Ag forms. The development of cross-reactivity after priming can be due to preferential selection of cross-reactive T cells by the second Ag form, as indicated by the equal expansion of the public clonotypic CDR3 sequence Vβ8.2-Jβ1.5 after priming with HEL or RCM-HEL. Identification of T cells using this CDR3 in response to both forms of Ag offers molecular evidence for the possibility that such cross-reactive T cells may be selectively activated upon challenge. The cross-activation of primed T cells could also be favored by functional and biochemical changes distinguishing naive and activated T cells, including requirements for costimulation (41, 42), as well as membrane organization and redistribution of signaling molecules within the immune synapse (43, 44). The result of these modifications is that the TCR avidity for peptide-MHC complexes is considerably higher in activated compared with naive T cells (45), possibly also due to increased cross-linking of the TCR (46), a mechanism that may be involved in memory maintenance (47).

Most importantly, the clonotypic cells appear functionally different, depending on the form of Ag used for priming. Cells obtained from mice primed with RCM-HEL display a proliferative response to HEL103–117 and, surprisingly, to native HEL stronger than the one exhibited by cells bearing the same clonotypic CDR3-β region, but obtained from mice primed with native HEL itself. It would be interesting to establish whether this heteroclitic response reflects the reorganization of membrane microdomains upon activation of the same TCR by distinct ligands. Although the CDR3 of the TCR β-chain engages most of the amino acid residues that are relevant for Ag recognition, the CDR3 region of the α-chain is also involved in determining the overall affinity of the TCR for peptide-MHC complexes and the flexibility of the TCR itself (48). Thus, although T cells obtained from mice immunized with HEL or RCM-HEL share identical TCR β-chains, they might differ in TCR α-chain usage, resulting in an altered affinity for the same immunodominant epitope. Indeed, it has been shown recently that resistance to Leishmania major infection in mice transgenic for the TCR β-chain of the Leishmania homologue of receptors for activated C-kinase Ag was associated with the selective expansion of high affinity CD4 T cells expressing a distinct TCR α repertoire and secreting high levels of IFN-γ (49).

The observation that irreversible disruption of the three-dimensional structure of a globular protein Ag, with defined modifications in TCR-contacting residues such as the addition of carboxymethyl groups to Cys¹¹⁵ in RCM-HEL, can lead to enhanced T cell proliferation and to a response skewed to the Th1 phenotype may be relevant for vaccine design. We could observe these effects independently from the adjuvant used, an important advantage for human vaccination. Indeed, reduction of disulfide bonds and carboxymethylation of cysteine residues in either intact protein or peptide has been shown to be critical for efficient T cell stimulation (50), and by studying the activation of several T cell hybridomas specific for HEL108–116/I-E^d we could observe that the effect of carboxymethylation at Cys¹¹⁵ varies depending on the T cell hybridoma tested (F.R. and J.-C.G., unpublished observations). Of note, one of these T cell clones that was modestly affected by the Cys¹¹⁵ modification expressed the BV8S2-J1S5 public TCR β-chain (33), consistent with our observation that HEL108–116/I-E^d-specific T cells from mice primed with RCM-HEL or HEL can share identical TCR β-chains. Our demonstration that in vivo induced bulk T cells specific for the structurally modified Ag can provide stable Th1-biased help for B cells may be particularly relevant, because the Th1-dependent Ig isotypes are the most efficient in mediating complement activation and Ab-dependent cellular cytotoxicity, key components in the protective response to pathogens. For example, immunity to the malaria sporozoite requires a Th1 response and is transferred by serum (51), most likely involving Th1-dependent isotypes.

In conclusion, a structurally modified Ag can induce an immune response qualitatively and quantitatively different from its native counterpart, although both Ag forms are recognized by a T cell population expressing the same public clonotype. This indicates that appropriate structural modifications of Ags, favoring the induction of desired responses, could improve the efficacy of vaccination and of Ag-based immunointervention in the treatment of autoimmune or allergic diseases.