Immunodominance refers to the restricted peptide specificity of T cells that are detectable after an adaptive immune response. For CD4 T cells, many of the mechanisms used to explain this selectivity suggest that events related to Ag processing play a major role in determining a peptide’s ability to recruit CD4 T cells. Implicit in these models is the prediction that the molecular context in which an antigenic peptide is contained will impact significantly on its immunodominance. In this study, we present evidence that the selectivity of CD4 T cell responses to peptides contained within protein Ags is not detectably influenced by the location of the peptide in a given protein or the primary sequence of the protein that bears the test peptide. We have used molecular approaches to change the location of peptides within complex protein Ags and to change the flanking sequences that border the peptide epitope to now include a protease site, and find that immunodominance or crypticity of a peptide observed in its native protein context is preserved. Collectively, these results suggest immunodominance of peptides contained in complex Ags is due to an intrinsic factor of the peptide, based upon the affinity of that peptide for MHC class II molecules. These findings are discussed with regard to implications for vaccine design.
When individuals are immunized with a complex Ag, most of the responding T cells are specific for a few of the many potential peptides contained in the Ag. These peptide determinants are termed immunodominant. Considering the diversity in the TCR repertoire and the number of peptides contained in complex proteins, it is remarkable that T cells apparently react to such a limited number of peptide epitopes, while seemingly ignoring the others. Many studies have found that complex proteins have many peptides that can bind to the MHC molecule and elicit T cell responses when administered as single peptides. These determinants, whose response is only detected after peptide immunization, are termed cryptic because they remain sequestered from the response after immunization with the intact proteins. Using a number of experimental model Ags, significant effort has been dedicated toward understanding the mechanisms that underlie immunodominance in CD4 (reviewed in Refs. 1, 2, 3, 4, 5, 6) and in CD8 (7, 8) T cell responses. There still remains considerable controversy regarding this issue, which may in part be due to conflicts in assay systems used, whether responses to extracellular proteins or pathogens that replicate within APC are studied, and whether one is analyzing responses to autoantigens or proteins completely foreign to the host. Despite this, there is undisputed value in being able to predict the major epitopes that will be recognized in the host T cell response to antigenic challenge. The unpredictability in the immunodominance hierarchy, especially when comparing outcomes between vaccinations and natural immunity, leads to significant impediments in our ability to rationally design vaccines (9).
A significant body of literature (4, 8, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20) has suggested that events related to Ag processing, such as tertiary structure of the Ag, protease susceptibility in the residues that flank the peptide, the localization of the peptide in the Ag, and competing peptides within the Ag, are major determinants of immunodominance. One may group all of these events as related to the context of the peptide. There are some experimental data that support each of these. For example, APC treated with inhibitors of endosomal proteolysis (21, 22, 23, 24) or APC deficient in defined proteases (25, 26) exhibit either enhanced or inhibited Ag presentation, depending on the epitope, whereas mutation of a flanking sequence by introduction of a protease recognition sequence can enhance epitope display (27) or modify the ability to recruit CD4 T cells (28). These data suggest that proteolytic release of a peptide may limit its epitope density with class II molecules at the cell surface, and thus, the recruitment of T cells during an immune response. Similarly, recent studies have observed that immunodominant epitopes cluster in limited regions of the Ag, often within solvent-exposed regions or at sites that are adjacent to protease-sensitive loops (29, 30, 31). These types of data, which are largely correlative in nature, suggest that Ags that have a high degree of tertiary structure may be poorly immunogenic, or peptides buried within protease-resistant regions of an Ag may sequester peptides from proteolytic release and subsequent association with class II molecules. Conversely, some studies have suggested that CD4 T cell epitopes are preferentially associated with structurally stable regions (32).
In contrast to the above studies that suggest location and thus proteolytic release determine immunodominance hierarchies, our laboratory has recently provided evidence that a biochemical property of the peptide:class II complex, its spontaneous kinetic stability, is a critical parameter that determines the immunodominance of that peptide:class II epitope in the developing immune response (3, 33). For these studies, we measured off-rates of peptides from the class II molecule as the most accurate measure of affinity of peptide:MHC class II and the only measure of affinity that has been identified that correlates with immunodominance. Additionally, we have shown that one of the key mechanisms that underlies this linkage between the kinetic stability and immunodominance of peptide:class II complexes is DM editing within APC (3, 34) because it strongly favors presentation of high stability complexes over those with lower stability. Our studies thus suggest it is the biochemistry of the peptide:class II complex that plays a critical role in determining immunodominance. Accumulated data from several laboratories have lent support to the view that the functional effects of DM can shape immunodominance hierarchies (6, 35, 36, 37, 38).
Despite these advances, it is imperative to more fully understand the molecular events that control immunodominance to complex Ags and in particular to more rigorously evaluate the role of the molecular context of a peptide in a developing immune response. This issue is important not only because of the value in enhanced understanding of the forces that drive the specificity of the adaptive immune response and central tolerance induction, but also from a practical standpoint. There have been significant new advances in epitope discovery that have allowed identification of CD4 and CD8 peptide epitopes from tumors, viruses, and other pathogenic organisms (1, 9, 39, 40, 41, 42, 43). Using both bioinformatics and peptide-scanning approaches, there is increasing promise in such strategies as multiepitope vaccines (43, 44, 45, 46, 47, 48, 49), in which candidate pathogen-derived or tumor-derived peptides of interest are linked together to make recombinant multiepitope proteins. Such strategies, frequently intended to target both CD8 and CD4 T cells, have been developed to enhance responses to tumors and pathogenic organisms. Also, advances in identifying some of the critical proteins that mediate uptake of Ags in dendritic cells, such as DEC-205 and dendritic cell-specific intracellular adhesion molecule-3-grabbing nonintegrin, have initiated efforts to target antigenic peptides to the appropriate dendritic cell by incorporating the peptides into Abs that engage these cell surface proteins (50, 51, 52, 53, 54). Despite the promise in these advances in epitope discovery and in peptide delivery, it is not clear whether the immunodominance patterns of these antigenic peptides will be preserved in a new protein context, or conversely, if the patterns of immunodominance will be re-established in each new protein.
In our previous studies, we observed the direct correlation of immunodominance and the kinetic stability of the peptide:MHC complexes when inserting multiple heterologous peptides into a single site in a shuttle protein. In the present study, we sought to explicitly evaluate whether such factors as peptide location within an Ag, competing peptides within the Ag, and/or differences in sensitivity of the peptide to the endosomal proteases that liberate it from the complex Ag are a deciding factor in establishing immunodominance hierarchies. The data presented in this study, using a comprehensive and systematic approach to change the context of peptide Ags, suggest that immunodominance hierarchies are largely independent of the molecular context in which they are carried, and therefore are intrinsic to the peptide:class II complex that elicits the CD4 T cells.
Materials and Methods
MalE protein purification
MalE constructs were designed and purified, as previously described (33). Briefly PAGE-purified synthetic oligonucleotides encoding the desired peptide were ligated into BamHI-digested MalE133, XhoI-digested MalE206, or BamHI-digested MalE303 vector. Sequenced clones were transformed into MalE (−/−) ER2507 Escherichia coli and expanded in culture, and the MalE was isolated after osmotic shock and purification using amylose columns (55), as described. Collected fractions containing MalE were identified by Bradford assay, and then pooled, dialyzed, concentrated, filter sterilized through a 0.2 μM syringe filter, quantified by Bradford assay (Bio-Rad) and SDS-PAGE, and stored at −20°C.
BALB/c mice (National Cancer Institute) were immunized in the hind footpad with 50 μl of 0.2 mg/ml protein Ags emulsified in CFA (Sigma-Aldrich). Ten days later, draining lymph nodes were harvested, pooled, and depleted with MKD6 (anti-I-Ad), M5/114 (anti-I-Ad, anti-I-Ed), and 3.155 (anti-CD8) Abs to enrich for CD4 T cells. CD4 T cells were stimulated with freshly isolated, T cell-depleted splenocytes, and IL-2 production was measured by an ELISPOT assay (56). The mean number of spots for each condition was determined in triplicate. All animal handling was performed according to the regulations set by the University Committee on Animal Care at the University of Rochester.
Abs and peptides
Purified rat anti-mouse IL-2 (clone JES6-1A12) and biotinylated rat anti-mouse IL-2 (clone JES6-5H4) Abs were obtained from BD Pharmingen. The hybridomas producing mAbs MKD6 (57), M5/114 (58), 3.155 (59), and J1j.10 (60) were acquired from the American Type Culture Collection. Synthetic peptides were obtained from BioPeptides.
Ag presentation assays
T cell hybridomas were derived and maintained, as previously described (61). In Ag presentation assays, 5 × 105 (splenocytes) or 5 × 104 (A20) APC were co-cultured with 5 × 104 T cell hybridomas in 0.2 ml in a 96-well flat-bottom plate with increasing concentrations of Ag. After 16–20 h, 50 μl of supernatant was removed, frozen, thawed, and then tested for IL-2 content using the CTLL indicator cell and an MTT assay to detect viable cells after a 16-h culture. Results appear as the mean OD in the MTT assay from triplicate wells read at 570–650 nm on a Vmax plate reader.
The proteins Leishmania homolog of receptors for activated C kinase (LACK)3 (a gift from D. Fowell, University of Rochester, Rochester, NY), MalE, OVA (Sigma-Aldrich), or hen egg lysozyme (HEL; Sigma-Aldrich) were diluted and transferred to 0.2 M phosphate/0.1 M citrate MacIlvaine buffers at a pH range between 4.0 and 7.0, at a final concentration of 2 μM, and incubated with 60 μM 8-anilino-1-naphthalenesulfonic acid (ANS; Sigma-Aldrich) for 10 min at room temperature. Measurement of bound fluorescent ANS was performed in a FLUOROMAX (Horiba Jobin Yvon), exciting the probe at 350 nm, and the fluorescence intensity was measured between 490 and 600 nm. Where indicated, the proteins were reduced in the presence of 1 μM DTT (Sigma-Aldrich) for 10 min before incubation with ANS probe.
The immunodominance of peptides track with their primary sequence rather than the protein in which they are contained
In the first series of experiments, we evaluated the role of a peptide’s molecular context by comparing its immunogenicity in vivo when contained within its native protein vs that displayed when it is encoded within a heterologous protein. As previously described (62), the MalE gene encodes a subunit of the E. coli maltose-binding protein and has several permissive sites that accept peptide inserts of 10–40 aa without changing its overall conformation. We used aa 133 as the site of insertion for the test peptides, which is known to be solvent exposed and to readily accommodate heterologous sequences (Table I). MalE is an attractive protein vector because it allows insertion of antigenic peptides in a new context and enables concurrent analyses of responses to dominant MalE peptides. In our experiments, peptides that were inserted within MalE contained several flanking residues of the native sequence on the N and C terminus to preserve potential TCR contact residues (63, 64). An independent set of peptides with previously characterized immunodominance hierarchies was compared in these experiments, including the immunodominant peptides Myo (102–118) (65); LACK (161–173) (66); the subdominant OVA epitope (273–288) (33, 67); the cryptic peptides HEL (20–35), HEL (11–25) (68, 69), and OVA (328–339) (33, 67, 70); and a modified cryptic form of the LACK (161–173) peptide with a substitution at the P4 pocket (163 I>A) that reduces its kinetic stability with I-Ad and correspondingly its immunodominance (33). BALB/c mice were immunized with these Ags either in their native context or encoded in MalE, and CD4 T cells were isolated from the draining lymph nodes 9–10 days after priming. CD4 T cells were analyzed for the number of IL-2 spots elicited in response to synthetic peptides or intact Ag to evaluate the number of CD4 T cells recruited into the response. The ELISPOT assay offers considerable advantages in direct enumeration of Ag-specific T cells immediately ex vivo, without any need for T cell proliferation or persistence in culture.
|Protein .||Region .||ID .||Solvent Accessibility .||Competing Peptidesa .||PDBb .|
|Sperm whale myoglobin||(102–118)||Dominant||Exposed||None known||1VXG|
|Insertion site||133||Exposedd||(69–84) (102–115) (269–285)|
|206||Exposedd||(69–84) (102–115) (269–285)|
|303||Buriedd||(69–84) (102–115) (269–285)|
|Chicken egg OVA||(323–339)||Dominant||Buriedc||(273–288)||1UHG|
|Protein .||Region .||ID .||Solvent Accessibility .||Competing Peptidesa .||PDBb .|
|Sperm whale myoglobin||(102–118)||Dominant||Exposed||None known||1VXG|
|Insertion site||133||Exposedd||(69–84) (102–115) (269–285)|
|206||Exposedd||(69–84) (102–115) (269–285)|
|303||Buriedd||(69–84) (102–115) (269–285)|
|Chicken egg OVA||(323–339)||Dominant||Buriedc||(273–288)||1UHG|
Peptides from several proteins were analyzed for solvent exposure using Protein Explorer 2.45, based upon known crystal structures submitted to the Protein Database, as referenced by the PDB numbers.
Potential competing peptides as defined in H-2d mice.
Peptides that were buried except for the N and C termini that were solvent exposed.
N terminus only determined because structure may alter depending on the inserting sequence.
Accessibility was determined in monomeric form.
Peptide that was buried except for the C terminus, which was solvent exposed.
Fig. 1 shows a comparison of CD4 T cell responses to peptides contained in their native protein context or inserted within MalE at aa 133. The data for each peptide are presented as the percentage of spots elicited by peptide relative to the total number of IL-2-producing cells that were elicited in response to the intact Ag. Strikingly, each of the peptides maintained their relative immunodominance observed in their native protein (Fig. 1, top panel) or when they were moved into the heterologous bacterial protein MalE (Fig. 1, bottom panel). The peptide Myo (102–118) and wild-type (WT) LACK (161–173), each binding very stably to class II (t1/2 > 150 h), remain immunodominant in MalE, recruiting ∼50% of the CD4 T cells that the intact protein Ag does. The OVA (273–288) peptide is subdominant, eliciting 15–25% of the total IL-2 response, both in its native OVA context and within MalE. In a similar fashion, the cryptic peptides OVA (328–339), HEL (11–25), and HEL (20–35), and the low stability LACK peptide variant consistently displayed an inability to prime CD4 T cell responses, independent of their molecular context, eliciting only 2–14% of the total IL-2 spots. These peptides have been shown to exhibit a t1/2 in association with I-Ad of less than 10 h at endosomal pH (33).
To determine whether there are differences in the conformational stability of the antigenic proteins, we assessed their sensitivity to acid-induced unfolding at the range expected to be encountered in endosomal compartments of APC. In Fig. 2, the fluorescent probe 8-anilino-1-naphthalenesulfonic acid (ANS), a fluorescent probe that binds to exposed hydrophobic patches (71, 72, 73, 74), was used to detect pH-dependent conformational changes. Each of the proteins studied, LACK, MalE, OVA, and HEL, displayed distinct sensitivities to low pH, as indicated by the variable ANS enhancement at each pH tested. For example, the LACK protein from Leishmania (Fig. 2,A, upper left) exhibited enhancement of ANS fluorescence at a modestly acidic condition of pH 5.5, MalE (Fig. 2,A, upper right) showed pH-dependent conformational changes only at pH 4.5, whereas, strikingly, HEL (Fig. 2,A, lower right) displayed almost no sensitivity to acidic pH, but did show evidence of unfolding in the presence of reducing agents (Fig. 2 B). These data suggest that during Ag processing within an APC, each Ag may be distinct in its progressive unfolding and differentially susceptible to proteolytic attack.
Collectively, these results, using unrelated foreign antigenic peptides, provide strong evidence that immunodominance is independent of the molecular context and that, conversely, the affinity for MHC class II is an accurate predictor of immunogenicity, irrespective of the protein context in which the peptide epitope is delivered.
The immunogenicity of a given peptide is not detectably altered by its location in a complex protein Ag
To further explore whether molecular context can influence the hierarchy of immunodominance, we asked whether there were differences in the immunodominance of antigenic peptides when they are located at different sites within a complex protein. We inserted a high stability variant of the well-defined cryptic peptide from influenza hemagglutinin (HA) (126–138) T128>V, with a substitution at the residue that anchors the peptide into the P1 pocket of I-Ad, into several distinct sites of MalE. Insertion of peptides into aa 133, 206, and 303 within MalE has been previously described (62, 75), who found that insertion of peptides at these sites does not affect the binding affinity for maltose, arguing that conformational integrity is largely intact, in agreement with crystallographic data (76). Table I shows the location of the peptides within the complex Ags, when known, from crystallographic studies. The structure of MalE shows that the segments around residues 133 and 206 are solvent exposed, whereas the region around aa 303 is buried.
One of the challenges in comparing individual responses with independent MalE constructs is that each molecular form is independently produced and purified in the laboratory. This could potentially introduce quantitative errors in estimation of concentration or purity of the constructs used for immunologic studies. To deal with this concern, before assessing the in vivo immunodominance of these test constructs, an in vitro Ag presentation assay was used to verify the functional concentrations of each Ag and to assess the release of the test HA peptide from the MalE constructs. When dose-response curves using T cell hybridomas specific for internal MalE peptides M1, reactive to MalE (269–285), and M3, reactive to MalE (69–84), respectively, were compared, similar dose-response curves between each set of constructs were observed (Fig. 3,A, top and bottom panels, respectively), indicating that the protein Ags were at similar functional concentrations. To assess whether release of the test HA peptide was similar in the different T128>V:MalE constructs, each construct was assessed in Ag dose-response assays using a T cell hybridoma specific for the HA peptide:I-Ad complex. T128>V:MalE (Fig. 3 B) 133 (squares), 206 (triangles), and 303 (circles) constructs displayed no differences in class II-restricted presentation of the HA peptide. This result suggests that the epitope density resulting from processing and presenting of the HA peptide encoded in the three different sites was similar. This in itself is informative because it suggests that the effective yield of peptide:class II at the cell surface of the APC was not significantly impacted by the localization of the HA peptide within MalE.
Having confidence that the three different constructs for the higher stability variant of HA were at the same effective concentration after their production and purification in vitro, it was possible to compare the immunogenicity of the HA peptides within the constructs in vivo. The CD4 response to internal MalE peptides was used to verify the quality of the immunization of different groups of mice, and thus to provide internal controls for the immunogenicity of the HA peptide. Fig. 4 shows that for each of the constructs tested, the MalE (69–84) peptide was the most immunogenic, followed by MalE (102–115), whereas the MalE (269–285) peptide was subdominant. Quantification of HA-specific CD4 T cells from the mice revealed that the responses elicited by the constructs bearing the HA peptide at the three different sites in MalE were very similar. These responses were approximately equal to the responses elicited by the MalE (269–285) peptide (Fig. 4).
Overall, these data involving MalE constructs with peptides at different sites show first, that the effective epitope density achieved on an APC in vitro of a given peptide is maintained when the peptide is located within different sites of a complex protein, and second, that the CD4 immunodominance patterns in vivo toward peptides are largely maintained when they are processed and released from different sites within the same protein context.
The immunogenicity of distinct registers of a single peptide
In the next series of experiments, we made side-by-side comparison of the immunodominance patterns of three different peptide epitopes in OVA. Each of these peptides is presented by I-Ad and was originally described by Grey and colleagues (77) as present in the proteolytic fragments of OVA recognized by T cell hybridomas, 3DO54.8, 3DO11.10, 8DO51.15, and 3DO18.3. The 3DO18.3 T cell recognizes a cyanogen bromide fragment (273–288), whereas 54.8, 51.15, and 11.10 all recognize peptides within a tryptic fragment (323–339). Both of these peptides are buried within the native OVA protein (Table I), although the segments flanking the 323–339 peptide are solvent exposed (78). More precise mapping of this segment of OVA (79) suggested that independent registers of the 323–339 peptide were stimulating the different T cell clones (Fig. 5 A). Peptide dissociation studies using I-Ad (70) indicate that binding of 323–339 to I-Ad displays complex kinetics, with the most N-terminal fragment (323–335) having the highest stability (t1/2 > 200 h), most likely corresponding to the register crystallized with I-Ad (80). Our own unpublished studies show that the N-terminal fragment is recognized by 3DO51.15 and 3DO54.8, whereas the most C-terminal fragment, 328–339, contains the register recognized by 3DO11.10 and the derived TCR transgenic mouse 11.10. Dissociation studies with purified I-Ad indicate that this peptide binds with very low stability (t1/2 < 5 h) (33).
The immunodominance of these different peptides in BALB/c mice using native OVA as the immunizing Ag is shown in Fig. 5,B. The (323–339) fragment, encompassing all of the different registers, stimulates the highest number of CD4 T cells, and virtually all of the responses seem attributable to the N-terminal (323–336) fragment, whereas the C-terminal (328–339) fragment recruits very few CD4 T cells above background (<5%). Also shown in Fig. 5 B is the immunodominance of the 273–288 peptide from OVA, which elicits significant, but somewhat lower numbers of CD4 T cells than does the 323–339 peptide. Therefore, in OVA, the 323–336 peptide is immunodominant, the 273–288 peptide is subdominant, and the 328–339 peptide is cryptic.
The OVA (323–339) peptide was then inserted into MalE, at aa 133, a site that is solvent exposed in MalE, or at aa 303, a site that is buried in MalE. Protein constructs bearing these peptides were produced and then tested in vivo for their immunodominance hierarchies. As can be seen in Fig. 5,C, the OVA (328–339) peptide remains cryptic in both sites of MalE, whereas, in contrast, the OVA (323–336) peptide maintains its dominant phenotype, recruiting similar numbers of cells as does the subdominant MalE (102–115) peptide. This finding indicates that distinct, but overlapping peptide epitopes can display quite discreet immunodominance patterns that correlate with their binding affinity to class II molecules rather than with their site within the native Ag. Finally, OVA (273–288) inserted at aa 133 and 303 in MalE maintained its subdominant phenotype observed with the native Ag (Fig. 5 D). Overall, the immunogenicity of the OVA peptides tracks with their primary sequence rather than with the protein context in which they are introduced into the immune system.
Addition of a dibasic motif increases the liberation of a cryptic peptide in vitro, but does not convert it to an immunodominant peptide in vivo
As a final strategy to addressing the impact of context in immunodominance, we sought to increase the efficiency of proteolytic release of a cryptic peptide and evaluate any changes in immunodominance. Dibasic motifs are recognized by a wide family of endoproteases (81, 82, 83, 84, 85, 86), and several reports have indicated that if these motifs flank antigenic peptides, class II:peptide presentation efficiency can be enhanced (27) or immunodominance can be conferred (28). For example, when native HEL protein was altered to contain a dibasic endosomal cleavage motif (82, 83, 85, 87) C terminus to the cryptic peptide HEL (20–35), class II-restricted presentation of the peptide by APC in vitro was enhanced (27), but this study did not examine whether this change was associated with any change in immunogenicity in vivo. Conversely, another report found that introduction of dibasic motifs flanking other epitopes increased immunogenicity in vivo, but did not examine Ag presentation efficiency. To address this issue more comprehensively, we asked whether the addition of an endopeptidase cleavage site (F34>R) flanking the cryptic peptide HEL (20–35) inserted in MalE leads to increased peptide liberation, and if so, using the same construct, whether this affects the immunogenicity in vivo. Examination of peptide-specific responses using the HEL (20–35)-specific T cell hybridoma indicated the amino acid substitution did not interfere with MHC class II or TCR binding (Fig. 6 A, inset), consistent with a modified residue being outside the core peptide-binding register (23–32) defined previously (27). To examine Ag presentation of this modified peptide, the WT and F34R variant sequences were encoded into the MalE shuttle protein at aa 133. The MalE proteins bearing the WT or modified peptide were produced, and dose-response curves were performed. As before, presentation of the endogenous MalE (269–285) and MalE (69–84) epitopes was used to evaluate the functional protein concentrations of the independently prepared Ags.
When dose-response curves using the intact MalE constructs bearing WT HEL (20–35) or the dibasic motif variant were performed, we observed a 4.5-fold increase in the presentation of the F34R peptide by A20 cells (Fig. 6,A, left) and a 6-fold increase in the presentation of the F34R peptide by freshly isolated splenocytes (Fig. 6,A, right). The observed differences were not due to protein concentration differences (Fig. 6, B and C), because there were no differences in any of the dose-response curves obtained to the internal MalE epitopes. We next asked whether the enhanced presentation was sufficient to alter the hierarchy of immunodominance in vivo. Fig. 6 D shows that the HEL (20–35) peptide is ∼2% of the total immune response, in agreement with previous studies demonstrating this peptide to be cryptic (68, 69, 88). When the responses to the dibasic motif variant F34R were similarly analyzed, it was found to maintain its crypticity in vivo, despite the increase in its Ag presentation by APC in vitro. The internal MalE control peptides from both constructs each elicited significant responses, and at similar levels among the immunized animals, confirming the mice were immunized equivalently. These data demonstrate that increased liberation of a cryptic peptide from a complex protein, at least at the levels possible through the introduction of a dibasic cleavage sequence, is not sufficient to confer immunodominance in vivo.
Although previous studies in our laboratory focused on evaluating the role the kinetic stability of the peptide:MHC complexes played in immunodominance (3, 33), the current studies were initiated to determine how significantly the molecular context of an antigenic peptide contributes to its immunogenicity in vivo. In this study, we use context to describe the elements within the intact foreign Ag that are independent of the primary sequence of the antigenic peptide. The question we sought to address is whether the potency of a peptide in recruiting CD4 T cells in vivo, its immunodominance, is seriously impacted by the protein context in which the peptide is contained or whether the immunodominance patterns are intrinsic to the primary sequence of the peptide. We have used a comprehensive, but empirical approach to assess this issue of the role of context in immunodominance, by changing the sites of expression of a series of unrelated peptides in the protein vector MalE and by comparing the immunogenicity of peptides that differ only in their affinity for class II within the same site of MalE. Comparing responses to peptides in their native context vs the same peptides inserted into MalE, we were able to change the competing peptides in the response and the secondary and tertiary structure of the Ag. By changing the localization of a peptide within a given protein or across different proteins, we anticipated that accessibility to proteases would vary, although we have not experimentally assessed this. Finally, in this study, we analyzed the potential of peptide liberation from the intact Ag to have a contribution to immunodominance through insertion of a dibasic motif adjacent to the cryptic peptide HEL (20–35). Collectively, these experiments revealed no differences in the immunogenicity of peptides when they are introduced in different molecular contexts, and find the immunodominance of a peptide tracks with the peptide itself, rather than the site in a given protein or the protein in which it is contained.
The results of our study raise the question of why a peptide’s context apparently plays such a minor role in immunodominance, as we have measured it experimentally. The first consideration is whether and by how much the yield of antigenic peptides varies among potential immunogenic peptides contained in the Ag and how these differences compare with the total quantity of competing peptides in endosomal class II peptide-loading compartments. Experimentally, microgram quantities of Ag are typically introduced s.c., and it can be expected that only a fraction of this will be taken up by APC. These Ag-derived peptides must compete for binding to class II molecules with peptides derived from a highly diverse and abundant pool of endogenously synthesized peptides. It is quite possible that under these conditions, even dramatic differences in yield of one peptide over another, perhaps as much as 10- to 20-fold within the Ag due to a favorable location and flanking sequences, will lead to only minor differences in the yield of cell surface-expressed peptide:MHC class II complexes because of competition with the tremendous excess of endogenously supplied peptides. These differences in ultimate yield may not offer a significant advantage of the more highly expressed complexes during CD4 T cell priming.
Secondly, DM editing may dramatically override the effects of differences in the initial yield of peptides from the intact Ag. Existing data on the function of DM (reviewed in Refs. 3 and 34) indicate that a subset of antigenic peptides is removed from the class II-binding pocket by DM, as CLIP is, before their export to the cell surface (35, 36, 89, 90, 91, 92, 93, 94). Biochemical studies with purified class II molecules (90, 91, 93, 94, 95) and Ag presentation studies (35, 94, 96) suggest that DM preferentially removes peptides that contain suboptimal side chains at their anchor positions, and that DM-antagonized peptides have lower stability interactions with the presenting class II molecule. Our laboratory has found that DM editing in APC can cause up to a 1000-fold change in initial epitope density, depending on the kinetic stability of the peptide:class II complex (96). This dramatic variability in the yield of peptide:class II complexes for different peptides due to differential DM editing is likely to profoundly diminish the impact of variation in initial proteolytic release of a peptide on its initial epitope density, and thus, its ability to successfully recruit CD4 T cells.
One can imagine that under some conditions of in vivo priming, molecular context may play a more dominant role. One of these conditions is when Ag is taken into APC by receptor-mediated uptake. Dendritic cells are known to take up Ag selectively by a number of cell surface receptors, as do B cells with their Ag-specific receptors. In vitro evidence suggests that the facilitated uptake of Ag can amount between 100- and 1000-fold enhancement in Ag presentation (97, 98, 99, 100, 101, 102). Also, some intracellular pathogens replicate rapidly within dendritic cells, and the abundance of their associated proteins may be quite high. Under these types of conditions, as also may occur when endogenously synthesized Ags are the target of CD4 T cells (2, 18), differences in yield among the peptides released from the target Ags may lead to a biologically relevant difference in the yield of peptide:MHC complexes at the cell surface that are determined by the initial yield of the peptide from proteolytic processing events. We have not yet established model systems to systematically examine immunodominance hierarchies under these types of conditions.
From the practical standpoint of vaccine design, the results of this study are highly encouraging and suggest that the immunodominance pattern of a given peptide will be maintained in different vaccine constructs, independent of its location within the Ag or the other potential competing peptides contained within the vaccine vector. This is particularly promising for efforts that seek to incorporate known pathogen or tumor-derived peptides into new complex proteins (44, 46, 49, 103, 104). There has been significant progress by many groups in epitope discovery using peptide vaccine vectors for pathogenic organisms (105, 106, 107, 108). Once identified, these peptides can be incorporated into different types of vaccine vectors, including those that selectively target dendritic cells (109, 110, 111). Additionally, the results of our studies completed to date indicate that the relative immunodominance of a peptide contained in foreign exogenous Ags can be modulated by alteration of its kinetic stability with its presenting class II molecule (3, 33). These results suggest that it is now possible to rationally promote a peptide’s immunogenicity by optimizing the anchor residues that bind to the MHC class II pockets. Such modified peptides can be incorporated into protein vaccines where the peptides will maintain their immunodominance. These vaccines can thus be implemented confidently to expand host CD4 T cells of the desired specificity.
We would like to thank Dr. Clara Kielkopf for the use of the FLUOROMAX Spectrofluorometer.
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 the National Institutes of Health Grants R01AI51542 and HHSN 266200700008C, which were awarded to A.J.S.
Abbreviations used in this paper: LACK, Leishmania homolog of receptors for activated C kinase; ANS, 8-anilino-1-naphthalenesulfonic acid; HA, hemagglutinin; HEL, hen egg lysozyme; WT, wild type.