The MHC-encoded cofactor DM catalyzes endosomal loading of peptides onto MHC class II molecules. Despite evidence from in vitro experiments that DM acts to selectively edit the repertoire of class II:peptide complexes, the consequence of DM expression in vivo, or a predictive pattern of DM activity in the specificity of CD4 T cell responses has remained unresolved. Therefore, to characterize DM function in vivo we used wild-type (WT) or DM-deficient (DM−/−) mice of the H-2d MHC haplotype and tested the hypothesis that DM promotes narrowing of the repertoire of class II:peptide complexes displayed by APC, leading to a correspondingly selective CD4 T cell response. Surprisingly, our results indicated that DM−/− mice do not exhibit a broadened CD4 T cell response relative to WT mice, but rather shift their immunodominance pattern to new peptides, a pattern associated with a change in class II isotype-restriction. Specifically, we found that CD4 T cell responses in WT mice were primarily restricted to the I-A class II molecule, whereas DM−/− mice recognize peptides in the context of I-E. The observed shift in isotype-restriction appeared to be due in part to a modification in the peripheral CD4 T cell repertoire available for peptide recognition.
Successful priming and expansion of Ag-specific CD4 T cells in response to exogenous Ag is a complex, multistep process. Specifically, within APCs, T cell priming requires productive proteolytic processing of the Ag in endosomal compartments, loading of the resulting peptides onto MHC class II molecules, and export of the peptide:class II complexes to the cell surface. A productive immune response also requires that the host possess a repertoire of circulating peripheral T cells capable of recognizing and engaging the peptide:class II complexes. Defects in the Ag presentation pathway or holes in the TCR repertoire can lead to the failure of a productive immune response to Ag. Interestingly, even when robust CD4 T cell responses are elicited in response to an antigenic challenge, the CD4 T cells typically display considerable selectivity in their peptide specificity, responding to a very limited number of peptides within the potentially vast peptide pool generated following Ag uptake: a phenomenon referred to as immunodominance (1, 2, 3, 4). Although poorly understood at the present time, the selectivity in T cell responses is thought to be due largely to intracellular events in Ag presentation.
Much research in the last decade on the control of class II-restricted Ag presentation has focused on the function and biochemistry of the MHC-encoded HLA-DM/H-2DM (DM)3 protein, a cofactor now known to be critical in endosomal peptide loading of class II molecules (5, 6, 7, 8). Accumulated biochemical studies (9, 10, 11, 12, 13, 14) have suggested that DM-activity favors the presentation of peptides that possess high stability interactions with class II therefore modifying the repertoire of peptides bound to class II molecules; a process that has become known as DM editing (1, 15). Functional studies (16, 17, 18, 19) from our laboratory and other laboratories lent support for the concept that DM expression edits the repertoire of peptides expressed on the surface of APC, thus functionally narrowing the specificity of the elicited T cells. Moreover, when we examined the effect of DM on the presentation of immunodominant or cryptic peptides (20), we found that immunodominant peptides were promoted in their presentation by the DM protein, whereas conversely, DM extinguished the presentation of cryptic peptides, or those peptides that do not elicit T cell responses when encoded within a complex Ag. We (21) and others (22) have also recently reported experiments from in vivo studies that show the kinetic stability of peptide:class II complexes is an important factor contributing to immunogenicity in CD4 T cell responses, a peptide-intrinsic parameter also found to have consequences on DM editing within APC when tested in vitro (23). However, little is known about the impact of DM in influencing immunodominance in vivo.
Initial studies on DM function in vivo were conducted in mice of the H-2b MHC haplotype. The absence of DM expression in these mice resulted in near normal class II expression. However, upon closer examination, it was discovered that nearly all surface-expressed class II molecules were occupied by a single peptide: CLIP (24, 25, 26). Furthermore, the CD4 T cell compartment, although reduced in number as compared with wild-type (WT) animals, was comprised of a significant population of CD4 T cells reactive to syngeneic class II molecules expressed by WT APC that bear the normal diversity of bound peptides. This finding was thought to reflect diminished expression of self-peptides during thymic selection due to the quantitative occupancy of class II molecules with the CLIP peptide (27, 28). The high occupancy of the I-Ab class II molecules by CLIP in DM-deficient cells made comparison of immunodominance patterns problematic. DM-deficient mice of the H-2d and H-2k MHC haplotypes not posing these limitations have since been created and characterized (29, 30). These animals express class II molecules capable of some spontaneous CLIP release (as a result of modest affinity for the CLIP peptide (31, 32), potentially allowing for endosomal loading of peptides onto class II molecules in the absence of DM. APC in these H-2d and H-2k mice also express both I-A and I-E class II molecules, allowing for additional analyses into isotypic requirements for DM-function. A recent report (33) using the H-2d and H-2k DM-deficient mice described the immunodominance response to the Leishmania Ag, Leishmania homologue of activated receptor for c-kinase (LACK). In the absence of DM expression, the repertoire of responding T cells was broadened to include new T cell epitopes. Thus, the conclusion from this body of work was that DM expression is required to focus the immunodominant response toward this particular epitope (33).
Collectively, these biochemical and functional studies have led to the view that DM promotes accumulation of a selected subset of the potential peptides derived from Ag, and it is these peptides that are available on the surface of APC to elicit T cell responses. One prediction of this model is that in the absence of DM, APC will express a broader array of peptides from the internalized Ag, perhaps including those peptides that have lower stability interactions with the class II molecule. These APC will, in turn, have the ability to elicit T cells of broader peptide specificity. In the present study, we sought to comprehensively test the hypothesis that the immunodominance in T cell responses to exogenous Ags in vivo would be expanded to include new peptide epitopes in the absence of DM expression. To test this hypothesis, DM−/− or WT mice of the H-2d haplotype were immunized with foreign protein Ags and tested for the specificity of the elicited T cells, using synthetic peptides and IL-2 ELISPOT assays to enumerate T cells of known peptide specificity. Contrary to expectations, we found that the specificity of the elicited T cell pool was not broadened to include additional epitopes, but rather shifted from I-A dominated responses to new dominant specificities associated with the alternate I-E isotypic form of the class II molecule expressed in H-2d mice. Further analysis revealed that the observed isotypic bias in DM−/− mice was, at least in part, due to an increased representation of I-E-restricted CD4 T cells within the periphery.
Materials and Methods
H-2d DM−/− mice were constructed as described (29) and bred onto the CB17 background, a strain congenic with BALB/c mice, but expressing the Igh b allotype from C57BL/Ka mice. CB17 WT animals were obtained from The Jackson Laboratory and maintained in pathogen-free facilities at the University of Rochester. C3H and CBA mice were obtained from National Cancer Institute-Frederick. All animal experiments were conducted according to the regulations set by the University Committee on Animal Care at the University of Rochester.
Cell culture reagents
Cell lines were maintained at 37°C and 5% CO2 in DMEM containing 5% FCS supplemented with 5 mM HEPES, 2 mM glutamine, 1 mM nonessential amino acids, and 5 × 10−5 M 2-ME. All medium and supplements were purchased from Invitrogen Life Technologies unless otherwise noted. The hybridomas producing mAbs reactive with I-Ad: MKD6 (34), I-E: 14.4.4S (35), or both class II molecules: M5/114 (36), were acquired from the American Type Culture Collection (ATCC). Other Abs used in this study were an anti-CD8 mAb 3.155 (37), anti-B220 mAb RA313A1/6.1, anti-Thy1.2 mAb J1j.10 (38), and an anti-rat κ mAb MAR 18.5 (39) obtained from ATCC. All Abs were used as culture supernatants unless otherwise noted. MHC class II staining of mouse splenocytes was performed using directly conjugated Abs (anti-I-E/PE, clone 14.4.4S, anti-I-Ad/FITC, clone AMS32.1, anti-I-Ad/biotin, clone 25.9.17, and anti-IgG2a/2b/FITC, clone R2–40 purchased from BD Pharmingin).
T cell hybridoma assays
MalE-specific T cell hybridomas were generated by fusion of lymphocytes from WT or DM−/− mice immunized subcutaneously with MalE (20 μg/hind-footpad) with the TCR negative BW5147 lymphoma cell line following a 3 day restimulation in vitro with 20 μg/ml MalE Ag and irradiated syngeneic APC (WT APC for CD4 T cells isolated from WT animals, and DM−/− APC for CD4 T cells isolated from DM−/− animals). After fusion, hybridoma cells were cloned by limiting dilution and Ag-reactive clones were further characterized for their peptide reactivity using individual MalE synthetic peptides and syngenic APC. The HA110–120 specific hybridoma (TS1) was created by fusion of peptide-activated LN cells from the 6.5-TCR TgN mouse (40) (TS1) with BW5147 lymphoma cells. For Ag presentation assays, 5 × 104 specific T cell hybridomas were mixed with 5 × 104 DAP cells (RT2.3B2 (I-Ad) or RT10.3C5 (I-Ed)) or 5 × 105 spleen cells as described (41) and peptide or protein at the indicated dose in flat-bottom 96-well dishes. After overnight culture, plates were frozen, then thawed and 50 μl supernatant was removed and added to 3.5 × 104 CTLL cells. After 16–20 h, CTLLs were incubated with 0.4 mg/ml MTT (Sigma-Aldrich) for 6 h, followed by 100 μl 10% SDS:0.01 N HCl. The OD was calculated from measurements at 570–650 nm.
Immunizations and IL-2 ELISpot assays
WT or DM−/− mice were immunized in hind-footpads with 50 μl of 400 μg/ml protein (MalE, TS1/MalE, β-galactosidase (Sigma-Aldrich), alcohol dehydrogenase (Sigma-Aldrich), LACK, or HEL (Sigma-Aldrich)) or 5 nM of peptide emulsified in CFA (Sigma-Aldrich). Ten days later, cells were isolated from draining popliteal lymph nodes. CD4 T cells were enriched following depletion of MHC class II and CD8 cells, using Ab supernatants specific for these cell surface markers (14.4.4S, M5/114 followed by MAR18.5, B220 and 3.155). Ab-coated cells were then depleted using rabbit complement (Cedarlane Laboratories) according to the manufacturer’s guidelines, typically resulting in 94–96% CD4 T cells, ascertained by FACs analysis. For certain experiments using class II transfectants as a source of APC, further purification using mixtures of sheep anti-mouse IgG, goat anti-rat IgG, and anti-B220 Dynal beads (Dynal Biotech/Invitrogen), or CD4 purification kits (MACs) were performed according to the manufacturer’s guidelines, to eliminate any contaminating APC from the T cell preparation. IL-2 or IFN-γ production was measured by ELISPOT assay as described previously (42), using DMEM medium with 10% FCS (16) instead of RPMI. For MalE peptide pool experiments, 32 pools consisting of four peptides each at a final concentration of 12.5 μM were used (synthetic peptides obtained from Mimetopes, Australia), with pools 1 and 2 representing negative controls spanning the MalE signal sequence, cleaved following translocation to the periplasm during protein production (43). For experiments involving MHC class II blocking, purified MKD6 (specific for I-Adβ) and/or 14.4.4S (I-Eα specific) were added at 80 μg/ml. Ag-loaded APC (WT APC for CD4 T cells isolated from WT animals, and DM−/− APC for CD4 T cells isolated from DM−/− animals) were incubated with the mAbs at 37°C for 30 min before addition of the CD4 T cells. Quantification of spots was performed on an ImmunoSpot reader series 2A using Immunospot software version 2.0 (Cellular Technologies). Mean number of spots for each condition was determined from triplicate wells.
MalE and LACK protein purification
For TS1/MalE constructs, PAGE-purified synthetic oligonucleotides encoding the desired peptide (LSSVSSFERFEIFPKESS) were obtained from IDT DNA Technologies and resuspended in 10 mM Tris/1 mM EDTA pH 8.0 at a concentration of 100 μM. Annealed dsDNA was phosphorylated in vitro and ligated into BamHI-digested MalE133 vector. For experiments involving MalE alone, the MalE vector with no additional insert at the BamHI site was used. Sequenced clones were transformed into MalE (−/−) ER2507 E. coli. A total of 40 ml of overnight bacterial cultures expressing the MalE construct were added to 4 L of LB:ampicillian with 0.2% maltose and grown for 6–7.5 h at 37°C. Bacteria were pelleted by centrifugation at 7000×g for 10 min at 4°C and resuspended in 800 ml of pH 8.0 30 mM Tris:Cl/20% sucrose/1 mM EDTA. After shaking for 10 min, bacteria were pelleted by centrifugation at 8000 × g for 10 min at 4°C and resuspended in 400 ml of ice cold 5 mM MgSO4 and shaken for 10 min on ice. After centrifugation at 8000 × g for 10 min at 4°C, the supernatant was collected and 8 ml of 1 M Tris:HCl (pH 7.4) was added. The osmotic shock fluid was filtered over a 0.45-μM membrane and applied to an amylose column (44) of 15-ml bed volume with a flow rate of ∼1 ml/min. The column was washed with 250 ml column buffer (20 mM Tris:HCl (pH 7.4); 0.2 M NaCl, and 1 mM EDTA) and protein was eluted in 50 ml of column buffer/10 mM maltose. Collected fractions positive for protein (Bradford analysis; Bio-Rad) were pooled, dialyzed against PBS, and concentrated with a Centricon μm-10 kDa cutoff filter to ∼1-ml final volume. LACK was synthesized and purified as described (21). Concentrated protein in PBS was sterile filtered through a 0.2 μM syringe filter, quantified by Bradford assay and SDS-PAGE, and stored at −20°C.
Alloreactive T cell assays
Purified peripheral lymph node CD4 T lymphocytes from DM−/− or WT mice at 1 × 105, 5 × 104, 2.5 × 104, or 1.25 × 104 cells/well were stimulated with T cell depleted C3H splenocytes (T cell depletion accomplished using anti-Thy1.2 mAb J1j.10) at 50 × 104, 12.5 × 104, or 3.1 × 104 cells/well in the presence of blocking Abs against I-Ak (10.2.16, ATCC), I-Ek (14.4.4S), both Abs, or control Ab (MK-S4). After overnight incubation, the number of IL-2 producing cells were quantified from four replicate wells by ELISPOT using ImmunoSpot counter software.
Identification of the immunodominance pattern to MalE Ag in WT or DM−/− mice
To determine the influence of DM expression on the selective specificity of CD4 T cells in response to Ag, we initially studied the immune response to a completely heterologous Ag, the maltose binding protein of Escherichia coli (MalE). Use of a foreign Ag was anticipated to broaden the peptides that would elicit a response by eliminating issues of self-tolerance. DM-deficient (DM−/−) or wild-type (WT) H-2d mice were immunized subcutaneously with MalE protein emulsified in CFA. Ten days after immunization, CD4 T cells isolated from draining lymph nodes were analyzed for their peptide specificity using a set of overlapping 15mer peptides representing the entire MalE sequence. Isolation of CD4 T cells at this time point from the lymph node environment is predicted to include factors within priming events relevant to immunodominance hierarchies, such as Ag processing, epitope density, T cell precursor frequency and T cell expansion that occurs over the 10 days of the response. MalE peptides were grouped into 32 pools consisting of four peptides each, as represented in Fig. 1 A. The observed reactivity to the MalE peptide pools is expressed as the percentage of the IL-2 spots obtained with the individual peptide pools relative to the total IL-2 spots cumulatively produced by all of the pooled peptides. To accurately assess the immunodominance pattern to the MalE Ag, we established arbitrary exclusion criteria, with responses consistently representing 4% or less of the total response being eliminated from further analysis. Using this criterion, we found that CD4 T cells from WT mice responded to a total of 9 peptide pools (pools 8, 11, 12, 14, 15, 16, 21, 26, and 30), including 8, 11, and 26, pools that contain peptides corresponding to previously characterized MalE epitopes (45, 46). Interestingly, when responses in DM−/− mice were analyzed, a similar number of peptide pools elicited a productive IL-2 response, with a total of 8 (pools 4, 7, 8, 10, 11, 12, 14, and 16) displaying reactivity above 4% of the total response. Surprisingly, this analysis revealed that several of the pools recognized in DM−/− mice were distinct from those eliciting responses in WT mice, associated with a gain in T cell activation to new peptide pools, including 4, 7, and 10, and the reciprocal reduction or loss of reactivity to pools 8, 11 and 26.
This pattern of epitope shift between WT and DM−/− T cell responses was further reinforced when individual peptides from pools 8 through 11, as well as 25 and 26 were used to restimulate CD4 T cells in IL-2 ELISPOT assays. As shown in Fig. 1 B, peptide68–82 from pool 8, as well as peptide98–112 and peptide101–115 from pool 11 restimulated T cell responses in WT mice, but elicited little to no reactivity with T cells from DM−/− mice. Conversely, peptide89–103 and peptide92–106 from pool 10 stimulated large numbers of IL-2 producing T cells from DM−/− mice. However, WT cells showed no reactivity to these two peptides. Peptides275–289 and peptides278–292 from pool 26 were capable of eliciting nearly equivalent responses from WT and DM−/− mice in this assay. Overall, the results from the peptide pool experiments in conjunction with the experiments involving individual peptides indicated that, contrary to expectation, the DM−/− response to the MalE Ag was not broadened, eliciting responses to more peptide epitopes relative to the number responded to in WT mice. Rather, the response in DM−/− mice was shifted, resulting in a gain in reactivity to certain peptides and a reciprocal loss in reactivity to others.
The shift in epitope recognition between WT and DM−/− mice is associated with a change in MHC class II isotype restriction
Because H-2d mice express both I-A and I-E MHC class II molecules, we compared the class II restriction pattern within the responding T cell populations in WT and DM−/− mice. To evaluate this, a panel of MalE-specific T cell hybridomas were derived by fusing MalE-primed T cells (isolated from either WT and DM−/− mice) with the TCR-negative cell line BW5147 after a brief in vitro restimulation of the CD4 T cells from the primed mice with MalE Ag and syngeneic APC. Hybridomas recognizing the immunodominant epitopes69–81, 103–118, and 269–285 from WT mice, as well as the epitope89–104 from DM−/− animals were identified, subcloned, and then analyzed for MHC class II restriction in T cell assays using transfected APC expressing either the I-Ad or I-Ed MHC class II molecules. This assay revealed an interesting and unexpected phenomenon: there was a striking disparity in the MHC class II restriction between CD4 T cells derived from WT and DM−/− mice. T cell hybridomas from WT mice recognized peptides in the context of I-A (Fig. 2,A, 69–81, 103–118, and 269–285), where as the T cell hybridoma obtained from DM−/− mice showed restriction to I-E peptide complexes (Fig. 2 A, 89–104).
The dramatic difference in the isotype-restriction of CD4 T cells from WT and DM−/− mice prompted us to develop an alternative strategy to further verify this important finding. Class II restriction differences were evaluated through the use of isotype-specific class II blocking Abs. In these assays, CD4 T cells from MalE-primed mice were restimulated with syngenic splenocytes and MalE Ag in the presence of I-Aβd (MKD6) or I-Eα (14.4.4S) specific Abs, a combination of both, or in the absence of any blocking Ab. Using this strategy, we found that 100% of the responding T cells from WT mice recognized epitopes from the MalE Ag in the context of I-A (Fig. 2,B). However, in DM−/− mice this pattern was reversed, with now only 7% of the responding T cells recognizing I-A-restricted peptides, and 92% reacting to peptides presented within the context of I-E. Importantly, this shift in Ag restriction cannot be explained by altered patterns in class II isotype expression. Both DM−/− and WT animals express similar levels of I-A and I-E as assessed by mAb staining (Fig. 2,C). In fact, the very small difference in expression detected by staining with the AMS32.1 Ab indicated slightly enhanced expression of I-Ad in the DM−/− mice relative to I-E. More comprehensive staining with the I-Ad-specific Abs MKD6, 25.9.17, and K24.119 (Fig. 2,D), as well as the I-E specific 14.4.4S (Fig. 2 C) demonstrate similar ratios of I-A to I-E on the surface of APC isolated from WT or DM−/− mice on the BALB/c background, and are in complete agreement with previously detailed studies (29). Therefore, the shift in peptide recognition by CD4 T cells in DM−/− mice associated with an altered isotype-restriction cannot be explained by changes in class II expression.
Isotypic bias in WT and DM−/− mice extends to multiple Ags
We next sought to determine whether the observed class II restriction bias was unique to the MalE Ag or whether it was a generalized trend applicable to multiple foreign Ags. Because it is prohibitively expensive to adopt the peptide scanning technique for multiple protein Ags, we used the same class II blocking strategy described above for MalE responses to assay responses to multiple exogenous Ags. WT and DM−/− mice were immunized separately with different foreign Ags (including alcohol dehydrogenase, β-galactosidase, and LACK), and CD4 T cells from the immunized mice were restimulated in vitro with intact Ag in the presence or absence of isotype-specific class II blocking Abs. As before, IL-2 ELISPOTs were used to enumerate the number of Ag-reactive cells. In WT mice, the majority of responding CD4 T cells (ranging from 63 to 97%) were restricted by the I-A class II molecule (Fig. 3). This I-A restriction bias within the responding CD4 T cell population was not a trait unique to CB17 mice, for we found that additional mouse strains with alternative background genes (BALB/c and B10.D2) displayed a similar pattern when challenged with these particular Ags (data not shown). In contrast, in DM−/− H-2d mice, the majority of responding T cells exhibited an I-E restriction bias, ranging from 65 to 100%, depending on the Ag. Interestingly, this altered isotypic restriction bias extended to animals immunized with the recently characterized LACK Ag (33). In agreement with a previous finding, nearly 100% of responding T cells in LACK-immunized WT animals were restricted to the I-A class II molecule, most likely due to CD4 T cells specific for the peptide156–173 that has been shown to be immunodominant in H-2d mice (33, 47, 48). However, in DM−/− animals, the class II restriction pattern was almost completely reversed, with 84% of responding T cells now recognizing peptides in the context of I-E. It was shown previously that two additional peptides (LACK peptide33–52 and peptide257–276) within the LACK protein, in combination with the response observed to the LACK156–173 peptide, are capable of eliciting responses in DM-deficient mice, with peptide [33–52] eliciting the majority of CD4 T cell responses (33). Through epitope/class II restriction studies of these LACK peptides we found that the immunodominant epitope [33–52] in DM−/− animals was indeed restricted to the I-Ed molecule (Fig. 3 B). Collectively, these results indicate that in the absence of DM expression, the predominating CD4 T cell responses following peripheral antigenic challenge are restricted to the I-E MHC class II isotype.
The I-E dominated responses in DM−/− mice persist when intracellular processing events are bypassed
One possible explanation for the shift in isotype specificity is that the presentation of peptides by I-E MHC class II molecules is less dependent on DM function in APC than peptide presentation by I-A, as has been suggested previously, determined in vitro through the activation of epitope-specific T cell hybridomas (29, 30). Therefore, we evaluated peripheral CD4 T cell responses in DM−/− and WT mice using an approach that would diminish any potential DM-mediated preferences in class II loading of peptides onto the I-A or I-E molecules. We used a strategy to bypass the need for intracellular processing of Ag, and therefore, diminish the impact of endosomal DM editing for cell surface expression of peptide:class II complexes. WT or DM−/− mice were immunized with either single peptides (Fig. 4,A), or a mixture of peptides (Fig. 4 B) that have known restrictions to either I-Ad or I-Ed class II molecules.
Ten days after peptide immunization, CD4 T cells isolated from immunized mice were tested for peptide specificity using IL-2 ELISPOTs. In single peptide immunization experiments, CD4 T cells were restimulated with APC isolated from both WT and DM−/− mice using various peptide doses. Mouse-to-mouse variability was also evaluated to compare peptide presentation by the two strains of mice (Fig. 4,A). Both WT and DM−/− animals elicited similar responses to the two I-Ad peptides MalE269–285 and MalE69–81 when administered alone, suggesting that both strains of mice possess a repertoire of T cells that can recognize these peptides. Immunizations with either of the I-Ed-restricted peptides, HEL103–117 or MalE89–104, resulted in enhanced CD4 T cells responses in DM−/− mice, with a 2-fold increase in total IL-2 spots produced as compared with CD4 T cells purified from WT mice. These results indicate that CD4 T cell repertoires to these individual peptides are present in both WT and DM−/− mice, with a possible an enrichment for T cells recognizing I-E restricted epitopes in DM-deficient mice. Most importantly, when we assessed the peptide presentation capacity of the DM+ vs DM-deficient APC, we noted no detectable difference in the ability of WT or DM−/− APC to present either I-Ad or I-Ed restricted peptides (Fig. 4 A). These results indicate that during polyclonal in vivo responses, peptide presentation by either of the MHC class II isotypes results in similar CD4 stimulation regardless of the presence or absence of the DM protein within APC.
The similar responses obtained using individual peptides prompted us to determine the preference in I-A and I-E-restricted responses when mixtures of peptides are used during immunization. Using this strategy, and in light of the single peptide experiments, one would predict that the loading of peptides onto class II molecules would be similar in WT and DM−/− mice, allowing for recruitment of both I-A and I-E-restricted T cells. However, unlike the single peptide immunization strategy, the use of mixtures of peptides would place the elicited CD4 T cells in competition with one another. Thus, mixtures of peptides should allow for immunodominance hierarchies to become established while bypassing the need for endosomal proteolysis and loading. The results of this peptide mixture experiment, as shown in Fig. 4b, revealed that the I-A-restricted response bias persisted in WT mice, with the CD4 T cells primarily reacting to the I-A presented MalE69–81 and HA126–138 peptides, as well as the two I-Ad-restricted OVA epitopes tested. The opposite pattern was observed with T cells isolated from DM−/− mice, which displayed an enhanced I-E restriction bias, responding predominantly to the I-E presented MalE89–104 and HEL103–117 peptides and showing very little reactivity to the I-A restricted epitopes included in the peptide mixture. This result, in conjunction with that of the individual peptide responses, suggests that the T cell repertoire in DM−/− mice may be biased toward I-E restricted epitopes, whereas the peripheral CD4 T cell compartment of WT animals predominantly favors I-A restricted peptides. Furthermore, these biases are most apparent when I-A and I-E restricted responses are competing within the lymph node environment containing the responding T cell populations.
DM-dependent I-E-restricted peptides recruit more T cells in DM−/− mice as compared with WT mice
We sought to further test whether the I-E restriction bias in DM−/− mice was a reflection of greater endosomal loading of I-E restricted peptides as compared with I-A in the absence of DM expression, as compared with an increased proportion of I-E restricted CD4 T cells in the periphery. There are reports indicating that I-E molecules are more readily available for peptide acquisition in the absence of DM expression (29, 30), suggesting that Ag priming might lead to an increased epitope density on I-E molecules as compared with I-A in DM−/− mice. Therefore, to address this possible complication in interpretation of the peptide priming experiments, we examined T cell responses in WT or DM−/− mice toward Ags bearing I-E restricted peptides whose presentation are known to be promoted by DM. As shown in Fig. 5, two classically defined I-E specific immunodominant epitopes, HEL103–117 and HA110–120 (49, 50), as well as the newly identified MalE89–104 epitope displayed dramatic shifts in their Ag dose-response curves with DM-positive APC in comparison to DM-negative APC. The I-E-specific presentation assays were tracked in parallel with I-A restricted T cell responses to assess the relative DM reliance between the two isotypic class II molecules. The results of these assays, shown in Fig. 5, indicate that I-A and I-E restricted responses were quite similar with regard to the impact of DM on presentation of Ag, indicating that both the I-A and I-E epitopes tested require the presence of DM for enhanced peptide:class II presentation. Accordingly, priming mice with Ags bearing epitopes whose presentation is enhanced by DM should result in a higher epitope density on WT APC as compared with DM−/− APC, which are available for CD4 T cell recruitment. We reasoned that if there were no restriction bias in the responding CD4 T cells within DM−/− mice, the higher epitope density of I-E restricted peptides on WT mice, due to the enhanced generation of peptide:class II complexes in the presence of the DM molecule, would promote a correspondingly larger response to these peptides in the WT animals as compared with DM−/− mice. To test this, we again used the MalE Ag, taking advantage of its structural flexibility, which permits the insertion of peptides up to 70 aa in length into selected sites (51). This allowed the immunodominant Influenza hemagglutinin peptide110–120 (“TS1”) to be inserted and expressed as part of a soluble Ag. WT or DM−/− mice were immunized with the TS1-MalE protein. Ten days later, purified CD4 T cells were restimulated with peptides either representing the inserted HA epitope or the MalE backbone peptides. IL-2 ELISPOT assays were used to enumerate peptide reactive T cells, and results are again expressed as the percentage of the total response. Fig. 6 A shows that CD4 T cells isolated from WT mice responded vigorously to the I-Ad restricted MalE epitopes69–81, 103–118, and 269–285, which collectively represent >85% of the total response. The I-E restricted HA110–120 epitope, which should have an increased epitope density on APC in WT mice, only elicited IL-2 production from ∼12% of the responding CD4 T cells in WT mice. As before, a response to the I-E-restricted MalE (89–104) epitope was undetectable in these mice. However, in CD4 T cell responses from DM−/− mice, the HA110–120 peptide recruited a much higher proportion of T cells, accounting for ∼35% of the response. The remaining responses were primarily restricted to the other I-E presented peptide (MalE89–104, 42%), while the I-Ad restricted MalE69–81, MalE103–118, and MalE269–285 peptides each accounted for <15% of the response. This result argues that the epitope density of I-E restricted peptides is not the primary factor dictating the isotype-biased response, but rather that the DM−/− mice may have a greater proportion of the CD4 T cell repertoire restricted to the I-E class II molecule that are available to respond to these peptides.
One potential caveat to the above interpretation is that CD4 T cells of differing class II restriction may compete during primary responses, particularly if there are both I-A and I-E immunodominant epitopes present within the immunizing Ag. In the experiments described above the insertion of HA110–120 into the MalE protein preserves the identified I-Ad restricted eptiopes (MalE peptides69–81, 103–118, and 269–285), potentially dampening responses to the I-E restricted HA110–120 epitope. To address this concern, two experimental strategies were used. First, similar experiments were conducted following immunization with the hen egg lysozyme (HEL) protein. The HEL Ag contains a limited number of epitopes for mice of the H-2d haplotype (49), including the dominant HEL103–117 I-Ed-restricted epitope as well as the cryptic I-Ad epitopes HEL11–25 and HEL20–35 (52, 53, 54). The minor responses to the I-Ad restricted peptides would not be expected to compete with the dominant I-Ed restricted response to the HEL103–117 epitope. Strikingly, these experiments revealed that mice lacking DM expression generated greatly enhanced responses to the HEL103–117 I-E restricted epitope compared with WT mice (Fig. 6 B), despite eliciting similarly low responses to the I-Ad restricted HEL (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25) peptide. These results suggest that the preimmune T cell repertoire in DM−/− mice is enriched for I-E restricted CD4 T cells.
A second strategy was undertaken to address the potential influence of competition between I-A and I-E restricted CD4 T cells on the observed I-A-restriction bias. Rather than using an Ag with few known epitopes, as in the case of the HEL protein, we engineered an Ag in which the major I-Ad-restricted immunodominant epitope was eliminated. Once more, the MalE Ag was exploited, introducing two substitutions within the immunodominant epitope69–81 at residues corresponding to the putative MHC class II pocket interactions, known to diminish class II binding (Fig. 7,A and Ref. 21). The Q72T, L75A mutations, at the P1 and P4 MHC anchors, resulted in a profound decrease (>50-fold) in peptide/class II stability, as determined by dissociation of peptide from soluble I-Ad over time (Fig. 7 A and data not shown). It was also shown using MalE-specific hybridomas that the MalE Q72T, L75A Ag was greatly diminished in the ability to stimulate the MalE69–81 specific hybridoma, despite retaining equivalent stimulatory capacity of the remaining MalE-epitope specific hybriomas (data not shown). This provides evidence that the introduced mutations solely affected the presentation of the epitope69–81. We could therefore ask whether the loss of the dominant I-A restricted response to epitope69–81 allows for I-E resticted responses to emerge in WT mice. To test this, WT MalE and MalE-Q72T, L75A proteins were used to immunize WT mice to determine the hierarchy of the characterized I-A and I-E peptide responses in the presence or absence of the potentially competitive immunodominant peptide69–81. The wild-type MalE protein again elicited a large proportion of the response against the three I-Ad-restricted epitope269–285, epitope103–118, epitope69–81, with responses to the epitope69–81 predominating. However, when WT mice were immunized with the variant Q72T, L75A protein, the epitope69–81 completely lost its immunodominance, while the epitope269–285 and epitope103–118 remained intact. Most importantly, the loss of the dominant I-A restricted peptide did not result in a gain in the I-Ed-restricted T cell response to the peptide89–104. The results of the experiments with HEL and the modified MalE led us to conclude that the failure to recruit I-E-restricted responses in WT mice cannot be accounted for by a robust I-A-restricted response.
CD4 T cell responses in WT and DM−/− mice display similar effector cytokine production
The numerous effects of the cytokine, IFN-γ, in an immune response are well established, including roles in CD8 T cell effector mechanisms and B cell Ab isotype-switching. This predominantly T cell-secreted soluble mediator is also known to facilitate macrophage activation, playing an important role in the destruction of both intracellular bacteria as well as extracellular pathogens. IFN-γ also enhances the surface expression of both MHC molecules and the costimulatory molecules CD80 and CD86, allowing APC to efficiently present antigenic-peptides to surveying T lymphocytes (reviewed in Ref. 55). Collectively, the profound impact IFN-γ has during an immune response makes it a potent and in some cases essential effector cytokine. To assess the effector function of I-E restricted T cells in DM−/− mice, we sought to determine whether IFN-γ production by the I-E-restricted immunodominant responses in DM-deficient mice were comparable to those of the I-A predominated responses in WT mice. To address this question, WT and DM-deficient animals were immunized subcutaneously with intact Ags (MalE, LACK, or HEL) and ten days later CD4 T cells isolated from the draining lymph nodes were evaluated for cytokine profiles following peptide restimulation in ELISPOT assays. In these assays, both IFN-γ and IL-2 producing T cells were evaluated in parallel to determine the relative ratios of these two cytokines in both WT and DM−/− mice, providing a standard by which the quantity of IFN-γ produced could be empirically ascertained. This type of analysis revealed that first and most importantly, the ratio of T cells producing IFN-γ to IL-2 for immunodominant peptides69–81 in WT mice vs peptides89–104 in the DM−/− mice is similar between the two strains (Fig. 8). Specifically, the immunodominant MalE69–81 epitope in WT animals elicited approximately twice as many IFN-γ producing T cells as IL-2 producing CD4 T cells. A similar ratio of T cells producing these two cytokines is obtained in DM-deficient animals when the now I-Ed-restricted immunodominant peptide89–104, is analyzed, indicating that the immunodominant peptides presented in the context of I-E are capable of eliciting effector functions. Secondly, the ratio of IFN-γ to IL-2 producing T cells increases proportionately to the immunodominance of the recall peptide and importantly this relationship between effector cytokine production and immunodominance is maintained between WT and DM−/− mice.
Allogeneic responses in DM−/− mice support a bias in the isotypic restriction of peripheral CD4 T cells to I-E
The results from the TS1-MalE and HEL priming experiments, as well as from the peptide mixture experiment, suggested that the preferential I-Ed restriction observed in the DM−/− mice may be due at least in part to the CD4 T cell repertoire available in these animals. To further evaluate the restriction bias in the peripheral CD4 T cell pool, we took advantage of the high preimmune repertoire that exists within the CD4 T cell compartment that is capable of responding to allogeneic MHC class II molecules. We speculated that anti-MHC alloreactivity to I-A vs I-E might reflect the isotype preferences of the T cell receptors expressed by the overall CD4 T cell population. If so, reactivity to I-A or I-E alloantigens could be used to assess the relative fraction of CD4 T cells that are restricted to the specific MHC class II isotypes. CD4 T cells from unimmunized WT or DM−/− mice were isolated and cultured with T cell-depleted splenocytes from C3H mice (H-2k background) in the presence or absence of isotype-specific blocking Abs. It is also important to note that the H-2k allo-Ag-bearing splenocytes in these assays express the DM molecule, eliminating any DM-dependent effects in the display of self-peptides. The number of alloreactive CD4 T cells was estimated using IL-2 ELISPOTS (Fig. 9). When CD4 T cells from WT mice were cultured with allogenic APCs, over half of the responding CD4 T cells could be mapped to the I-A isotype. When CD4 T cells from DM−/− mice were similarly analyzed, the preferences in allorecognition were reversed and a majority (65%) of the responding CD4 T cells were restricted to the I-E isotype. Nearly identical results were obtained when a purely naive T cell population was used (data not shown). This result suggests that in the absence of DM expression, the peripheral CD4 T cell pool is altered, with a greater proportion displaying restriction to the MHC class II I-E isotype.
The I-A restriction bias observed in H-2d mice extends to animals of the H-2k haplotype
The striking I-A restriction bias observed in WT mice of the H-2d haplotype encouraged the examination of epitope specificity within CD4 T cell responses in mice expressing different MHC class II alleles. Mice of the H-2k haplotype are among the few available inbred strains expressing both I-A and I-E isotypes, permitting this type of isotype-specific analysis. The previously described set of overlapping 15mer peptides representing the entire MalE Ag was once again used, and peptides were arranged into a peptide matrix, a strategy designed to quickly identify single CD4 T cell-reactive peptides (56, 57, 58). WT mice were immunized with the MalE Ag and ten days later peptide-specific CD4 T cells responses from pooled draining lymph nodes were enumerated in IL-2 ELISPOTs. Positive peptide pools possess stimulatory peptides, while negative pools were considered to be devoid of any CD4 T cell epitopes and were eliminated from any further analyses. A total of 42 individual candidate peptides were identified using this strategy (data not shown) and were further evaluated individually. As shown in Fig. 10, iterative or sequential analyses of T cell responses to the individual peptides revealed a total of 8 peptides capable of consistently generating CD4 T cell responses (10, 11, 68, 72, 77, 78, 88, and 89). Of these, five peptides were found to be restricted to I-Ak (peptides 68, 77, 78, 88, and 89) with peptides 77, 78, 88, and 89 predominating, whereas peptide 10 was presented within the context of I-E. It is likely that peptides 77 and 78, and peptides 88 and 89 represent single CD4 T cell epitopes, due to the amount of overlap between the individual peptides (Table I). Overall, the results obtained from the MalE peptide pool arrays indicate that the immunodominant CD4 T cell epitopes (MalE203–220 and 236–253) in H-2k mice are restricted to the I-A isotype.
|MalE Peptide .||Amino Acid Sequence .||MHC Class II Restriction .|
|MalE Peptide .||Amino Acid Sequence .||MHC Class II Restriction .|
To further validate the finding that the immunodominant MalE epitopes in H-2k mice are restricted to the I-A isotype, we once again performed class II blocking studies using the I-Ak specific Ab 10.2.16, the I-E Ab 14.4.4S, or a combination of the two. This alternative assay also allowed us to conclude that the majority of epitopes generated from the MalE Ag and recognized by CD4 T cells are presented by the I-A isotype (Fig. 11). The pattern of isotype bias was also evaluated using a panel of different foreign Ags, including LACK, β-galactosidase, alcohol dehydrogenase, and PPD (Fig. 11). In these studies, I-A restricted T cell responses represented on average ∼100% of the total Ag-specific responses, whereas I-E-restricted CD4 T cell response within this panel of exogenous Ags represented the minority of responses. These studies indicate that wild type CD4 T cell responses in H-2k mice are also predominantly restricted to the I-A isotype.
In the present study, we sought to determine the influence of DM on the immunodominance pattern of peripheral CD4 T cell responses following antigenic challenge with foreign Ags, specifically testing the hypothesis that in the absence of DM expression, CD4 T cell responses would broaden to include reactivity to newly presented peptides, perhaps now including peptides with lower affinity for class II molecules. Surprisingly, results obtained during our initial peptide scanning studies with the MalE Ag revealed that the DM−/− mice did not possess broadened CD4 T cell activity, but rather shifted the response to new specificities, gaining reactivity to new peptides and losing reactivity to others. Even more striking was the observation that the shift in Ag presentation observed in DM−/− mice was associated with an altered restriction toward the I-E MHC class II molecule.
There were several, nonmutually exclusive explanations for the shift in isotype preference in CD4 T cells in WT verses DM−/− mice that we considered and experimentally addressed in the present study. The first and most straightforward explanation is that I-E molecules depend less on DM for endosomal peptide loading than I-A molecules. Indeed, data in the literature has lent support for this possibility (29, 30), demonstrating enhanced presentation of Ag-derived I-Ed and I-Ek epitopes over I-Ad and I-Ak, respectively, in the absence of DM expression. As reported here, and in agreement with other published findings (59), this phenomenon may not be shared by all I-E-associated peptides, for the presentation of MalE89–104, HA110–120, and HEL103–117 by I-E molecules to T cell hybridomas in vitro were greatly dependent on the expression of DM within APC. Additionally, when Ag processing was bypassed, eliminating the need for intracellular processing and class II loading, the responses elicited toward peptides within the immunizing mixture that are presented by I-A molecules persisted in WT animals, whereas in DM−/− mice, CD4 T cell responses against peptides restricted to I-E again predominated. This suggests that there are additional parameters contributing to the observed MHC restriction bias in the CD4 responses elicited following antigenic challenge, most likely an isotypic bias in the circulating, preimmune CD4 TCR repertoire. Indeed, this possibility was supported by examination of preferences for I-A vs I-E class II molecules in alloreactive responses in the H-2d WT or DM−/− mice and by examination of responses to the I-E restricted HEL103–117 and HA110–120 peptides, shown to display enhanced epitope density by DM-positive APC.
If I-E dominated responses in DM−/− mice are due at least in part to a bias in the TCR repertoire, what might account for this? DM activity in the thymus, specifically within medullary thymic epithelial cells, has been proposed to act in a manner similar to that during peripheral Ag presentation (60), potentially removing CLIP and other peptides displaying intrinsic low stability interactions with MHC and enhancing the presentation of stably binding self-peptides. One possibility for a bias toward I-E-driven selection in DM−/− mice is availability of peptide-receptive I-E molecules in comparison to I-A, perhaps related to differences in CLIP occupancy. The association of CLIP within the I-Ad molecule has been determined through dissociation assays performed in our laboratory (21, 61) and others (32), and has been found to display an intermediate stability, with a dissociation half-life of ∼10 h at endosomal pH in the absence of DM. We have found in preliminary experiments that the I-E molecule has a less favorable association with the CLIP peptide, allowing for rapid dissociation of this peptide, estimated to be less than one hour. This is in agreement with previously published studies (32). Differences in the stability of CLIP may ultimately lead to a quantitatively larger pool of available, or peptide accessible I-E class II molecules within the thymus as compared with I-A. Therefore, with its absence in the DM-deficient mice, the difference in CLIP affinity between I-A and I-E may influence the presentation of self-peptides to developing thymocytes, with I-E preferentially driving selection as a result of its expression with a more diversified pool of peptides. This could significantly influence the isotype restriction of the CD4 T cell repertoire, creating a large emigrating population biased toward peptides presented within the context of I-E. Such a large I-E restricted peripheral T cell population could account for the observed enhancement to I-E presented eptiopes (HA110–120, MalE89–103, HEL107–118) in DM−/− mice. This is further supported by immunizations with individual I-A and I-E-restricted peptides. Such experiments presumably recruit any available CD4 T cells expressing polyclonal T cell receptors in the absence of either endosomal DM-mediated enhancement in peptide:class II loading, or competition from T cells of differing specificity. Therefore, the enhanced CD4 T cell responses to the I-Ed-restricted MalE89–104 and HEL103–118 peptides in DM−/− mice (shown by an increase in the magnitude of IL-2 responses) possibly reflect an increased preimmune repertoire. Recent reports using MHC class II tetramers are in support of such a hypothesis, demonstrating that when saturating doses of peptide are delivered in the absence of interclonal peptide competition, the total expansion of T cell populations parallel the CD4 T cell precursor number (62). Thus, an increased precursor frequency of I-E-restricted T cells may overcome or compensate for the decrease in I-E:peptide complexes on the surface of APC due to the absence of the DM protein following Ag uptake.
It is equally important to consider the preferential recognition of peptides within the context of I-A in WT mice, where I-Ad and I-Ed will both display a diverse repertoire of peptides during thymic selection events and during peripheral CD4 T cell responses to foreign Ags. One potential explanation for the observed bias toward I-A molecules is a higher cell surface expression of I-A class II molecules as compared with I-E. There is some evidence that the alternative isoforms of the MHC class II master-regulator CIITA may interact disproportionately with the class II isotypes, however, this is in support of enhanced HLA-DR (I-E) expression rather than HLA-DQ (I-A) (63, 64, 65, 66, 67, 68, 69). Furthermore, overall surface expression has been found to be equivalent between I-Ad and I-Ed (this article and Ref. (29)), making this an unlikely parameter contributing to the observed bias. A second mechanism that may account for the dominant I-A restriction bias in WT animals is a favored interaction between the CD4 coreceptor and the I-Ad molecule relative to I-Ed. Previous work in our laboratory (70) as well as several accounts in the literature (71, 72, 73, 74) support the theory of asymmetrical CD4 association with the different MHC isotypes, with I-A class II molecules displaying enhanced CD4 coreceptor interactions. This, along with the very interesting observation that several mouse strains have independently lost I-E gene expression (75), implicates that a fundamental difference in the use of MHC class II isotypes.
Overall, the results presented in this manuscript highlight a striking preference in WT mice for the recognition and response of the peripheral CD4 T cell pool to epitopes within the context of I-A, a phenomenon that is eliminated by the absence of DM expression, and subsequently superseded by a predominating I-E-restricted response in the DM−/−mice. It is interesting to speculate about the factors that may contribute to this observation, either directly or indirectly due to the presence or absence of the DM molecule itself. Although we have discussed only a few of the parameters that may contribute to the observed isotype-restriction bias, several others, such as kinetic stability of the peptide within the class II binding groove, affinity/avidity of the TCR repertoire for I-A vs I-E molecules or differential recruitment of regulatory T cells by I-A and I-E which may also be influenced by DM activity in APC, highlighting the importance of further analysis into isotype-specific DM requirements.
Independent of the mechanism(s) that may underlie the isotype preferences revealed in this study are the implications of our findings on immune responses and epitope discovery, particularly in humans. Due to the genetic complexity of MHC class II genes in humans, with multiple DR β loci and heterozygosity at alleles that encode MHC proteins, a typical human host might express as many as 10–15 individual class II molecules. In anticipating the peptides that may be responded to in humans, it is common to assume that expression of a particular MHC protein makes it a logical candidate for epitope screening or for study of T cell repertoire, using such strategies as MHC-peptide tetramer staining. Such logic has been applied in mapping epitopes in commonly expressed and biochemically tractable MHC molecules such as HLA-DR1. Indeed, examination of the National Institutes of Health epitope data base (76) (http://immuneepitope.org) reveals that >1500 peptides presented by HLA-DR1-0101 have been identified, whereas <100 epitopes identified are restricted to the most commonly studied DQ allele, HLA DQ-0301. Our results in WT mice suggest that it may be very ill advised and misleading to screen humans for expression of commonly expressed (or studied) class II molecules, and assume that the relative numbers of CD4 T cells present in that human are in any way related to the menu of MHC molecules expressed. Moreover, if the dramatic preferences in mice for I-A over I-E are due to relative affinities in interactions with the CD4 protein and whether this extends to human class II and CD4 proteins, then it is possible that a disproportionate fraction of the peripheral CD4 repertoire in humans is restricted to the poorly studied HLA-DQ proteins, analogous to the murine I-A molecules. Further analyses of these issues in TCR repertoire preferences using both human and HLA transgenic mice should be highly illuminating and are the subject of current investigations.
We thank Dr. Deborah Fowell for helpful comments on this 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 National Institutes of Health Grants R01 AI51542, R21 AI059898 (to A.J.S.), AII19047 (to E.K.B.), training Grants T32 AI 007285 (to P.R.M.), and T32 HL66988 (to S.A.J.), and by the Wellcome Trust.
Abbreviations used in this paper: DM, HLA-DM or H-2DM; CLIP, class II-associated invariant chain peptide; WT, wild type; LACK, Leishmania homologue of activated receptor for c-kinase; MalE, maltose-binding protein of Escherichia coli; HEL, hen egg lysozyme.