As organ-specific autoimmune diseases do not become manifest until well-advanced, interventive therapies must inhibit late-stage disease processes. Using a panel of immunogenic peptides from various β cell Ags, we evaluated the factors influencing the efficacy of Ag-based therapies in diabetes-prone NOD mice with advanced disease. The ability of the major β cell autoantigen target determinants (TDs) to prime Th2 responses declined sharply between 6 and 12 wk of age, whereas the ability of immunogenic ignored determinants (IDs) of β cell Ags to prime Th2 responses was unaffected by the disease process. The different patterns of TD and ID immunogenicity (even from the same β cell Ag) may be due to the exhaustion of uncommitted TD-reactive, but not ID-reactive, T cell pools by recruitment into the autoimmune cascade. Therapeutic efficacy was associated with a peptide’s immunogenicity and ability to promote Th2 spreading late in the disease process but not its affinity for I-Ag7 or its expression pattern (β cell specific/nonspecific or rare/abundant). Characterization of some IDs revealed them to be “absolute” cryptic determinants. Such determinants have little impact on T cell selection, leaving large precursor T cell pools available for priming by synthetic peptides. Traditional Ag-based therapeutics using whole autoantigens or their TDs cannot prime responses to such determinants. These findings suggest a new strategy for designing more efficacious Ag-based therapeutics for late-stage autoimmune diseases.

A promising strategy for inhibiting organ-specific autoimmune diseases has been based on identifying the autoantigens and their major target determinants (TDs)3 and administering them in modalities that induce regulatory responses (1, 2, 3, 4, 5, 6, 7). The induced regulatory responses spread rapidly to other target cell Ags, bypassing the spreading hierarchy along which T cell autoimmunity usually gradually spreads (8, 9). This rapid spreading of anti-inflammatory autoreactivity may exhaust the pool of uncommitted TD-reactive T cells that could otherwise contribute to the pathogenic response, as well as generate more extensive bystander suppression (9). This therapeutic approach is appealing because it may inhibit autoimmune diseases with little interference with immune system function. As human organ-specific autoimmune diseases do not become clinically apparent until the autoimmune processes is well advanced, Ag-based therapies should be rationally designed to induce strong regulatory responses late in an autoimmune disease processes. However, currently, there is little understanding of the factors that influence the efficacy of Ag-based therapies, especially when delivered at advanced stages of organ-specific autoimmune diseases.

The NOD mouse provides a useful animal model with which to study interventive strategies for spontaneous organ-specific autoimmune disease. We and others (10, 11, 12, 13, 14) have shown that several different β cell autoantigens are highly effective in preventing type 1 diabetes (T1D) in NOD mice when administered early in the disease process in modalities that prime anti-inflammatory Th2 responses. However, these treatments lose efficacy when administered at later stages of the disease process. There is some evidence suggesting that late in the NOD disease process, a greater autoantigen-induced Th2 response is associated with a better clinical outcome: 1) a β cell autoantigen treatment that induced greater Th2 responses prolonged the survival of syngenic islet grafts in diabetic NOD mice longer than autoantigens that induced weak Th2 responses (10); and 2) treatment of NOD mice with a combination of glutamic acid decarboxylase (GAD)65 peptides inhibited T1D development better than treatment with any of the peptides alone, and this protection required IL-4 production (15).

There are many open questions concerning how to better design Ag-based therapies to inhibit late-stage autoimmune disease. Should early or late targets of the spontaneous autoimmune response be administered? Does the expression pattern of the Ags (e.g., β cell specific vs ubiquitously expressed) affect its therapeutic efficacy? Could immunization with peptides having higher affinity for MHC class II improve therapeutic efficacy? Can Ags that are expressed in the target cell but are not targets of the autoimmune response also induce effective regulatory responses?

Another important factor to consider is how the kinetics of the ongoing autoimmune response affects the size of the naive T cell precursor pool that can be primed by Ag immunization. We previously observed that there is a progressive decline in the ability of autoantigen treatments to prime Th2 responses with autoimmune disease progression in NOD mice (16). We believe that this attenuation is due to the recruitment of TD-reactive T cells into the spontaneous autoimmune response, which exhausts the pool of uncommitted TD-reactive T cells that are available for priming by autoantigen treatment. Given this attenuation in the ability of autoantigens to prime responses, how can Ag-based therapies be designed to induce strong regulatory responses late in an autoimmune disease process? We hypothesized that there may be β cell Ag determinants that are ignored by the autoimmune response, which have had little impact on T cell selection and should have large pools of uncommitted T cells available at late stages of the disease process. Theoretically, the ability of such ignored determinants (IDs) to induce Th2 immunity should not attenuate during the disease process. Conceivably, once activated, these T cells can respond to determinants that they had functionally ignored as naive cells. For example, T cells reactive to cryptic determinants of myelin ignore these determinants until they are experimentally activated by immunization with a peptide containing a cryptic myelin determinant, after which they recognize the presented determinant, promote the spreading of autoimmunity among myelin Ags, and can mediate experimental allergic encephalomyelitis (EAE) (17, 18).

However, it may be difficult to identify β cell Ags that are ignored by the spontaneous autoimmune response and have large pools of uncommitted cognate T cells. The majority of proteins expressed in β cells appears to be ignored by the spontaneous autoimmune response. This suggests that T cells recognizing most β cell Ag dominant/subdominant determinants have been inactivated/deleted, leaving only low avidity cognate T cells that fail to activate and expand even in the context of inflammation (18, 19, 20). Large high-avidity precursor T cell pools should exist to the cryptic determinants of β cell Ags, which are poorly presented and have had little impact on T cell selection (20). However, in the context of the inflammation, the presentation of cryptic determinants often becomes sufficient to activate cognate T cells, such that cryptic determinants become TDs and contribute to disease pathogenesis (20, 21). Hence, the prevalence of β cell Ag determinants that are both highly immunogenic and ignored by the autoimmune process (i.e., are IDs) is an open question. Moreover, ID-based therapies may be no more effective than TD-based therapies in older NOD mice because: 1) a highly pathogenic autoreactive response has already built up; and 2) there should be little spreading of induced Th2 responses to TDs because the autoimmune process has exhausted the pool of naive T cells that recognize β cell TDs.

Herein, we describe our identification of immunogenic IDs within an autoantigen (GAD65) as well as in other (nontargeted) β cell Ags. The selected IDs came from Ags with different expression patterns (i.e., β cell specific and nonspecific, rare and abundant). We tested these IDs, as well as the major early and late TDs of β cell autoantigens, for their affinity for I-Ag7, their ability to prime Th2 responses and Th2 spreading, and their ability to prevent diabetes in NOD mice with advanced autoimmune disease. We discuss the insights our findings provide toward the rational design of Ag-based therapies for late-stage organ-specific autoimmune disease.

NOD mice (Taconic Farms) were bred under specific pathogen-free conditions. Only female mice were used for these studies. The University of California, Los Angeles, Animal Research committee approved all animal care and experimental procedures.

The panel of TDs from major NOD mouse β cell autoantigens included HSPp277 (22), insulin (9–23) (13), GAD (524–543) (also termed GADp35), GAD (530–543), and GAD (217–236) (also termed GADp15) (23, 24, 25). All of these peptides contain dominant TDs and have previously been shown to efficiently prevent T1D when administered early in the NOD mouse disease process (Refs. 14 , 15 , and 22 and our unpublished observations).

To screen for immunogenic GAD65 determinants that were ignored by the spontaneous autoimmune response, we tested the immunogenicity of synthetic peptides from an overlapping set of 20-mer peptides spanning GAD65 with 5-aa overlaps (previously described in Refs. 23 and 24). Six-week-old NOD mice were immunized with peptide/CFA and tested for recall responses as described below. We did not test GAD65 peptides known to be targets of autoreactive T cells (23). We identified two IDs from GAD65: GAD (260–279) (also known as GADp18, PEVKEKGMAALPRLIAFTSE) and GAD (398–420) (also known as GADp27, VPLQCSALLVREEGLMONCNQ) (23, 24).

Using a consensus I-Ag7-binding sequence as a guide (26, 27, 28), we also synthesized peptides containing candidate IDs from different β cell Ags, which are not known to be targets of autoreactive T cells in NOD mice and which have very different patterns of tissue expression. Three peptides were synthesized from mouse calbindin, which is ubiquitously expressed at relatively high levels (calbindin (101–114) EEFMKTWRKYDTDHS, calbindin (122–141), ELKNFLKDLLEKANKTVDDT, and calbindin (214–228) LKDLCEKNKQELDIN); two peptides from mouse reduced expression in cancer (REC) protein, which is widely expressed at relatively low levels (REC (133–148), LIKPDRCHHCSVCDKC, and REC (149–163), ILKMDHHCPWVNNCV); one peptide from mouse neuropeptide Y precursor, which is widely expressed and secreted (45–59) MARYYSALRHYINLI; and three from mouse islet amyloid polypeptide, which is only expressed in β cells and is secreted ((IAPP (19–33) NHLRATPVRSGSNPQ, (IAPP (46–60), TQRLANFLVRSSNNL, and IAPP (75–91) GKRNAAGDPNRESLDFL).

An immunogenic hen egg lysozyme (HEL) peptide (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25) (29) was used as a control foreign peptide, and a mouse serum albumin (MSA) peptide (560–574) (26), which contains a nontargeted determinant, was used as a control self-Ag.

The TD and ID peptides, except GAD (260–279), were synthesized by NeoMPS at ≥95% purity. GAD (260–279) was synthesized by Biosynthesis at >95% purity. Whole mouse GAD65 was purified as described previously (23).

NOD mice were immunized at the indicated age in their hind footpads with 200 μg of Ags in CFA (Invitrogen Life Technologies). Nine to 10 days later, lymph node mononuclear cells were plated in 96-well microtiter plates at 8 × 105 cells/well in HL-1 medium (BioWhittaker) and supplemented with antibiotics and 1% FCS (HyClone), in the presence of a dose range (2–20 μM) of each test Ag. Medium alone was used as the negative control, and purified PPD (10 μg/ml) or anti-CD3 (1 μg/ml) was used as the positive control for each mouse. During the last 12 h of the 96-h culture period, 1 μCi of [3H]thymidine was added into each well. Incorporation of label was measured by liquid scintillation counting.

Six- and 12-wk-old female NOD mice were injected i.p. with 100 μg of the indicated Ag peptide in IFA. Ten days later, the treatment was repeated. After another 10 days, the frequency of Ag-specific splenic T cells secreting IL-4 or IFN-γ was determined using a modified ELISPOT technique as described previously (10, 30). Briefly, 106 splenic mononuclear cells were added/well (in triplicate) of an ELISPOT plate (Millipore) that had been coated with cytokine capture Abs and incubated with peptide (20 μM) for 24 h for IFN-γ or 40 h for IL-4 detection. After washing, biotinylated detection Abs were added, and the plates incubated at 4°C overnight. Bound secondary Abs were visualized using HRP-streptavidin (DakoCytomation) and 3-amino-9-ethylcarbazole. Abs 11B11/BVD6-24G2-biotin and R4-6A2/XMG 1.2-biotin (BD Pharmingen) were used for capture and detection of IL-4 and IFN-γ, respectively. Groups of two to four experimental and control mice were tested simultaneously in at least two separate experiments for a total of four to eight mice in each group.

To determine peptide affinity for NOD I-Ag7, a binding assay was performed on each of the tested Ags as described previously (27). The peptide affinity for I-Ag7 was expressed as the peptide concentration required to inhibit the binding of biotinylated HEL (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25) peptide to purified I-Ag7 molecules by 50% (IC50). The peptides were simultaneously analyzed in at least three experiments.

Groups of 10–15 female NOD mice were immunized i.p. at 12 wk of age with the indicated control or β cell Ag peptide (100 μg of peptide in 50% IFA). The animals were boosted once 10 days later. Additional control groups received IFA alone or were unmanipulated. The mice were tested three times weekly for urine glucose levels by Tes-tape (Lilly Research Laboratories). After detection of abnormal glucose levels in the urine, blood glucose levels was monitored daily by testing tail blood with a Glucometer Elite instrument (Bayer). Two consecutive daily blood glucose levels > 300 mg/dl were considered to be disease onset. The mice were monitored up to 45 wk of age, and the data were analyzed using Kaplan-Meier survival analysis with the log-rank test (SAS Institute).

To identify β cell Ags that were both immunogenic and ignored by the spontaneous immune response, we tested synthetic peptides of β cell Ags rather than whole β cell Ags because only by bypassing whole Ag processing could we study cryptic determinants. By using synthetic peptides we also avoided contaminants that could copurify with whole Ags isolated from different tissues or recombinant cells that would confound the interpretation of our results.

We screened for IDs within GAD65 by testing a set of peptides, which span GAD65 for immunogenicity in NOD mice (23). Female NOD mice were immunized at 6 wk of age with a peptide (200 μg in CFA) and tested for lymph node mononuclear cell recall responses. We identified two peptides, mouse GAD (260–279) and GAD (398–420), which were highly immunogenic in NOD mice (stimulation indexes (SIs) of 68 and 57, respectively; Fig. 1,A). These peptides did not elicit proliferative splenic T cell responses in our previous study of spontaneous T cell responses during the development of NOD mice (23). To further verify that these immunogenic GAD65 peptides were ignored by the spontaneous autoimmune response, we again tested for spontaneous T cell responses, using both a proliferation and ELISPOT assays. Consistent with our previous study (23), both assays detected spontaneous T cell responses to the TD GAD (524–543) early in NOD development but did not detect significant responses to GAD (260–279) or GAD (398–420) in female NOD mice tested from 3 to 18 wk in age (Fig. 1, B and C). These observations define GAD (260–279) and GAD (398–420) as GAD65 IDs.

FIGURE 1.

Identification of candidate IDs. A, Six weeks old NOD mice were immunized with 200 μg peptide/CFA in a hind footpad, and 10 days later, lymph node mononuclear cells were tested for recall responses over a dose range of Ags. Data shown is the mean SI ± SEM at the optimal Ag concentration (n = 3/group). B and C, Longitudinal analysis of spontaneous responses to GAD65 determinants in unmanipulated NOD mice. The TD GAD (524–543), but not GAD (260–279) or GAD (398–420), elicits responses from splenic mononuclear cells using proliferation (B) and ELISPOT (C) assays.

FIGURE 1.

Identification of candidate IDs. A, Six weeks old NOD mice were immunized with 200 μg peptide/CFA in a hind footpad, and 10 days later, lymph node mononuclear cells were tested for recall responses over a dose range of Ags. Data shown is the mean SI ± SEM at the optimal Ag concentration (n = 3/group). B and C, Longitudinal analysis of spontaneous responses to GAD65 determinants in unmanipulated NOD mice. The TD GAD (524–543), but not GAD (260–279) or GAD (398–420), elicits responses from splenic mononuclear cells using proliferation (B) and ELISPOT (C) assays.

Close modal

Conceivably, spontaneous autoimmune responses may ignore dominant, subdominant, or cryptic autoantigen determinants (hence, we have not assumed a priori that all IDs are cryptic determinants). To begin to understand why the GAD65 IDs are immunogenic yet ignored by spontaneous autoimmune responses, we tested whether they contained dominant or cryptic determinants. We immunized NOD mice with whole mouse GAD65 and tested for recall response to the GAD65 IDs and vice versa. Mice immunized with whole GAD65 displayed strong recall responses to whole GAD65 but no detectable recall responses to GAD (260–279) or GAD (398–420) (Fig. 2). Mice immunized with GAD (260–279) or GAD (398–420) mounted strong recall responses to the injected peptide but none to whole GAD65 (Fig. 2). These results define GAD (260–279) and GAD (398–420) as “absolute” cryptic determinants (see Discussion).

FIGURE 2.

GAD (260–279) and GAD (398–420) are cryptic determinants. NOD mice (8–10 wk of age) were immunized with the indicated Ag (200 μg/CFA in the hind footpad), and 10 days later, lymph node monocular cells were tested for recall responses to the indicated Ag over a dose range. Data shown are mean stimulation index ± SD (n = 4 mice/group).

FIGURE 2.

GAD (260–279) and GAD (398–420) are cryptic determinants. NOD mice (8–10 wk of age) were immunized with the indicated Ag (200 μg/CFA in the hind footpad), and 10 days later, lymph node monocular cells were tested for recall responses to the indicated Ag over a dose range. Data shown are mean stimulation index ± SD (n = 4 mice/group).

Close modal

To expand the panel of β cell Ag IDs, we searched cDNA and protein sequence databases for mouse cDNA and protein sequences that are expressed in β cells (31, 32). We chose several β cell Ags with different expression patterns for study. Theoretically, even a widely expressed abundant protein may contain determinants that are immunogenic and ignored by the autoimmune response.

Using a consensus I-Ag7-binding sequence as a guide (26, 27, 28), we synthesized three peptides from mouse calbindin (an abundant intracellular protein expressed in many tissues), two peptides from the mouse REC protein (which is expressed in many tissues at relatively low levels), one peptide from mouse neuropeptide Y (which is widely expressed and secreted), and three peptides from mouse islet amyloid polypeptide (IAPP, which is only expressed in β cells and is secreted) (see Materials and Methods and Table I).

Table I.

Properties of β cell Ag TDs, IDs, and control Ags

PeptideIC50 μMTD/IDDominant/CrypticExpression PatternReference
Insulin (9–23) 0.6 TD Dominant β Cell specific, secreted 1357  
HSPp277 10 TD Dominant Ubiquitous 22  
GAD (217–236) 0.7 TD Dominant Low level, in a few tissues 15  
GAD (524–543) 0.8 TD Dominant Low level, in a few tissues 2358  
GAD (260–279) 0.5 ID Cryptic Low level, in a few tissues 23 , herein 
GAD (398–420) 0.5 ID Cryptic Low level in a few tissues 23 , herein 
Calbindin (122–141) ID ND Ubiquitous Herein 
IAPP (19–33) ID ND β Cell specific, secreted Herein 
IAPP (46–60) >50 ID ND β Cell specific, secreted Herein 
REC (133–148) >50 ID ND Low level, in many tissues Herein 
HEL (11–23) 0.7 NA Dominant Foreign 29  
MSA (560–574) 0.4 NA Dominant Secreted self-Ag 26  
PeptideIC50 μMTD/IDDominant/CrypticExpression PatternReference
Insulin (9–23) 0.6 TD Dominant β Cell specific, secreted 1357  
HSPp277 10 TD Dominant Ubiquitous 22  
GAD (217–236) 0.7 TD Dominant Low level, in a few tissues 15  
GAD (524–543) 0.8 TD Dominant Low level, in a few tissues 2358  
GAD (260–279) 0.5 ID Cryptic Low level, in a few tissues 23 , herein 
GAD (398–420) 0.5 ID Cryptic Low level in a few tissues 23 , herein 
Calbindin (122–141) ID ND Ubiquitous Herein 
IAPP (19–33) ID ND β Cell specific, secreted Herein 
IAPP (46–60) >50 ID ND β Cell specific, secreted Herein 
REC (133–148) >50 ID ND Low level, in many tissues Herein 
HEL (11–23) 0.7 NA Dominant Foreign 29  
MSA (560–574) 0.4 NA Dominant Secreted self-Ag 26  

We then tested whether these peptides were immunogenic in NOD mice by immunizing mice and testing proliferative recall responses to the injected Ag over a dose range. Five of the nine synthesized peptides were not immunogenic. Three of the peptides were quite immunogenic (IAPP (19–33), IAPP (46–60), and REC (133–148)), with recall response SIs ranging from 16 to 48 (Fig. 1 A). One peptide, calbindin (122–141), was very weakly immunogenic (SI = 4) and was retained in our test panel as an internal control of a weakly immunogenic β cell Ag.

Next, we tested whether these four immunogenic peptides were ignored by the spontaneous autoimmune response in NOD mice using both proliferation and ELISPOT assays. There were no detectable spontaneous T cell responses to these four peptides in female NOD mice when tested from 3 to 18 wk of age (data not shown), thereby defining them as IDs of their respective β cell Ags.

Purified REC and IAPP are not commercially available, so to determine whether their IDs contained dominant or cryptic determinants, we subcloned their cDNAs into bacterial expression vectors and purified their respective proteins using several different chromatographic techniques. However, the purified proteins always induced some nonspecific T cell responses such that we were not able to unambiguously determine whether whole REC and IAPP fail to recall responses primed by their IDs.

Studies of immune responses to foreign determinants have shown that a major factor defining their immunogenicity is the efficiency with which they are displayed on MHC (33). However, the same factors that cause a foreign determinant to be immunogenic should promote the induction of T cell tolerance to self-determinants. To evaluate the factors influencing the therapeutic efficacy of TDs and IDs in Ag-based therapies, we determined the affinity of each peptide for I-Ag7. We found that the peptides containing control non-β cell determinants, as well as TDs and IDs, displayed a range of binding affinities for I-Ag7 (Table I). There was no correlation between the affinity of the peptide for I-Ag7, its immunogenicity, or whether it was a TD or an ID. These findings are consistent with previous reports that the ability of autoantigen determinants to prime immune responses in NOD mice is unrelated to their binding affinity for I-Ag7 (15, 34).

We tested the ability of each ID to induce Th2 responses in NOD mice shortly after the onset of autoimmunity and in NOD mice with advanced disease. We also tested the major TDs of well-characterized β cell autoantigens. The panel of autoantigen TDs included insulin (9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23), HSPp277, GAD (524–543), and GAD (206–217), all of which are dominant determinants that have been shown previously to inhibit T1D when administered early in the NOD mouse disease process (Refs. 14 , 15 , and 22 and our unpublished observations). GAD (524–543) contains an early TD of the spontaneous autoimmune response, HSPp277 contains an intermediate target, and insulin (9–23) and GAD (206–217) contain late TDs (23, 25, 35). Control immunogenic Ags included HEL (11–23) and the nontargeted self-Ag MSA (560–574) (26, 29). The characteristics of the panel of peptides are listed in Table I.

We found that immunization with HEL (11–23) and MSA (560–574) primed a similar number of Ag-specific, IL-4-secreting T cells when NOD mice were immunized at 6 or 12 wk of age (Fig. 3). Hence, autoimmune disease progression has little effect on the ability of these control Ags to prime Th2-type immune responses.

FIGURE 3.

The ability of IDs to induce Th2 responses is unaffected by disease progression in NOD mice. NOD mice were immunized either at 6 or 12 wk of age with the indicated Ag in IFA i.p. Subsequently, the frequency of IL-4-secreting splenic T cells responding to the injected Ag was determined by ELISPOT. Data shown are the mean number of IL-4-secreting SFCs (n = 8 mice/group).

FIGURE 3.

The ability of IDs to induce Th2 responses is unaffected by disease progression in NOD mice. NOD mice were immunized either at 6 or 12 wk of age with the indicated Ag in IFA i.p. Subsequently, the frequency of IL-4-secreting splenic T cells responding to the injected Ag was determined by ELISPOT. Data shown are the mean number of IL-4-secreting SFCs (n = 8 mice/group).

Close modal

Although treatments using the major TDs of β cell autoantigens induced vigorous Th2 responses in 6-wk-old NOD mice, their ability to induce Th2 immunity in 12-wk-old mice was greatly reduced (Fig. 3). For example, the TDs of GAD (GAD (217–236) and GAD (530–543)) primed 55 and 45 Ag-specific, IL-4 secreting colonies/million splenic T cells (respectively) when administered at 6 wk of age but only 11 colonies/million when administered at 12 wk of age, an 80 and 75% reduction (respectively). The rapid decline in the ability of the immunodominant determinants of different β cell autoantigens to prime Th2 responses with disease progression may reflect the recruitment and commitment of their cognate T cells to the spontaneous proinflammatory autoimmune response, which depletes the T cell repertoire of naive cognate T cells that can be primed by autoantigen treatment.

In contrast, we observed that the ability of all six IDs to prime Th2 responses in NOD mice remained constant, or slightly increased, between 6 and 12 wk of age (Fig. 3). Thus, as with control Ags HEL (11–25) and MSA (560–574), the ability of IDs to prime anti-inflammatory responses was not affected by autoimmune disease progression. It is notable that at 6 wk of age several TDs primed more frequent Th2 responses than IDs, but by 12 wk of age, all of the tested IDs induced greater Th2-type responses than the TDs (Fig. 3).

It is also interesting that within a single autoantigen, GAD65, that both TDs ((GAD (530–543) and GAD (217–236)) and IDs (GAD (260–279) and GAD (398–420)) induced vigorous Th2 responses when injected at 6 wk of age. However, at 12 wk of age, the ability of the GAD TDs to prime Th2 responses was reduced by 80–75%, whereas the ability of the GAD IDs to induce Th2 responses had increased by ≈10% (Fig. 3). Thus, the “immunogenicity” of different determinants from the same protein can change dramatically over the course of the autoimmune disease process. These data further support our contention that the autoimmune processes depletes the availability of uncommitted T cells reactive to targeted determinants but that large pools of precursor T cells may be available at late stages of the disease process to β cell Ag IDs.

NOD mice were treated with a control, TD, or ID peptide at 12 wk of age (100 μg i.p. in IFA). This treatment was repeated 10 days later, and 10 days later, their splenic mononuclear cells were tested for IL-4 and IFN-γ responses to other (uninjected) β cell Ags. Previous studies have shown that at 12 wk of age NOD mice already have well-established, Th1-polarized T cell responses to all of the tested TDs (23, 25, 36). Mice immunized with control HEL (11–25) or MSA (560–574) developed Th2 responses to the injected Ag (i.e., IL-4 spot-forming colonies (SFCs)) but not to any of the tested β cell Ags (Fig. 4), indicating that the Th2 responses primed by these control Ags did not spread to β cell Ags, which is consistent with previous studies (8, 9). In mice treated with a TD, we detected very little spreading of Th2-type autoimmunity to other TDs (Fig. 4). We did not detect spreading of Th2 immunity to any ID (data not shown).

FIGURE 4.

Late in the disease process, immunogenic IDs promote more spreading of Th2 autoimmunity to other β cell Ags. NOD mice treated with the indicated Ag at 12 wk of age (100 μg i.p. in IFA) and tested 3 wk later for the frequency of IL-4-secreting T cells reactive to the injected Ag, as well as other TDs, IDs, and control Ags (n ≥ 8 mice/group). Only spreading responses to uninjected TDs are shown. There was no detectable spreading after immunization with any TD to any ID (data not shown),

FIGURE 4.

Late in the disease process, immunogenic IDs promote more spreading of Th2 autoimmunity to other β cell Ags. NOD mice treated with the indicated Ag at 12 wk of age (100 μg i.p. in IFA) and tested 3 wk later for the frequency of IL-4-secreting T cells reactive to the injected Ag, as well as other TDs, IDs, and control Ags (n ≥ 8 mice/group). Only spreading responses to uninjected TDs are shown. There was no detectable spreading after immunization with any TD to any ID (data not shown),

Close modal

ID-induced Th2 responses spread to TDs but not any other IDs (Fig. 4). In general, IDs induced more extensive spreading of Th2 immunity to TDs (compared with TD treatments). However, even ID-induced Th2 spreading was of low magnitude, as expected due the presumed depletion of the naive TD-reactive T cell pools at this late stage. Thus, ID treatment can promote more extensive spreading of Th2 immunity, and this infectious Th2 autoimmunity is confined to the β cell Ag TDs that are the natural targets of spontaneous autoimmunity.

We did not discern any significant differences in the magnitude of TD-specific IFN-γ responses in control, TD-, and ID-treated mice (data not shown). This is not surprising given that at the time of treatment the mice already had well-established Th1 responses to the TDs, and their splenic T cells were tested only 3 wk after treatment. This finding is also consistent with previous findings that Ag/IFA-induced Th2 responses have little short-term effect on the magnitude of already established Th1 responses (37, 38). It will be of interest to further characterize Th1 responses at later time points in treated mice that do and do not progress to T1D. Although most studies have observed reduced Th1 responses long-term after Ag-based therapy (e.g., Refs. 10 and 39), in some cases, Th1 responses have been associated with protection from autoimmune disease (40, 41).

Control NOD mouse groups that were untreated, treated with IFA alone, or MSA (560–574) at 12 wk of age had similar rates of spontaneous T1D (Fig. 5). Although all of the TDs in our test panel can effectively inhibit T1D in NOD mice when administered early in the disease process, when administered late in the disease process, none of them provided significant protection from disease compared with the control groups (Fig. 5). In contrast, treatment with each of the IDs reduced the overall disease incidence relative to unmanipulated NOD mice (Fig. 5)—the only exception being calbindin (122–141), which was included in the test panel as an internal control ID with low immunogenicity. Thus, at late stages of the disease process, only β cell Ag determinants that could prime strong Th2 responses displayed any therapeutic efficacy.

FIGURE 5.

T1D incidence in NOD mice immunized with a TD, ID, or control Ag. Mice were immunized at 12 wk of age with the indicated Ag (n = 10 mice/vaccinated group, n = 15 for unmanipulated group). ∗, p < 0.05 and ∗∗, p < 0.01 relative to unmanipulated NOD mouse group.

FIGURE 5.

T1D incidence in NOD mice immunized with a TD, ID, or control Ag. Mice were immunized at 12 wk of age with the indicated Ag (n = 10 mice/vaccinated group, n = 15 for unmanipulated group). ∗, p < 0.05 and ∗∗, p < 0.01 relative to unmanipulated NOD mouse group.

Close modal

There was a correlation between an ID’s immunogenicity (see Fig. 1,A) and its therapeutic efficacy (i.e., the T1D incidence in immunized mice at 45 wk of age; see Fig. 5) (R = 0.87, p = 0.02; Fig. 6,A). There is also a correlation between the ability of a TD or an ID to prime Th2 responses at 12 wk of age (i.e., the number of IL-4 SFC in Fig. 3) and its therapeutic efficacy (R = 0.86, p = 0.002; Fig. 6 B)).

FIGURE 6.

Correlation between the ability of a β cell Ag to prime responses in NOD mice with advanced pre-T1D and the treatment’s therapeutic efficacy. A, Correlation between an ID’s immunogenicity (as shown in Fig. 1) and its therapeutic efficacy (T1D incidence at 45 wk in Fig. 5) (R = 0.87, p = 0.02). B, Correlation between the ability of a TD or an ID to prime Th2 responses at 12 wk of age (i.e., the number of IL-4 SFC in Fig. 3) and its therapeutic efficacy (R = 0.86, p = 0.002).

FIGURE 6.

Correlation between the ability of a β cell Ag to prime responses in NOD mice with advanced pre-T1D and the treatment’s therapeutic efficacy. A, Correlation between an ID’s immunogenicity (as shown in Fig. 1) and its therapeutic efficacy (T1D incidence at 45 wk in Fig. 5) (R = 0.87, p = 0.02). B, Correlation between the ability of a TD or an ID to prime Th2 responses at 12 wk of age (i.e., the number of IL-4 SFC in Fig. 3) and its therapeutic efficacy (R = 0.86, p = 0.002).

Close modal

The traditional approach for developing Ag-based therapies for organ-specific autoimmune diseases has been to identify the autoantigens and their TDs, and then use these Ags to induce regulatory responses. Although immunization with the TDs used in this study efficiently prevents T1D when administered to young NOD mice (10, 11, 12, 13, 14), these treatments lose their efficacy when administered at late stages of the disease process. In the present study, we evaluated how Ag-based therapies could be better geared toward inhibiting advanced autoimmune disease, which is the clinically relevant situation. In particular, we hypothesized that β cell Ags that were highly immunogenic and ignored by the autoimmune response may have large pools of naive T cells available and may induce larger magnitude regulatory responses at late stages of an autoimmune disease. Consistent with this hypothesis, a previous study noted that a peptide of heat shock protein 60 that was not a target of spontaneous autoimmunity had some therapeutic efficacy in NOD mice (42).

To identify NOD mouse β cell Ag determinants that were both immunogenic and ignored by the spontaneous autoimmune response, we screened peptides from GAD65, as well as from β cell Ags that are not targets of autoreactive T cells. Many of the tested peptides were not immunogenic, suggesting that they did not bind MHC well or that the self-determinants had effectively inactivated/deleted reactive T cells. We did not detect spontaneous autoimmune responses to any of the tested peptides from REC, neuropeptide Y, calbindin, or IAPP during NOD mouse development, suggesting that the spreading of autoimmunity may be limited to a relatively small number of β cell autoantigens. We were readily able to identify immunogenic IDs, even in the abundant and secreted IAPP and in the widely expressed REC protein. We refer to these determinants as “IDs” rather than cryptic determinants because it is possible that some immunogenic β cell Ag-dominant determinants may be ignored by the spontaneous autoimmune responses (and hence are IDs), and because many cryptic determinants become involved in the autoimmune response (and hence are TDs). Based on the high immunogenicity of some of the IDs, it does not appear that these IDs escape recognition by the autoimmune response due to having purged cognate T cells.

Although TDs of β cell autoantigens induced vigorous Th2 responses when administered to NOD mice at 6 wk of age, they primed only weak Th2 responses when administered at 12 wk of age, confirming and extending previous observations (16). Indeed, in 12-wk-old NOD mice, TD-primed responses were only just above background levels, regardless of whether the TD contained an early or late target of the spontaneous autoimmune response. This rapid attenuation in the ability of TDs to prime responses may explain some inconsistencies in the literature concerning the dominant determinants of GAD65 in NOD mice. For example, GAD (524–543) and GAD (255–269) (which has sequence similarity with Coxsackievirus) can elicit recall responses after immunizing 3-wk-old NOD mice with whole GAD65, indicating that they contain dominant/subdominant determinants (Ref. 43 and unpublished observations). In contrast, peptides containing these regions do not elicit discernible recall responses after immunizing 9-wk-old NOD mice with whole GAD65 (44). Our findings suggest that immunization/recall assays aimed at determining the dominant determinants of whole autoantigens should be conducted before autoimmunity arises to the Ag of interest.

In contrast to TDs, the ability of β cell Ag IDs to prime Th2 responses was unaffected by disease progression. Notably, within a single β cell Ag, GAD65, the ability of its TDs to prime Th2 responses declined by 80–75% between 6 and 12 wk of age, whereas the ability of its IDs to prime Th2 immunity increased slightly (Fig. 3). These data demonstrate that the immunogenicity of different determinants within an autoantigen can change dramatically during the course of autoimmune disease. The clear attenuation of inducible Th2 immune responses to TDs, but not IDs, from within the same β cell protein supports the notion that the autoimmune disease process exhausts the pool of naive TD-reactive T cells available for priming by Ag-based therapy. Because ID-reactive T cells do not become involved in the disease process, some IDs may have large precursor pools of cognate T cells available for priming by Ag-based therapy at advanced stages of the disease.

Along with priming larger magnitude Th2 responses, the IDs also promoted greater spreading of Th2 immunity to TDs but not to any other IDs. This spreading of Th2 immunity was very limited, as is expected if there is a paucity of naive TD-reactive T cells at late stages of the disease. Nevertheless, this spreading of Th2 immunity among TD-reactive T cells may be an important factor for therapeutic efficacy because it both increases bystander suppression and exhausts the pool of autoantigen-reactive T cells that could otherwise develop toward a pathogenic phenotype (9). We did not observe significant changes the frequency of IFN-γ responses to TDs, which is not unexpected because NOD mice have well-established Th1 responses to TDs at 12 wk in age, and we characterized T cell responses a short time after treatment. We also examined whether ID treatments induced CD4+CD25+ T cells that can be potent regulators of autoimmunity (6, 45, 46). FACS analysis of splenic T cells from ID and control Ag-treated mice did not detect any significant differences in their frequencies of CD4+CD25+ T cells when tested 3 wk postimmunization (data not shown). Whether CD4+CD25+ T cell frequencies subsequently increase, or whether other regulatory cell types (e.g., Tr1 and Th3 (2, 3)) are involved in the protective mechanism, remains to be investigated.

We found that NOD mice treated at 12 wk of age with a major TD from key β cell autoantigens developed spontaneous T1D at essentially the same rate as control groups that were untreated, treated with MSA (560–574), or IFA alone (Fig. 5). The lack of TD therapeutic efficacy in NOD mice with advanced disease is likely to be due to the large pathogenic response that is already established, the paucity of TD-reactive T cells available for priming, and the reduced spreading of anti-inflammatory responses. In contrast, all of the selected IDs of β cell Ags displayed significant protection compared with both control and TD-treated groups (with the exception of calbindin (122–141), which was included in the test panel as an internal control for a weakly immunogenic ID).

The immunogenicity of each ID, as well as the extent to which each TD and ID induced Th2 immunity at 12 wk of age, correlated with the efficacy of that Ag-based therapy to inhibit T1D development (Fig. 6). This suggests that these parameters may be used to predict the therapeutic efficacy of a candidate β cell Ag treatment to inhibit late-stage disease in NOD mice. However, this contention requires additional testing, as for example, other cytokines such as IL-10 and TGF-β have been associated with Ag-based therapy-induced immunoprotective responses (3, 47, 48), and in some studies, protection from autoimmune disease has been associated with Th1 responses (40, 41). Nevertheless, in contrast to a study in which transfusion of Th2-polarized cells led to β cell damage in NOD-scid mice (49), our study suggests that in immunocompetent NOD mice the extent of induced Th2 immunity correlates with the extent of long-term protection from disease.

Characterization of the GAD65 IDs revealed that they are cryptic determinants; whole GAD65 did not prime responses to these determinants, but these determinants are very immunogenic when administered in synthetic peptides that bypass the natural processing of whole Ags (20). Such cryptic determinants should have little impact on T cell selection, which explains why they are highly immunogenic. The IDs from REC and IAPP may also be cryptic determinants (see below), but we cannot exclude the possibility that they contain dominant/subdominant determinants. The different IDs we studied displayed a range of affinities for I-Ag7, which is expected if they are ignored due to poor processing from whole Ags. This is also consistent with a previous study of TDs, which found no correlation between the binding affinity of TD for I-Ag7 and its immunogenicity (15, 34).

The protective mechanism in ID-immunized mice may have parallels with the pathogenic process underlying EAE. EAE can be initiated by immunization with a cryptic determinant of a myelin Ag. Once experimentally activated in the periphery, T cells reactive to the cryptic myelin determinant can then respond to Ags in the CNS that they were previously ignoring (20). Activated T cells are stimulated at much lower levels of peptides than when they are naive, express higher levels of accessory molecules, and require less TCR ligations than naive cells (50, 51, 52, 53). These activated T cells then promote the spreading of (Th1-type) autoimmunity to other myelin Ags, leading often to EAE. In a similar manner, although naive ID-reactive T cells ignore their cognate β cell Ag, once they are activated by immunization with the ID, they can respond to presented Ag in the islets/lymph nodes and promote the spreading of autoimmunity (in this case, Th2-type) to other β cell TDs.

Studies of EAE have shown that the spreading of autoimmunity to cryptic myelin determinants is important for disease pathogenesis (21). It is thought that in the context of inflammation some cryptic myelin determinants are more efficiently presented, leading to the activation and expansion of cognate T cells such that these cryptic determinants become TDs. The cryptic determinants that become targets of autoreactive T cells have been termed “latent” cryptics (54). The ability of such latent cryptic determinants to prime Th2 responses should, as with other TDs, attenuate with disease progression. In contrast, the GAD65 IDs we have identified are “absolute” cryptic determinants (54), which do not become involved in the autoimmune cascade. In unmanipulated NOD mice, these IDs may be processed from whole Ags and presented to a limited extent, but it is insufficient to prime naive cognate T cells during the spontaneous autoimmune process. However, their presentation may be sufficient to elicit responses from regulatory T cells that are activated by ID immunization. Alternatively, the vaccine-induced T cells may exert their protective action indirectly through their ability to spread regulatory responses to TD-reactive T cells. This “infectious tolerance” (36, 55) can arise from Ag-primed T cells “educating” APCs (56). The educated APCs then influence naive T cells, which recognize other Ags, to produce the same cytokines as those produced by the Ag-primed T cells (Ref. 56 and also Ref. 37). This both increases bystander suppression and helps exhaust the TD-reactive T cell precursor pool.

Because Th2 autoimmunity did not spread to any of the tested IDs, it suggests that the presentation of IDs by educated APCs is insufficient to activate ID-reactive T cells. This, together with being ignored by the spontaneous autoimmune response, despite having large naive precursor T cell pools, is consistent with the IDs being absolute cryptic determinants. Immunization with peptides containing IDs bypasses whole Ag processing, allowing the absolute cryptic determinants to prime large cognate T cell precursor pools that exert regulatory functions through one of the above mechanisms.

We chose to study IDs from β cell Ags with different patterns of expression and to give the mice a minimum number of treatments in order to obtain a proof-of-principle. Immunogenic β cell Ag IDs appear to be commonplace, and a new catalog of thousands of cDNAs expressed in β cells should allow the identification of many additional β cell Ag IDs (32). Because there are no pre-existing spontaneous autoimmune responses to IDs, it may be easier to control the phenotype of ID-primed immune responses by the adjuvant and route of administration (as opposed to giving TDs to animals with established TD-reactive pathogenic responses). Accordingly, ID administration may be inherently safer than whole autoantigen or TD-based therapies.

In conclusion, we found that while the ability of TDs to prime Th2 responses attenuates during disease progression, the ability of immunogenic IDs to prime Th2 immunity was unaffected by disease progression—even when the IDs and TDs are from the same autoantigen. Late in the disease process, immunogenic ID treatments promoted greater spreading of Th2 immunity to TDs and more efficiently inhibited disease progression. Therapeutic efficacy late in the disease process was associated with the extent to which IDs and TDs induced Th2 responses and Th2 spreading and was unrelated to the affinity of the peptide for I-Ag7, the expression pattern of the Ag (abundant/rare, secreted/intracellular, and β cell specific/nonspecific), or whether it was an early or late target of spontaneous autoimmunity. Although further study is needed to confirm a relationship between a peptide’s immunogenicity and therapeutic efficacy late in the NOD mouse disease process, our results demonstrate that treatments with immunogenic β cell Ag IDs can be significantly more protective than TD-based treatments. The greater therapeutic efficacy of IDs is likely to involve the larger magnitude responses they induce, leading to greater education of APCs, Th2 spreading, bystander suppression, and exhaustion of naive TD-reactive T cell precursors. These findings run counter to the traditional concept that Ag-based therapies should be based on autoantigens or their TDs. Although the capability of the ID-based treatments to prevent T1D was only modest in our study, it should be noted that the mice received only one boost immunization and that such treatments are less invasive than other therapeutic strategies that broadly affect T cell function. Moreover, as autoimmunity can reoccur even after more invasive therapeutic approaches, combined treatments that include an Ag-based therapy may more effectively establish long-term functional tolerance. Taken together, our observations may aid in the rational design of Ag-based therapies that are better geared toward slowing the progression of late stage organ-specific autoimmune diseases.

We thank Drs. Eli Sercarz, Kamal Moudgil, and Mark Atkinson for their comments on the manuscript. We also thank Cindy Chau, Dan Zekzer, and Celeste Chong-Cerrillo for their technical assistance.

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.

1

This work was supported by a National Institutes of Health grant (to D.L.K.).

3

Abbreviations used in this paper: TD, target determinant; T1D, type 1 diabetes; ID, ignored determinant; EAE, experimental allergic encephalomyelitis; GAD, glutamic acid decarboxylase; REC, reduced expression in cancer; IAPP, islet amyloid polypeptide; HEL, hen egg lysozyme; MSA, mouse serum albumin; SI, stimulation index; SFC, spot-forming colony.

1
Liblau, R. S., S. M. Singer, H. O. McDevitt.
1995
. Th1 and Th2 CD4+ T cells in the pathogenesis of organ-specific autoimmune diseases.
Immunol. Today
16
:
34
.
2
Weiner, H. L..
1997
. Oral tolerance: immune mechanisms and treatment of autoimmune diseases.
Immunol. Today
18
:
335
-343.
3
Groux, H., A. O’Garra, M. Bigler, M. Rouleau, S. Antonenko, J. E. de Vries, M. G. Roncarolo.
1997
. A CD4+ T cell subset inhibits antigen-specific T cell responses and prevents colitis.
Nature
389
:
737
-742.
4
Powrie, F., R. L. Coffman.
1993
. Cytokine regulation of T cell function: potential for therapeutic intervention.
Immunol. Today
14
:
270
-274.
5
Tisch, R., H. O. McDevitt.
1994
. Antigen-specific immunotherapy: is it a real possibility to combat T cell-mediated autoimmunity?.
Proc. Natl. Acad. Sci. USA
91
:
437
-438.
6
Bach, J. F., L. Chatenoud.
2001
. Tolerance to islet autoantigens in type 1 diabetes.
Annu. Rev. Immunol.
19
:
131
-161.
7
You, S., C. Chen, W. H. Lee, T. Brusko, M. Atkinson, C. P. Liu.
2004
. Presence of diabetes-inhibiting, glutamic acid decarboxylase-specific, IL-10-dependent, regulatory T cells in naive nonobese diabetic mice.
J. Immunol.
173
:
6777
-6785.
8
Tian, J., P. V. Lehmann, D. L. Kaufman.
1997
. Determinant spreading of T helper 2 (Th2) responses to pancreatic islet autoantigens.
J. Exp. Med.
186
:
2039
-2043.
9
Tian, J., A. P. Olcott, D. L. Kaufman.
2002
. Antigen-based immunotherapy drives the precocious development of autoimmunity.
J. Immunol.
169
:
6564
-6569.
10
Tian, J., M. Clare-Salzler, A. Herschenfeld, B. Middleton, D. Newman, R. Mueller, S. Arita, C. Evans, M. A. Atkinson, Y. Mullen, et al
1996
. Modulating autoimmune responses to GAD inhibits disease progression and prolongs islet graft survival in diabetes-prone mice.
Nat. Med.
2
:
1348
-1353.
11
Elias, D., I. R. Cohen.
1994
. Peptide therapy for diabetes in NOD mice.
Lancet
343
:
704
-706.
12
Elliott, J. F., H. Y. Qin, S. Bhatti, D. K. Smith, R. K. Singh, T. Dillon, J. Lauzon, B. Singh.
1994
. Immunization with the larger isoform of mouse glutamic acid decarboxylase (GAD67) prevents autoimmune diabetes in NOD mice.
Diabetes
43
:
1494
-1499.
13
Daniel, D., D. R. Wegmann.
1996
. Protection of nonobese diabetic mice from diabetes by intranasal or subcutaneous administration of insulin peptide B-(9-23).
Proc. Natl. Acad. Sci. USA
93
:
956
-960.
14
Muir, A., A. Peck, M. Clare-Salzler, Y. H. Song, J. Cornelius, R. Luchetta, J. Krischer, N. Maclaren.
1995
. Insulin immunization of nonobese diabetic mice induces a protective insulitis characterized by diminished intraislet interferon γ transcription.
J. Clin. Invest.
95
:
628
-634.
15
Tisch, R., B. Wang, D. V. Serreze.
1999
. Induction of glutamic acid decarboxylase 65-specific Th2 cells and suppression of autoimmune diabetes at late stages of disease is epitope dependent.
J. Immunol.
163
:
1178
-1187.
16
Tian, J., D. L. Kaufman.
1998
. Attenuation of inducible Th2 immunity with autoimmune disease progression.
J. Immunol.
161
:
5399
-5403.
17
Steinman, L..
1990
. The T lymphocyte in experimental allergic encephalomyelitis.
Annu. Rev. Immunol.
8
:
579
-621.
18
Lehmann, P. V., O. S. Targoni, T. G. Forsthuber.
1998
. Shifting T cell activation thresholds in autoimmunity and determinant spreading.
Immunol. Rev.
164
:
53
-61.
19
Ashton-Rickardt, P. G., S. Tonegawa.
1994
. A differential-avidity model for T cell selection.
Immunol. Today
15
:
362
-366.
20
Sercarz, E. E., P. V. Lehmann, A. Ametani, G. Benichou, A. Miller, K. Moudgil.
1993
. Dominance and crypticity of T cell antigenic determinants.
Annu. Rev. Immunol.
11
:
729
-766.
21
Lehmann, P. V., T. Forsthuber, A. Miller, E. E. Sercarz.
1992
. Spreading of T cell autoimmunity to cryptic determinants of an autoantigen.
Nature
358
:
155
-157.
22
Elias, D., T. Reshef, O. S. Birk, R. van der Zee, M. D. Walker, I. R. Cohen.
1991
. Vaccination against autoimmune mouse diabetes with a T cell epitope of the human 65-kDa heat shock protein.
Proc. Natl. Acad. Sci. USA
88
:
3088
-3091.
23
Kaufman, D. L., M. Clare-Salzler, J. Tian, T. Forsthuber, G. S. Ting, P. Robinson, M. A. Atkinson, E. E. Sercarz, A. J. Tobin, P. V. Lehmann.
1993
. Spontaneous loss of T cell tolerance to glutamic acid decarboxylase in murine insulin-dependent diabetes.
Nature
366
:
69
-72.
24
Atkinson, M. A., M. A. Bowman, L. Campbell, B. L. Darrow, D. L. Kaufman, N. K. Maclaren.
1994
. Cellular immunity to a determinant common to glutamate decarboxylase and coxsackie virus in insulin-dependent diabetes.
J. Clin. Invest.
94
:
2125
-2129.
25
Kaufman, D. L., R. Tisch, N. Sarvetnick, L. Chatenoud, L. C. Harrison, K. Haskins, A. Quinn, E. Sercarz, B. Singh, M. von Herrath, et al
2001
. Report from the First International NOD Mouse T Cell Workshop and the follow-up mini-workshop.
Diabetes
50
:
2459
-2463.
26
Reich, E. P., H. von Grafenstein, A. Barlow, K. E. Swenson, K. Williams, C. A. Janeway, Jr.
1994
. Self-peptides isolated from MHC glycoproteins of non-obese diabetic mice.
J. Immunol.
152
:
2279
-2288.
27
Harrison, L. C., M. C. Honeyman, S. Trembleau, S. Gregori, F. Gallazzi, P. Augstein, V. Brusic, J. Hammer, L. Adorini.
1997
. A peptide-binding motif for I-Ag7, the class II major histocompatibility complex (MHC) molecule of NOD and Biozzi AB/H mice.
J. Exp. Med.
185
:
1013
-1021.
28
Corper, A. L., T. Stratmann, V. Apostolopoulos, C. A. Scott, K. C. Garcia, A. S. Kang, I. A. Wilson, L. Teyton.
2000
. A structural framework for deciphering the link between I-Ag7 and autoimmune diabetes.
Science
288
:
505
-511.
29
Deng, H., R. Apple, M. Clare-Salzler, S. Trembleau, D. Mathis, L. Adorini, E. Sercarz.
1993
. Determinant capture as a possible mechanism of protection afforded by major histocompatibility complex class II molecules in autoimmune disease.
J. Exp. Med.
178
:
1675
-1680.
30
Forsthuber, T., H. C. Yip, P. V. Lehmann.
1996
. Induction of Th1 and Th2 immunity in neonatal mice.
Science
271
:
1728
-1730.
31
Neophytou, P. I., E. M. Muir, J. C. Hutton.
1996
. A subtractive cloning approach to the identification of mRNAs specifically expressed in pancreatic β cells.
Diabetes
45
:
127
-133.
32
Kaestner, K. H., C. S. Lee, L. M. Scearce, J. E. Brestelli, A. Arsenlis, P. P. Le, K. A. Lantz, J. Crabtree, A. Pizarro, J. Mazzarelli, et al
2003
. Transcriptional program of the endocrine pancreas in mice and humans.
Diabetes
52
:
1604
-1610.
33
van der Most, R. G., A. Sette, C. Oseroff, J. Alexander, K. Murali-Krishna, L. L. Lau, S. Southwood, J. Sidney, R. W. Chesnut, M. Matloubian, R. Ahmed.
1996
. Analysis of cytotoxic T cell responses to dominant and subdominant epitopes during acute and chronic lymphocytic choriomeningitis virus infection.
J. Immunol.
157
:
5543
-5554.
34
Chao, C. C., H. K. Sytwu, E. L. Chen, J. Toma, H. O. McDevitt.
1999
. The role of MHC class II molecules in susceptibility to type I diabetes: identification of peptide epitopes and characterization of the T cell repertoire. [Published erratum in 1999 Proc. Natl. Acad. Sci. USA 96: 12970].
Proc. Natl. Acad. Sci. USA
96
:
9299
-9304.
35
Tian, J., S. Gregori, L. Adorini, D. L. Kaufman.
2001
. The frequency of high avidity T cells determines the hierarchy of determinant spreading.
J. Immunol.
166
:
7144
-7150.
36
Tian, J., A. P. Olcott, L. R. Hanssen, D. Zekzer, B. Middleton, D. L. Kaufman.
1998
. Infectious Th1 and Th2 immunity in diabetes prone mice.
Immunol. Rev.
164
:
119
-127.
37
Tian, J., Y. Lu, L. Hanssen, H. Dang, D. L. Kaufman.
2003
. Memory and effector T cells modulate subsequently primed immune responses to unrelated antigens.
Cell. Immunol.
224
:
74
-85.
38
Yip, H. C., A. Y. Karulin, M. Tary-Lehmann, M. D. Hesse, H. Radeke, P. S. Heeger, R. P. Trezza, F. P. Heinzel, T. Forsthuber, P. V. Lehmann.
1999
. Adjuvant-guided type-1 and type-2 immunity: infectious/noninfectious dichotomy defines the class of response.
J. Immunol.
162
:
3942
-3949.
39
Elias, D., A. Meilin, V. Ablamunits, O. S. Birk, P. Carmi, S. Konen-Waisman, I. R. Cohen.
1997
. Hsp60 peptide therapy of NOD mouse diabetes induces a Th2 cytokine burst and down-regulates autoimmunity to various β cell antigens.
Diabetes
46
:
758
-764.
40
Billiau, A., H. Heremans, F. Vandekerckhove, R. Dijkmans, H. Sobis, E. Meulepas, H. Carton.
1988
. Enhancement of experimental allergic encephalomyelitis in mice by antibodies against IFN-γ.
J. Immunol.
140
:
1506
-1510.
41
Kumar, V., E. Sercarz.
1998
. Induction or protection from experimental autoimmune encephalomyelitis depends on the cytokine secretion profile of TCR peptide-specific regulatory CD4 T cells.
J. Immunol.
161
:
6585
-6591.
42
Bockova, J., D. Elias, I. R. Cohen.
1997
. Treatment of NOD diabetes with a novel peptide of the hsp60 molecule induces Th2-type antibodies.
J. Autoimmun.
10
:
323
-329.
43
Tian, J., P. V. Lehmann, D. L. Kaufman.
1994
. T cell cross-reactivity between coxsackievirus and glutamate decarboxylase is associated with a murine diabetes susceptibility allele.
J. Exp. Med.
180
:
1979
-1984.
44
Chao, C. C., H. O. McDevitt.
1997
. Identification of immunogenic epitopes of GAD 65 presented by Ag7 in non-obese diabetic mice.
Immunogenetics
46
:
29
-34.
45
Chatenoud, L., B. Salomon, J. A. Bluestone.
2001
. Suppressor T cells: they’re back and critical for regulation of autoimmunity!.
Immunol. Rev.
182
:
149
-163.
46
Shevach, E. M..
2001
. Certified professionals: CD4+CD25+ suppressor T cells.
J. Exp. Med.
193
:
F41
-F46.
47
Weiner, H. L., A. Friedman, A. Miller, S. J. Khoury, A. al-Sabbagh, L. Santos, M. Sayegh, R. B. Nussenblatt, D. E. Trentham, D. A. Hafler.
1994
. Oral tolerance: immunologic mechanisms and treatment of animal and human organ-specific autoimmune diseases by oral administration of autoantigens.
Annu. Rev. Immunol.
12
:
809
-837.
48
Zekzer, D., F. S. Wong, L. Wen, M. Altieri, T. Gurlo, H. von Grafenstein, R. S. Sherwin.
1997
. Inhibition of diabetes by an insulin-reactive CD4 T-cell clone in the nonobese diabetic mouse.
Diabetes
46
:
1124
-1132.
49
Pakala, S. V., M. O. Kurrer, J. D. Katz.
1997
. T helper 2 (Th2) T cells induce acute pancreatitis and diabetes in immune-compromised nonobese diabetic (NOD) mice.
J. Exp. Med.
186
:
299
-306.
50
Cai, Z., A. Brunmark, M. R. Jackson, D. Loh, P. A. Peterson, J. Sprent.
1996
. Transfected Drosophila cells as a probe for defining the minimal requirements for stimulating unprimed CD8+ T cells.
Proc. Natl. Acad. Sci. USA
93
:
14736
-14741.
51
Pihlgren, M., P. M. Dubois, M. Tomkowiak, T. Sjogren, J. Marvel.
1996
. Resting memory CD8+ T cells are hyperreactive to antigenic challenge in vitro.
J. Exp. Med.
184
:
2141
-2151.
52
Dubey, C., M. Croft, S. L. Swain.
1995
. Costimulatory requirements of naive CD4+ T cells: ICAM-1 or B7-1 can costimulate naive CD4 T cell activation but both are required for optimum response.
J. Immunol.
155
:
45
-57.
53
Bachmann, M. F., K. McKall-Faienza, R. Schmits, D. Bouchard, J. Beach, D. E. Speiser, T. W. Mak, P. S. Ohashi.
1997
. Distinct roles for LFA-1 and CD28 during activation of naive T cells: adhesion versus costimulation.
Immunity
7
:
549
-557.
54
Moudgil, K. D., E. E. Sercarz.
2000
. The self-directed T cell repertoire: its creation and activation.
Rev. Immunogenet.
2
:
26
-37.
55
Cobbold, S., H. Waldmann.
1998
. Infectious tolerance.
Curr. Opin. Immunol.
10
:
518
-524.
56
Alpan, O., E. Bachelder, E. Isil, H. Arnheiter, P. Matzinger.
2004
. “Educated” dendritic cells act as messengers from memory to naive T helper cells.
Nat. Immunol.
5
:
615
-622.
57
Tian, J., D. L. Kaufman.
1998
. Insulin selectively primes Th2 responses and induces regulatory tolerance to insulin in pre-diabetic mice.
Diabetologia
41
:
237
-240.
58
Quinn, A., B. McInerney, E. P. Reich, O. Kim, K. P. Jensen, E. E. Sercarz.
2001
. Regulatory and effector cd4 T cells in nonobese diabetic mice recognize overlapping determinants on glutamic acid decarboxylase and use distinct vβ genes.
J. Immunol.
166
:
2982
-2991.