| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Genes1
*
Division of Immune Regulation, La Jolla Institute for Allergy and Immunology, San Diego, CA 92121,
Anergen, Redwood City, CA 94063; and
Department of Microbiology, Immunology, and Molecular Genetics, University of California, Los Angeles, CA 90095
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene use. CD4+ T cells arise
spontaneously in young NOD mice to an apparently dominant determinant
found within the GAD65 peptide 530–543 (p530); however, T cells to the
overlapping determinant 524–538 (p524) dominate the response only
after immunization with GAD65(524–543). All p530-responsive T cells
used the V4 gene, whereas the
V12 gene is preferentially used to
encode the TCR of p524-responsive T cell populations. T cell clones and
hybridomas from both of these T cell groups were responsive to APC
pulsed with GAD65(524–543) or whole rGAD65. p524-reactive cells
appeared to be regulatory upon adoptive transfer into young NOD mice
and could inhibit insulin-dependent diabetes mellitus development,
although they were unable to produce IL-4, IL-10, or TGF upon
antigenic challenge. Furthermore, we found that i.p. injection with
p524/IFA was very effective in providing protection from
cyclophosphamide-induced insulin-dependent diabetes mellitus. These
data demonstrate that the regulatory T cells elicited by immunizing
with GAD65(524–543) are unique and distinct from those that arise from
spontaneous endogenous priming, and that T cells to this limited region
of GAD65 may be either regulatory or pathogenic. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells of the pancreas. It has previously been
shown that glutamic acid decarboxylase (GAD65) is the first of several
cell Ags to be recognized by spleen cells from naive, prediabetic,
NOD mice (1, 2). The proliferative T cell response is
initially confined to the carboxyl-terminal region of GAD65 (residues
509–543) followed by responsiveness to additional GAD65 determinants
(1). These autoimmune responses to GAD65 develop
concurrently with the onset of lymphocytic infiltration into the islets
(insulitis) of NOD mice (1), and schemes designed to
tolerize or deviate the anti-GAD65 response in young mice seem to
provide some measure of protection from diabetes (1, 2, 3).
Similarly, humoral and cellular responses to GAD65 have also been
detected in the peripheral blood of recent-onset IDDM patients
(4). These findings clearly indicate that immune responses
to GAD65 are pivotal in the development of IDDM.
In the effort to discover which T cells become naturally involved in
bringing about the IDDM disease state, many different features of
diabetes in the NOD mouse have had to be considered. Which T cells are
relevant to disease, those found by their spontaneous appearance in the
young NOD mouse or those able to be evoked by immunization with Ag? Do
effector cells differentiate into regulatory cells by a type of
Th1
Th2 deviation, or are there numerous functionally distinct T
cells? Attempts have been made by many groups to approach answers to
these questions and others. Suffice it to say that none of the above
questions has a definitive answer at this point. However, there are a
few issues that can serve as a foundation for further progress. First,
in NOD mice, the proliferative activities of CD4+
T cells directed against GAD65 are first detectable at 
3 wk of age
(1, 2). Second, two almost nonoverlapping families of
GAD65-specific CD4 T cells have been reported in NOD mice, those
appearing spontaneously (1) and those induced by
immunization with GAD65 or GAD65 peptides (5, 6, 7). Lastly,
the prevention of autoimmune diabetes by cell-specific suppression
of GAD expression in two lines of antisense GAD transgenic NOD mice
suggests that GAD65 is required, as a self Ag, for spontaneous diabetes
to occur in NOD mice (8).
Upon immunization, a number of self Ags, and peptides derived from
them, are fully capable of inducing severe autoimmune disease in mice.
However, despite its prominence as a target during disease progression,
it is only with rare exception that autoimmune diabetes has been
associated with the delivery of GAD65 or GAD65 peptides into NOD mice
(9, 10). In this report, we address the distinction
between the GAD65(524–543)-specific CD4+ T cells
which spontaneously arise in the NOD mouse (1) and those
that are activated as a result of immunization with this peptide
(11). We demonstrate that there are actually two
functional sets of such T cells, each directed against one of two
overlapping registers within the 20-mer peptide GAD65(524–543). One
set arises spontaneously and is specific for a GAD65 determinant within
the 530–543 sequence (p530) while the other is induced only after
immunization with peptides GAD65(524–543) or (524–538) (p524). We
show that p524- and p530-reactive T cells use distinct TCR V
families, and that the 524–538 moiety induces T cells that are capable
of playing a regulatory role in IDDM.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Female NOD mice were purchased from Taconic Farms (Germantown, NY) and bred at the La Jolla Institute for Allergy and Immunology (San Diego, CA). NOD/Caj mice were kindly provided by Dr. Charles A. Janeway, Jr. (Yale University School of Medicine, New Haven, CT) and were maintained in a breeding colony at Anergen (Redwood City, CA). BALB/c and NOR1/Lt mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The mice were age- and sex-matched in all experiments.
Peptides and Ags
GAD65 peptides were synthesized at the University of California (Los Angeles, CA) Peptide Synthesis Laboratory on an Advanced Chemtech 395 synthesizer using f-moc chemistry and purified using HPLC. Peptide purity was determined by capillary electrophoresis and the composition was verified by mass spectrometry. OVA was purchased from Sigma (St. Louis, MO).
Recombinant mGAD65 was produced from BL21(DE3)pLysS E. coli
containing the bacterial expression vector pET3a
10-mGAD65. The
bacteria and vector were kindly provided by Agnes Lehuen and Nicolas
Glaichenhaus. The vector contains the full-length mGAD65 cDNA preceded
by an N-terminal hemagglutinin and a 6x histidine tag. The recombinant
protein was purified on Ni-NTA beads as described in the product
literature (Qiagen, Valencia, CA). Briefly, 1 L of Luria-Bertani media
is inoculated from an overnight culture of the above bacteria in the
presence of 100 µg/ml ampicillin and 10 µg/ml
chloramphenicol/ethanol and cultured for 4 h at 37°C/225 rpm in a
bacterial shaker until the OD590 equaled 0.4. The
expression of the recombinant protein was then induced using 5 mM
isopropyl -D-thiogalactoside, which was
followed by culture for an additional 6 h. The bacteria were then
pelleted via centrifugation and bound onto Ni-NTA columns in batches in
a solution containing 6 M guanidine. Contaminating proteins were then
removed in a series of 8 M urea-containing buffers of decreasing pH
before the elution of GAD65 in 8 M urea (pH 4.5). The presence and
purity of GAD65 were confirmed by SDS-PAGE and Western blot analysis
using purified anti-GAD65 mAb (GAD65–6; American Type Culture
Collection, Manassas, VA) (12). Those fractions containing
purified GAD65 were then pooled and dialyzed 10 times against PBS to
remove the urea. The purified GAD65 was then lyophilized, and
resuspended to a concentration of 1 mg/ml.
T cell proliferation assay
Spontaneous and induced proliferative responses were determined as previously described, (1) and (11), respectively. For spontaneous responses, spleen cells were plated at 8 x 105 nucleated cells per well in 96-well flat-bottom plates, using HL-1 serum-free medium (Bio-Whittaker, Walkersville, MD). Peptides were added to a final concentration of 10–40 µg/ml for 5 days at 37°C in 7% CO2, and 1 µCi/well [3H]thymidine (International Chemical and Nuclear, Irvine, CA) was added for the last 16 h. The cells were harvested from microtiter plates using a Micro Cell Harvester (Skatron Instruments, Sterling, VA) and incorporation of label was measured by liquid scintillation counting in an LKB 1205 Betaplate counter (LKB Instruments, Gaithersburg, MD). The results were read as mean cpm of triplicate wells; the SD was <15% in all experiments.
For induced responses, 6- to 10-wk-old mice were immunized s.c. in the hind footpad with 20 µg peptide emulsified in CFA (Difco, Detroit, MI). Popliteal and inguinal lymph nodes and spleens were removed 9–11 days later to prepare single-cell suspensions. Lymph node cells and spleen cells were plated in 96-well microtiter plates at 5 x 105 and 8 x 105 cells/well, respectively, in serum-free medium (X-VIVO-10; Bio-Whittaker) supplemented with 2 x 10-5M 2-ME. The 10 µg/ml level of peptide was usually optimal for inducing proliferation and tuberculin-purified protein derivative (1:40) was used as a positive control for proliferation. [3H]-tritiated thymidine was added for the last 16 h of a 4-day culture.
T cell lines, clones, and hybridomas
T cell lines and clones were produced by the in vitro restimulation of in vivo-primed splenic or lymph node cells using irradiated syngeneic spleen cells plus the peptide (10 µg/ml) of interest. Three to 5 days later, the cells were further expanded in 10 U/ml rIL-2-containing culture medium (RPMI 1640 or Click’s medium supplemented with penicillin/streptomycin, 2 mM L-glutamine, sodium pyruvate, nonessential amino acids, 2-ME, and 10% FBS) and then screened for antigenic specificity 10–14 days later. T clones were isolated by limiting dilution in 96-well round-bottom plates containing 2 x 105 peptide-pulsed, irradiated, syngeneic spleen cells in IL-2-containing culture medium.
Ag-specific T cell hybridomas were created by fusing lymphocytes from
peptide-immunized mice with the 
TCR (-/-) BW5147 cell line
(American Type Culture Collection), as previously described
(11). Ag-responsive hybridomas were identified by IL-2
production using the HT-2 bioassay (11).
Cytokine ELISA
Supernatants collected from 24–72 h after antigenic stimulation
of T cell lines and clones were tested for the presence of IFN-
,
IL-4, IL-5, and IL-10 by ELISA and IL-2 was measured using the HT-2
bioassay as described previously (11). Recombinant murine
cytokines were used as standards.
Cyclophosphamide-induced IDDM
To accelerate and synchronize the onset of diabetes, NOD mice were treated with cyclophosphamide as previously described (13). Briefly, 10- to 13-wk-old mice were given a single i.p. injection with cyclophosphamide (Sigma), 200 mg/kg of body weight. The incidence of IDDM in mice was determined by daily urine analysis (Chemstrip µG; Boehringer Mannheim, Indianapolis, IN) for 3 wk, with those mice testing positive for glucosuria being confirmed by measurements of blood glucose using an Encore glucometer (Bayer, Elkhart, IN). Those mice with blood glucose >250 mg/dl were considered diabetic. Pancreatic insulitis was judged blindly from H&E stained sections of formalin-fixed tissue.
To attempt to block cyclophosphamide-induced diabetes, neonatal NOD mice were injected twice i.p. with 100 µg of peptide in IFA, 14 and 21 days after birth, and then given cyclophosphamide as adults.
Adoptive-transfer experiments
To assess the regulatory capacity of p524-reactive T cells, T cell clones and lines were restimulated with Ag-fed irradiated spleen cells for 3–5 days, and then expanded in IL-2-containing medium. Before transfer, viable T cells were collected by Histopaque separation (Sigma), washed twice, and then resuspended in PBS. Two-week-old NOD mice were injected i.p. with 1 x 107 T cells in 0.2 ml of saline or saline alone. The mice were monitored weekly for glucosuria.
Flow cytometry analysis
For analysis of TCR V expression and cell surface markers, T
cell lines, and clones were stained with FITC or PE-conjugated Abs
(PharMingen) specific for CD3, CD4, TcR, and TCR V-specific
chains 2, 3, 4, 5.1, 5.2, 6, 7, 8.1, 8.2, 8.3, 9, 10, 12, 13, 14, and
17a. Stained cells were analyzed on a FACScan flow cytometer (Becton
Dickinson, Mountain View, CA) using CellQuest software (Becton
Dickinson).
Statistical analysis
The Statview software package (Abacus Concepts, Berkeley, CA) was used for statistical analysis. The proliferative responses to GAD65 peptides and insulitis scores were compared with controls by Student’s t test, and the incidence of IDDM in Ag-pretreated groups was compared with control groups by the Fisher’s Exact test. Values of p < 0.05 were considered significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
usagePreviously, we characterized the T cell responses in NOD mice immunized with the 20-mer peptide GAD65(524–543) (11). Six- to 10-wk-old NOD mice were immunized in the footpad with 20 µg of peptide GAD65(524–543) emulsified in CFA and then tested 9–10 days later for proliferative responses to overlapping GAD65 peptides. Although GAD65 peptides 524–538 (p524), 527–541 (p527), and 530–543 (p530) were all able to induce recall responses in GAD65(524–543)-immunized mice, p524 contained the dominant determinant, whereas p530 was the least effective at inducing proliferation in such mice (11). Previously, we showed that p524-specific T cell hybridomas produced from the spleen cells of mice immunized with GAD65(524–543) were not responsive to p530 in the context of I-Ag7 (11), thus indicating that p530 and p524 were distinct T cell determinants.
P530-reactive T cells hybridomas use the V4 TCR chain
To characterize the response to p530 in the NOD model, a panel of
T cell hybridomas was produced from the spleen cells of prediabetic
mice immunized s.c. with p530, as described in Materials and
Methods. It should be noted that the p530-primed spleen cells were
restimulated in vitro with peptide GAD65(524–543) before fusion, and
were selected based on reactivity with this same peptide. Although all
of these hybridomas were reactive with p530, none was responsive to
p524 (Table I
). Because the determinant
recognized by these hybrids could be processed and presented from the
GAD65 molecule by syngeneic APC (Table I), these results confirmed that
there were two overlapping I-Ag7-restricted
determinants within the diabetes-associated GAD65(524–543) sequence
and that NOD APC were able to bind the GAD65(524–543) peptide with
reasonable affinity in either of the two registers. When TCR use by
p530-reactive T cell hybridomas was characterized, we found that the
V4 TCR chain was almost exclusively selected in this response (Table I). Although a larger set of p530-reactive T cell hybridomas was
produced, Table I only contains the hybridomas that we have
conclusively determined are unique and are not sister clones, based on
their response profile to altered peptide ligands and RT-PCR analysis
of their V and V TCR gene expression.
|
12 TCR chain
In contrast to the pattern shown by p530-reactive T cells, our
previously described hybridomas, which were recovered from
GAD65(524–543)-immunized NOD mice, preferentially expressed V12
TCRs (Table II). Three of four unique
hybrids tested used the V12 TCR chain. In addition, two T cell
clones, GAD35Z.a and GAD35Z.b, produced from a different set of
GAD65(524–543)-immunized NOD mice, also used the V12 TCR chain
(Table II). RT-PCR analysis of the TCR--chain genes (data not shown)
showed that the hybrids and clones were each unique and, therefore, the
preponderance of V12+ T cells was not owing to
the analysis of sister clones. Moreover, when an additional panel of
p524-reactive T clones and hybridomas were independently produced
(prepared at Anergen) from GAD65(524–543)-immunized mice, their
specificity and V12 TCR usage were identical with those described
above (Table II). OVA-specific T cell clones recovered from
OVA-immunized NOD mice did not show a similar preference for V4 or
V12 TCR usage (data not shown). These results show that each of the
two I-Ag7-restricted determinants present within
the GAD65(524–543) sequence recruits a unique set of T cells that
characteristically express a particular TCR V-chain. Furthermore,
both determinants can be processed from the GAD65 molecule to stimulate
their respective T cells (Tables I
and II).
|
To determine which of the T cell populations was responsible for
the spontaneous proliferation to GAD65(524–543) in young prediabetic
NOD mice, spleen cells from female mice were cultured with the
overlapping GAD65 peptide 521–535 (p521), p524, p527, p530, and
167–181 (p167) (all 15-mers). We found that peptides p530 and p527
were able to induce proliferative responses (Fig. 1); however no response was seen with
peptide p524 or any of the other GAD65 peptides tested (Fig. 1).
Although the results from this experiment were repeated several times,
spontaneous T cell responses to GAD peptides were sometimes difficult
to demonstrate on a routine basis. We have now discovered that a more
consistent response can be shown when short-term lines are produced
from the spleen cells of untreated NOD mice, using two to three cycles
of in vitro stimulation with irradiated syngeneic APCs pulsed with
peptide GAD65(524–543). The response pattern of such
GAD65(524–543)-reactive T cell lines and clones (Fig. 2) was dominated by the same p530
specificity profile as that shown in the initial spontaneous
proliferative studies (Fig. 1) and in the T cell hybridomas produced
from p530-immunized mice (Table I). Because p530-specific T cells are
among the first to arise spontaneously in the NOD mouse
(1), in this context they are dominant. Contrastingly, the
p524-specific response was undetectable in naive NOD mice (p524 was
tested over a large concentration range, 0.1–200 µg/ml, for its
ability to induce spontaneous T cell proliferative responses, data not
shown).
|
|
4+ T cells increased in the bulk population
(Fig. 3A) while T cells
expressing other V family members (i.e., V12 and V8) were
reduced in frequency (Fig. 3A) or were absent (data not
shown). As would be expected, p530-reactive T cell clones produced from
naive NOD mice showed a similar lack of response to p524 and were also
V4+ T cells (data not shown). The selection of
this particular TCR V-chain was not due to an in vitro artifact
because identical culture conditions using lymphocytes recovered from
GAD65(524–543) or p524-immunized NOD mice led to a bias toward the
expression of V12 TCR chains (Table II and Fig. 3B,
respectively). Fig. 3B shows that after 2 rounds of in vitro
restimulation with GAD65(524–543), a T cell line produced from the
lymph node cells of p524-immunized mice (524B) is composed primarily of
V12+ T cells (44%).
V4+ T cells were not significantly represented
in this line (Fig. 3B). These lines of evidence point
clearly to the existence of two distinct repertoires of T cells and the
need for an extensive characterization of each.
|
Pretreatment with GAD65(524–543) has been shown to protect NOD
mice from spontaneous (1) and cyclophosphamide-induced
IDDM (14), presumably by rendering the effector T cells
tolerant. In this study, we show that while IDDM was reduced in NOD
mice neonatally treated with GAD65 peptides p524, p530, or 524–543,
compared with PBS/IFA-treated animals, p524 was the most efficacious
and significantly better at providing protection from disease (Table III). Only 25% of the p524-treated mice
became diabetic, compared with 88–100% IDDM in the controls treated
with IFA or peptide 167–181, respectively. Seven of 12 mice became
diabetic after treatment with peptide p530 (Table III) and the 50%
reduction in disease incidence in the GAD65(524–543)-treated group
(Table III) was similar to that seen in a previous report
(14). Despite their strong immunogenicity in NOD mice,
GAD65 peptides 206–220 (5) and 167–181 (our unpublished
data) afforded little to no amelioration from cyclophosphamide-induced
IDDM (Table III), thus demonstrating that immunogenicity was not
sufficient to provide protection from the accelerated form of
IDDM.
|
A) while mice pretreated with
p530-IFA responded strongly to peptides p530 and GAD65(524–543) (Fig. 4B). Evidently, the i.p. injections with peptide did not
eliminate the proliferative responses to cognate peptides but rather
seemed to prime responses, particularly in the case of GAD65(524–543)
because the 524–538 response is not seen in unimmunized NOD mice (Fig. 1). Thus, the protection afforded by pretreatment with p524 can likely
be attributed to the priming of a p524-reactive regulatory T cell
rather than to the deletion of a pathogenic T cell population.
|
) these p524-reactive T cells may emanate from determinant
spreading and endogenous priming to this subdominant determinant. Protection from IDDM by adoptive transfer of p524-specific T cells
In previous reports in which GAD65 peptides were delivered to NOD
mice in IFA, the most significant change in the recall response was
observed in the Th2 arm, which increased, while the Th1 arm of the
response remained (3). In such a situation it is difficult
to determine whether protection is mediated by the enhanced Th2
response, the loss of pathogenic Th1 clones, the arisal of a new
regulatory repertoire (15), or a combination of these
events. To directly determine whether p524-reactive Th cells could
regulate IDDM, NOD-derived, peptide p524-reactive,
V12+ T cell clones and lines were produced
from NOD mice immunized with GAD65(524–543). The
CD4+ T cell clone GAD35Z.a (Table III), which
proliferates (Fig. 5A) and
secretes IFN- and IL-5, but not IL-4 or IL-10, (Fig. 5B,
Table IV) in the presence of syngeneic
APC and rGAD65 or GAD65(524–543), was adoptively transferred into
2-wk-old NOD mice and the course of IDDM was followed (male mice were
excluded from the experiment at the time of weaning). NOD recipients of
T cell clone GAD35Z.a were not only protected from IDDM, but at 20–24
wk of age, less than 4% of the islets showed any sign of inflammation.
More striking, in the mice receiving GAD35Z.a T clones, less than 1%
of the islets showed the invasive/destructive form of insulitis (Fig. 5C, Table V), while over 50%
of the islets in the saline control mice had a massive infiltration of
lymphocytes (Fig. 5D) (Table V). The littermate control mice
all succumbed to IDDM by 20 wk of age.
|
|
|
). Line GAD53Y
also secreted IL-5 and IFN- upon antigenic challenge but was unable
to produce IL-4 and IL-10 (Table IV). IL-4, IL-10, and TGF
expression were undetectable by RT-PCR analysis of Ag stimulated
GAD35Z.a and GAD35Y T cells (data not shown) although each produced
IL-2, as demonstrated in the HT-2 bioassay (data not shown). Therefore,
the p524-reactive T cells were able to arrest the effector mechanism(s)
set in play by crucial diabetogenic T cells, or alternatively, induced
a regulator that was reactive with the TCR associated with p524
responsiveness.
To further determine whether regulation of IDDM was a characteristic
associated with the p524-specific response in these mice we performed
additional adoptive transfer experiments using a third
V12+, p524-reactive T cell line. The T cell
line GAD35F was raised from lymph node cells of
GAD65(524–543)-immunized mice and was only restimulated three times in
vitro with the homologous peptide (as with the other p524-reactive T
cell lines, GAD35F produced IL-2, IL-5 and IFN- upon antigenic
challenge but no IL-4 or IL-10 (data not shown)). Pancreata from
recipient animals were harvested at 12 wk of age to compare the
severity of insulitis. Similar to previous experiments, at 12 wk of age
GAD35Z.a T cells significantly protected recipients from insulitis
(Table V). In mice receiving line GAD35F, the inflammation in the
pancreas was greatly reduced, compared with controls, such that only
4% of the islets showed any infiltration and none exhibited the
invasive form (Table V). To determine whether the regulation of IDDM
was induced by the adoptive T cells themselves, some recipient animals
were given GAD35F cells that had been irradiated (3000 rad) just before
transfer. The protection provided by GAD35F cells was lost when they
were irradiated (Table V), suggesting that the GAD35F T cells needed to
be completely functional in order for the regulation to work. In
contrast to those mice receiving p524-reactive T cells, mice receiving
OVA-1 T cells were similar to saline-treated mice with regard to their
insulitis profile (Table V). The OVA-1 T cells produced IL-2, IL-5, and
IFN- upon antigenic challenge (data not shown). These combined data
show that three independently derived populations of p524-reactive T
cells, with similar cytokine profiles, provided protective regulation
of IDDM in NOD mice, and suggest that the protection does not require
an additional effector cell (e.g., anti-idiotypic T cells) to be
generated in the recipient.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
For several years, there has been a concerted effort directed at
discovering the specificity of the CD4+ and
CD8+ T cells that influence the course of IDDM in
humans and NOD mice. An understanding of how GAD65-specific T cells can
be engaged and regulated is urgently needed in the light of efforts to
use preemptive GAD therapy in humans. Much attention has been devoted
to the various members of the GAD-specific CD4+ T
cell populations. GAD65 immunization has been reported to consistently
induce 206–220- and 221–235-specific CD4+ T
cell responses (5, 6) while a distinct set of
CD4+ T cells, specific for GAD65(509–528),
524–543, and 247–267 were shown to arise in naive NOD mice
(1). T cells of additional GAD65 specificities also may
become involved in this response (7). These findings
suggest that T cell repertoires raised after immunization may greatly
differ from those that arise spontaneously to endogenous priming. In
this study, we have shown that two unique CD4+ T
cell repertoires are recruited by the diabetes-associated region within
GAD65(524–543). One repertoire is primed spontaneously in vivo
(against GAD(530–543) (p530)) and is associated with cell
autoimmunity, the other is regulatory and inducible after immunization
with GAD65(524–543) or GAD(524–543) (p524). Although GAD-reactive T
cells may possess diabetogenic activity (10) and GAD
expression appears to be required for cell damage to occur in the
NOD mouse model (8), GAD65 or certain of its peptides also
have the ability to prevent the spontaneous development of IDDM in NOD
mice when they are administered in a specific manner (i.v., mucosally,
or i.p.) (1, 2, 9, 16). Therefore, the two different
effector functions of cellular destruction and immune regulation appear
to be dissociable within GAD65-specific T cell responses, even to a
circumscribed area of the molecule.
Spontaneous vs induced GAD65-specific populations
We have pursued an intensive investigation of the peptide
GAD65(524–543). The experiments reported in this study show that the
early spontaneous appearance of a proliferative response to the
dominant determinant GAD65(524–543) can now be further defined to
involve the 530–543 moiety of the larger peptide (although the core
appears to be contained within the 530–541 sequence). In some
instances, spontaneous responses were also seen to p527 (Fig. 1). Most
p530-specific T cells also recognized p527 but many p527-specific T
cells (from p527 immunized mice) do not respond to p530 (data not
shown). In addition, p527 was always less effective than p530 at
stimulating responses in naive mice, or p530-specific T cell lines and
hybridomas. Therefore, because the spontaneous response to p527 is
likely mediated by p530-specific T cells, spontaneous responses to
GAD65(524–543) were considered to be synonymous to p530-specific
responses.
The fact that this spontaneous proliferative response is confined to
p530, and does not include p524-reactive T cells, provides evidence for
its dominance with respect to processing and presentation from the
native GAD65 molecule. In contrast, within the scope of immunization
with peptide GAD65(524–543), the 524–538 moiety appears to be
dominant. Therefore, it is evident that within the larger 20-mer
peptide, GAD65(524–543), two functionally distinct dominant
determinants exist, apparently with completely opposite roles. The
dominantly arising spontaneous response to 530–543 is characterized by
V4+ T cells and is indicative of disease
progression (1, 17, 18); meanwhile,
V12+ T cells clones directed toward 524–538
can actually protect against diabetes. Our findings demonstrate that
there is a discrepancy between the dominant determinants that are
selected during lymphocyte activation with autonomously provided
peptide:MHC ligands and those determinants that are selected following
immunization with homologous Ag, even within a 20-mer peptide
It is noteworthy that the preferential selection of the distinct
GAD65(524–543)-specific repertoires, V4+ vs
V12+ T cells, is controlled by an endogenous
mechanism(s) during in vivo priming because GAD65(524–543) was always
used as the Ag during in vitro stimulation and in the initial screening
of T cells for antigenic specificity. Whether this difference in
determinant selection is operating through particular APCs is difficult
to determine because it is still unclear which APC are responsible for
the spontaneous priming of islet cell Ags. Although it is likely that
dendritic cells are a major contributor to T cell priming after
immunization, B cells and macrophages are necessary for spontaneous
autoimmunity to GAD65 and the development of IDDM in NOD mice
(19, 20, 21). Each of the three APC types can be found in the
inflamed pancreas of prediabetic NOD mice. Preferential targeting of Ag
to particular APC populations after immunization may explain why p530
induces a proliferative response in naive NOD mice (1),
which is curiously absent in the responses of GAD65-immunized mice
(5, 6). Interestingly, we found that the determinants p524
and p530 could both be processed from the GAD65 protein by NOD spleen
cells, although p530 was always more efficiently processed (only the
p530-specific clones can be detected in naive NOD mice and
p530-specific hybridomas are more sensitive to activation when rGAD65
is provided to APC in vitro). The data presented in this study, and
those of additional studies (K. Jensen, N. Che, B. Pederson, M.
Sadauski, P. Van Endert, A. Lehuen, S. Schoenberger, A. Quinn, and E.
E. Sercarz, manuscript in preparation), are in support of the view that
while the p530 determinant is dominant, and thus easily processed from
GAD65, p524 represents a subdominant determinant that is poorly
processed from intrinsic or exogenously provided GAD65. It is possible
that p524-specific T cells produced after immunization with cognate
peptide, have high affinity Ag receptors that allow them to detect the
suboptimal expression of p524:MHC complexes on GAD65-pulsed APC.
Furthermore, the small response to p524 seen in p530/IFA treated mice
(Fig. 4B) may result from the in vivo inactivation of high
affinity p530-reactive T cells, which consequently allows the
revelation of a less affine spontaneous p524-specific repertoire.
Consistent with this view, it has been shown that previously undetected
responses to subdominant determinants can become uncovered when
responses to dominant determinants, on the same molecule, are
interrupted (22, 23).
I-Ag7 MHC-binding motifs and GAD65(524–543)
It is of interest that the two overlapping peptides, p524 and
p530, each fit the I-Ag7 motif described by
Harrison et al. (Ref. 24 ; Fig. 6). In p524, the MHC anchor residues
could be I533 and R536, to fit I-Ag7 pockets p6
and p9, respectively. Likewise, for p530, the putative p6 residue could
be M537, while Y540 would fit pocket 9. The precise boundaries of
neither determinant is completely defined as yet, although the minimal
determinant required by the majority of V12+ T
cells is 526–538 (Ref. 11 and data not shown). According
to the above predictions about anchor residues in each determinant,
peptide 527–541 might very well include the amino acids necessary to
occupy both registers for stimulating each of the T cell sets.
|
4 in p530-specific clones and
BDC2.5
Strikingly, all of the p530-specific hybridomas use V4 chains
in their Ag receptors. This preferential V usage was also observed
in T cell lines and clones derived from unimmunized prediabetic NOD
mice. The focus on V4 is of interest given the fact that the
oft-studied diabetogenic T cell clone BDC-2.5 also expresses the V4
TCR chain (25). The BDC-2.5 clone has been shown to be
responsive to pancreatic islet cells and possesses diabetogenic
activity. The deconvolution of BDC-2.5 by Judkowski et al.
(29) implicates GAD65 as the autoantigen recognized by
this clone and brings the discussion closer to tying a knot connecting
spontaneous GAD65(524–543) responses with the disease process (a
combinatorial peptide library study deconvolutes the specificity of
BDC2.5 to GAD65 peptide 526–541; Judkowski et al. (29)).
Accordingly, the link has been forged between the earliest appearing
spontaneous T cell clones in the NOD mouse model and a TCR-Tg mouse
model of type I diabetes.
Regulation by p524-specific V12+ clones
The mechanisms by which the p524-reactive T cells control IDDM are
not yet clear. Numerous T cells have been reported to have regulatory
functions that can modulate autoimmune disease. These include T cells
that are characterized by particular cell surface markers (e.g., CD45B,
CD25, etc.), as well as those that display apparently unique cytokine
secretion patterns (e.g., Th2, Th3, Tr1). The regulatory cells
described in this study are CD4+ TCR
V12+ T cells that are potent producers of
IL-2, IL-5, and IFN-. IL-4, IL-10, and TGF have been described as
anti-inflammatory cytokines that are antagonistic to Th1 cells;
however, because none of the p524-reactive T cell lines were able to
produce detectable levels of IL-4, IL-10, or TGF, they cannot be
categorized as Tr1 (26) or Th3 type clones
(27). A similar protective role has not been previously
ascribed directly to IL-5 in autoimmune diabetes, but rather it has
been thought to be indicative of specific tolerance induction and
immune deviation (3). The idea that IFN--producing
cells can play a protective role in an inflammatory disease might seem
counterintuitive, although it was recently shown that a Th1 phenotype
was required in generating TCR-specific regulatory T cells in the EAE
model (28). Although the roles of IL-5, IFN-, and the
V12 TCR chain in the regulatory function of p524-reactive T cells
are currently being investigated, the transfer of an OVA-specific
V8+ clone or a p530-specific
V4+ T cell clone (data not shown) did not
alter the course of diabetes in recipients.
In summary, this work demonstrates that the response to an apparently unique determinant in a self-molecule instead represents engagement of at least two functionally distinct T cell repertoires; one distinct determinant spontaneously and dominantly activates a putatively autoaggressive T cell response early in the life of the NOD mouse, while the other overlapping determinant evokes a regulatory response that may down-regulate the aggressive response and interfere with IDDM. Interestingly, because the dominant p530 response can be interrupted by the induction of p524-specific T cells, the response to GAD65 may depend on the fine details of Ag processing in influencing determinant choice, and thereby disease outcome. In any case, it may be possible to actively induce Ag-specific regulatory T cells that recognize determinants distinct from those that arise during the natural course of autoimmune disease. Their recruitment could avoid the inadvertant activation of pathogenic clones, a potential problem associated with therapies designed to deviate or anergize disease-inducing T cells.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Eli E. Sercarz, Division of Immune Regulation, La Jolla Institute for Allergy and Immunology, 10355 Science Center Drive, San Diego, CA 92121.
3 Abbreviations used in this paper: IDDM, insulin-dependent diabetes mellitus; NOD, nonobese diabetic; GAD65, glutamic acid decarboxylase
Received for publication June 16, 2000. Accepted for publication December 13, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells. Science 284:1183.-cell antigens: [Published erratum appears in 1998 Diabetes 47:303.]. Diabetes 46:1975.[Abstract]
cells of nonobese diabetic mice. J. Immunol. 163:743.This article has been cited by other articles:
|
|
 |
|
  A. R. Burton, E. Vincent, P. Y. Arnold, G. P. Lennon, M. Smeltzer, C.-S. Li, K. Haskins, J. Hutton, R. M. Tisch, E. E. Sercarz, et al. On the Pathogenicity of Autoantigen-Specific T-Cell Receptors Diabetes, May 1, 2008; 57(5): 1321 - 1330. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. Martinic, A. E. Juedes, D. Bresson, D. Homann, K. Skak, C. Huber, E. Ling, M. Ejrnaes, T. Wolfe, L. Togher, et al. Minimal Impact of a De Novo-Expressed {beta}-Cell Autoantigen on Spontaneous Diabetes Development in NOD Mice Diabetes, April 1, 2007; 56(4): 1059 - 1068. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Quinn, M. McInerney, D. Huffman, B. McInerney, S. Mayo, K. Haskins, and E. Sercarz T cells to a dominant epitope of GAD65 express a public CDR3 motif Int. Immunol., June 1, 2006; 18(6): 967 - 979. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Fallarino, U. Grohmann, S. You, B. C. McGrath, D. R. Cavener, C. Vacca, C. Orabona, R. Bianchi, M. L. Belladonna, C. Volpi, et al. The Combined Effects of Tryptophan Starvation and Tryptophan Catabolites Down-Regulate T Cell Receptor {zeta}-Chain and Induce a Regulatory Phenotype in Naive T Cells. J. Immunol., June 1, 2006; 176(11): 6752 - 6761. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. D. Dai, K. P. Jensen, A. Lehuen, E. L. Masteller, J. A. Bluestone, D. B. Wilson, and E. E. Sercarz A Peptide of Glutamic Acid Decarboxylase 65 Can Recruit and Expand a Diabetogenic T Cell Clone, BDC2.5, in the Pancreas J. Immunol., September 15, 2005; 175(6): 3621 - 3627. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. P. Olcott, J. Tian, V. Walker, H. Dang, B. Middleton, L. Adorini, L. Washburn, and D. L. Kaufman Antigen-Based Therapies Using Ignored Determinants of {beta} Cell Antigens Can More Effectively Inhibit Late-Stage Autoimmune Disease in Diabetes-Prone Mice J. Immunol., August 1, 2005; 175(3): 1991 - 1999. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Han, Y. Li, J. Wang, R. Wang, G. Chen, L. Song, R. Xu, M. Yu, X. Wu, J. Qian, et al. Active Tolerance Induction and Prevention of Autoimmune Diabetes by Immunogene Therapy Using Recombinant Adenoassociated Virus Expressing Glutamic Acid Decarboxylase 65 Peptide GAD500-585 J. Immunol., April 15, 2005; 174(8): 4516 - 4524. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. K. Gregg, R. Jain, S. J. Schoenleber, R. Div |
cpm (mean experimental cpm - medium control
cpm).
