| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
*
Torrey Pines Institute for Molecular Studies, San Diego, CA 92121; and
Department of Immunology, The Scripps Research Institute, La Jolla, CA 92037
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cells similar to type 1 diabetes
mellitis in humans. Insulitis also occurs in the BDC2.5 TCR transgenic
line of NOD mice that express the rearranged TCR 
- and -chain
genes of a diabetogenic NOD CD4 T cell clone. When activated with
syngeneic islet cells in culture, BDC2.5 T cells adoptively transfer
disease to NOD recipients, but the identity of the islet cell Ag
responsible for pathogenicity is not known. To characterize the
autoantigen(s) involved, BDC2.5 T cells were used to screen a
combinatorial peptide library arranged in a positional scanning format.
We identified more than 100 decapeptides that stimulate these T cells
at nanomolar concentrations; they are then capable of transferring
disease to NOD-scid mice. Surprisingly, some of the
peptides include sequences similar (8 of 10 residues) to those found
within the 528–539 fragment of glutamic acid decarboxylase 65.
Although this 12-mer glutamic acid decarboxylase 65 fragment is only
slightly stimulatory for BDC2.5 T cells (EC50 > 100
µM), a larger 16-mer fragment, 526–541, shows activity in the low
micromolar range (EC50 = 2.3 µM). Finally, T cells
from prediabetic NOD mice respond spontaneously to these peptide
analogs in culture; this finding validates them as being related to a
critical autoantigen involved in the etiology of spontaneous diabetes
and indicates that their further characterization is important for a
better understanding of underlying disease
mechanisms. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In both species, type 1 diabetes involves an occult lymphocytic
infiltration in the pancreatic islets (insulitis) followed by
destruction of insulin-producing cells mediated by T lymphocytes,
ultimately leading to the failure to metabolize glucose. T cell
involvement is indicated by the finding that spontaneous disease does
not occur in congenic athymic or neonatally thymectomized NOD mice
(3, 4, 5), and disease is diminished or prevented in NOD mice
pretreated with anti-T cell reagents (6, 7, 8, 9, 10, 11).
Both CD4 and CD8 subsets of T cells are thought to be involved in the
spontaneous development of insulitis and cell destruction in the
NOD mouse; however, under some circumstances T cells of either subset
alone, including cloned T cell lines, are capable of causing adoptive
insulitis and disease in asymptomatic young NOD mice and in T
cell-deficient NOD-scid recipients (12, 13, 14, 15, 16, 17, 18, 19, 20).
The CD4+ T cell clone, BDC2.5 is highly
diabetogenic in NOD mice (16). The best evidence for a
role for CD8 T cells derives from studies showing that T cells from
young, prediabetic NOD mice are capable of transferring disease to MHC
class I-positive NOD-scid mice, but not to
NOD-scid mice backcrossed to
2-microglobulin-deficient mice that lack MHC
class I expression (21). In addition, a murine
CD8+ T cell clone (G9C8), reactive to a peptide
fragment of insulin B-chain, has been described that is capable of
adoptive transfer of diabetes in the absence of CD4 T cells to
recipients expressing H-2Kd (22, 23).
Several candidate proteins have been considered as potential
autoantigens in the etiology of diabetes both in humans and the NOD
mouse. These include an islet granule protein (24),
insulin (25), the 64-kDa isoform of glutamic acid
decarboxylase (GAD) 65 (26), and heat shock protein
(27). However, much of the evidence for the involvement of
these putative autoantigens is indirect, based either on data showing
proliferative or cytokine responses to various peptide fragments of
these proteins by T cells or clones derived from prediabetic animals,
or on studies showing that various pretreatments of NOD animals with
these peptides/proteins reduces disease severity (25, 26, 27, 28, 29, 30, 31).
Recently, it was reported that selective suppression of GAD expression
in the cells of a transgenic line of NOD mice bearing antisense GAD
transgenes significantly reduced insulitis and blocked development of
disease (32). However, it appears to be only a rare
exception that direct administration of GAD65 or one of its fragments
results in diabetes in the mouse model (26, 31).
A transgenic strain of NOD mice expressing the TCR - and -chain
genes of the BDC2.5 T cell clone (33) has been a useful
model for studies of the immunobiology of a single T cell specificity
in the pathogenesis of diabetes. The original BDC2.5 T cell clone was
derived from a NOD mouse, is reactive to an unknown islet cell Ag in
association with H-2g7, and is highly
diabetogenic in young asymptomatic NOD mice (16). The
TCR-transgenic BDC2.5 mouse develops an aggressive form of diabetes in
some (33) but not all colonies (34), and
activated T cells from these donors transfer disease adoptively to
NOD-scid recipients (34).
In this study, we sought to identify a series of peptides specific for T cells from the BDC2.5-transgenic NOD mouse by screening a synthetic combinatorial peptide library. We report here an extensive array of decamer peptides that stimulate these transgenic T cells at concentrations in the 700 pM-10 nM range and, when activated, these T cells cause adoptive disease in NOD-scid mice. T cells derived from prediabetic NOD mice respond spontaneously to many of these peptides in culture, a direct demonstration that T cells activated during the spontaneous onset of insulitis and diabetes include subsets having specificity for these peptide mimics. Surprisingly, some of these peptides include sequences similar (8 of 10 residues) to those found within the 528–539 fragment of the 64-kDa protein GAD65 (KVAPVIKARMME). This 12-mer GAD65 fragment is only slightly stimulatory for BDC2.5 T cells (EC50 > 100 µM), although a longer 16-mer fragment, 526–541 (LSKVAPVIKARMMEYG) is effective in the low micromolar range.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
NOD/shi and NOD/Lt-scid/scid female mice and NOD.BDC2.5 TCR-transgenic mice were obtained from the mouse colony maintained at The Scripps Research Institute. The generation of the BDC2.5 mouse has been described by others (33). In this colony, BDC2.5 mice do not develop spontaneous diabetes, but do develop insulitis accompanied by the appearance of activated islet-specific T cells (34).
Libraries and peptides
Libraries and biased sublibrary mixtures were prepared at Mixture Sciences, Inc. (San Diego, CA) as described previously (35). PCL 97-4 is a synthetic N-acetylated, C-terminal amide, L-amino acid combinatorial peptide library arranged in a positional scanning format. It consists of 200 mixtures in the OX9 format where O represents one each of the 20 natural L-amino acids in a defined position and X represents all of the natural amino acids, with the exception of cysteine, in each of the remaining positions. For example, the first mixture has alanine (A) in position 1 (A1X9) whereas mixture number 200 has tyrosine (Y) in position 10 (X9Y10). Each OX9 mixture consists of 3.2 x 1011 (199) different decamer peptides in approximate equimolar concentration, and the total X10 library consists of 6.4 x 1012 (20 x 199) different peptides. Assuming an average m.w. of 1200 for a decapeptide mixture and a concentration of 100 µg/ml (83 µM), the concentration of each individual peptide is 2.6 x 10-16 M.
Individual peptides were synthesized by the simultaneous multiple peptide synthesis method (36). Purity and identity of each peptide were characterized using an electrospray mass spectrometer interfaced with a liquid chromatography system.
Culture conditions
Whole spleen cells from 4- to 6-wk-old BDC2.5 TCR-transgenic mice were depleted of erythrocytes and cultured (200–300 x 103 cells per well) in microtiter plates in standard T cell medium containing PCL 97-4 mixtures at 100 µg/ml. Varying consecutive dilutions were used for stimulation with sublibrary mixtures and individual peptides. Cultures were harvested at 72 h following overnight exposure to 0.5 µCi [3H]TdR (6.7 Ci/mM) and incorporated radioactivity was assessed by scintillation counting. Medium used for mouse T cells consisted of RPMI 1640 (Fischer Scientific, Pittsburgh, PA) supplemented with 8% FCS (J.R. Scientific, Woodland, CA), HEPES buffer (10 mM; Sigma, St. Louis, MO), 2-ME (50 µM 2-ME; Bio-Rad, Richmond, CA), penicillin-streptomycin (5 U/ml and 50 µg/ml, respectively; BioWhittaker, Walkersville, MD), and glutamine (2 mM; Sigma).
Determination of EC50 values of peptides and mixtures
T cell populations were cultured using conditions described above with varying dilutions of peptides and sublibrary mixtures. The concentration causing a half-maximal proliferative response (EC50) was determined by curve-fitting using a scientific graphics software program (GraphPad Prism; Graph Pad Software, San Diego, CA). The maximal proliferative response was fixed using the mean of the highest values obtained in each experiment.
Adoptive transfer of activated BDC2.5 T cells
Whole spleen cells from young BDC2.5 TCR-transgenic mice were cultured in mouse T cell medium for 4 days with selected stimulators at the specified concentrations. On day 4, 5 x 106 cells were injected i.v. into groups of five NOD-scid mice 10 wk of age. Control animals received a similar number of nonstimulated BDC2.5 T cells. Glucose levels were determined from tail vein blood samples using Glucometer Elite test strips (Bayer, Elkhart, IN) on a standard glucometer with a range of 20–400 mg/dl. Mice were monitored twice each week after cell transfer; mice with two consecutive values greater than 300 mg/dl were considered diabetic.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
T cells from BDC2.5 mice were stimulated in duplicate microwell
cultures in four different experiments conducted at different times
with the 200 mixtures of the PCL 97-4 positional scanning combinatorial
decapeptide library; control cultures were left unstimulated. The
results of a typical experiment (Fig. 1
)
show unambiguous responses with stimulation indices (SI) 3–4 times
over background levels to mixtures with proline defined at position 4
(P4) and methionine at position 9 (M9). At other positions a number of
mixtures having different defined amino acids showed responses
substantially over background levels (>1.5x). For these positions,
rankings of the five most active mixtures were recorded for each
experiment. The results of this analysis are summarized in Table I as "library selections"; here the
amino acids recorded at each position represent the most active
mixtures in at least two experiments, shown in order of decreasing
activity. Because all of the possible sequences based on the selected
amino acids listed in Table I would require the synthesis of >40
million different peptides, a different strategy for identifying the
active peptides was required.
|
|
shows the activity of these two
peptide mixtures with dual fixed positions in comparison to the single
fixed position mixtures of the PCL 97-4 library. The P4 M9 mixture was
substantially more active (EC50 = 43 µg/ml)
compared with the others (>100 µg/ml).
|
in the other positions. The results with this second set of
mixtures, including SI and EC50 values (Table II), indicated that the mixture with both P4 and M9 fixed was more
active (EC50 = 1 µg/ml) than those with P4 or
M9 only; Y9 mixtures with or without P4 were only slightly active.
Having established the requirement for the amino acids P4 and M9 in the
most active mixtures for stimulating BDC2.5 T cells and the fact that
the N37-103-biased sublibrary, itself, is active, the next step in the
deconvolution strategy was to determine the peptides responsible for
the activity of N37-103P4 M9. This was accomplished with the synthesis
and screening of a positional scanning sublibrary (N66) based on the
biased 103P4 M9 mixture. The EC50 values of each
of the mixtures in this positional scanning biased sublibrary are shown
in Table III, and the most active of
these, selected to make individual peptides, are indicated in Table IV. In addition to activity, these
selections were based on chemical character. For positions 1, 2, 3, and
5, where most of the mixtures were active, we selected those with
EC50 values <2.1 µg/ml (a value three times
the EC50 of the 103P4 M9 control mixture). M at
position 3 was omitted because of its chemical similarity to L; E and G
were both included at position 10 due to their different chemical
character. From these selections, 120 different peptides (1040 series)
were synthesized, and their activity on BDC2.5 T cells is shown in
Table V. More than half of these peptides
were highly active with EC50 values ranging from
700 pM to 10 nM; these are listed, along with their sequences and
EC50 values, in Table VI.
|
|
|
|
Having available 120 individual peptides (1040 series), many of
which are highly active, makes it possible to assess the relative
contributions of the different amino acids in their ability to activate
proliferative responses by BDC2.5 T cells. Of the thirty most active
peptides shown in Table VI, all use E at position 10; most use V rather
than G at position 7; and, they use either L or R at position 3. For
the remaining positions, there was no clear preference of possible
amino acid residues.
A more complete motif analysis of all 120 individual BDC2.5 ligands is
shown in Table VII. Because all peptides
contain P4, L5, W6, R8, and M9 (Table IV), they were grouped on the
basis of the amino acids defined at positions 3, 7, and 10. The results
reveal an obvious motif correlation with activity, the most optimal
motifs being R3V7E10, L3V7E10, and L3G7E10; 42 of 62 peptides with
EC50 values <10 nM use these motifs. In
contrast, the least optimal motif was R3G7G10; 10 of 13 peptides with
EC50 values >100 nM use this motif.
|
Searches of the SWISS-PROT and TrEMBL data bases with the
ScanProsite program (available at
http://www.expasy.ch/tools/scnpsit2.html; February, 2000) using the
amino acid predictions (Table I) used in the synthesis of the N37-103P4
M9 mixture, with P4 and M9 defined, revealed no relevant matches.
However, a search of the MHC motif binding database at the SYFPEITHI
website (http://www.uni-tuebingen.de/uni/kxi/) for peptides that bind
to the MHC class II molecule of the H2-g7
haplotype revealed that, except for a 2-aa insert, a fragment of GAD65
524–543 (SRLSKVAPVIKARMMEYGTT) has 9 residues
(noted in bold) in common with the residues that comprise the N37-103P4
M9 mixture (Table VIII). Based on this
finding, we assessed the activity on BDC2.5 T cells of a series of
overlapping decapeptides within the GAD65 524–541 sequence. None of
these peptides were found to be active (Table VIII). However, the
N37-103P4 M9 mixture was active in stimulating BDC2.5 T cells
(EC50 = 1 µg/ml; SI = 16), an indication
that altered peptide sequences imposed on a basic GAD backbone sequence
might be effective. Accordingly, a series of five different decamer
peptides were synthesized (Table VIII; 1047 series, peptides 1–5) For
these peptides, seven positions were defined from the GAD sequence: V2,
A3, P4, V5, R8, M9, and M10; amino acid selections for the other
positions were based on the preferences defined from the 103P4 M9 scan
(Table IV) and activity of the 1040 series (Tables VI and VII). K1 was
included because of its presence (K528) in the GAD65 sequence.
EC50 activity of these peptides was found to be
in the 30 ng/ml to >1 µg/ml range.
|
; series 1047, peptides
6–10) were synthesized with E at position 10. This substitution of M10
with E10 was based on the fact that most of the active peptides derived
from the library deconvolution (Table VI) used a terminal E10. This
synthesis resulted in GAD-like peptides having a 1000-fold improvement
in their EC50 activities on BDC2.5 T cells in the
low nanomolar range.
As an attempt to maximize similarity to GAD while retaining the ability
to stimulate BDC2.5 T cells, we synthesized a 12-mer peptide (1068-55)
having five N-terminal amino acids and five C-terminal amino acids of
the GAD65 sequence while substituting I and K at positions 6 and 7 with
W and V. This peptide proved to be only slightly active
(EC50 > 100 µg/ml; SI = 11). However, a
similar, 11-mer peptide, in which A8 was deleted, showed activity in
the 100-nM range. Finally, we explored the possibility that a longer
fragment of this GAD65 region might be effective. An
N-terminal-acetylated 16-mer fragment, 526–541 (LSKVAPVIKARMMEYG),
was shown to be active with an EC50 of 4.3
µg/ml (2.3 µM) and an SI of 17 at 1 µg/ml (Table VIII).
Fig. 2 shows a summary of the
similarities of active peptides and the GAD65 528–539 sequence. The
principal findings are 1) the 12-mer GAD 528–539 sequence (p1040–133)
is only marginally active (>100 µg/ml); 2) substituting W and V in
positions 6 and 7 (p1068-55) results in no improvement; 3) an 11-mer
peptide (p1068-56) with these substitutions and a deletion of A8 in the
GAD sequence shows substantially improved activity (100 ng/ml); and 4)
a decamer peptide (p1047-10), also with these W6 and V7 substitutions,
and A8 and M9 deletions, terminating with E10, shows even greater
improvement in activity (4.2 nM). Finally, the native GAD65 sequence
526–541 shows activity in the low micromolar range. Also shown for
comparison are two of the most active superagonist peptides (0.7–1.2
nM) derived from the library deconvolutions.
|
Pathogenicity of peptide-stimulated BDC2.5 T cells
Results shown in Table IX explore
the pathogenicity in vivo following adoptive transfer of BDC2.5 T cells
after stimulation in culture with selected peptides or a biased
sublibrary mixture. Cells activated with the marginally active peptide,
GAD65 528–539 (1040-133, EC50 > 100 µg/ml),
and cells stimulated with the biased library mixture (103P4 M9) caused
diabetes in all of NOD-scid recipients within 2 wk following
transfer. Disease onset appeared more promptly in recipients of T cells
stimulated with the superagonist peptide (p1040-63).
|
To determine whether T cell populations from prediabetic NOD mice
that are spontaneously activated during the course of disease onset
contain subsets that are specific for and respond to the BDC2.5 ligands
defined from the library deconvolution in this study, we isolated
draining pancreatic lymph node cell populations from 7-wk-old
prediabetic NOD mice; these were cultured with three different
superagonist mimics (30 µg/ml). The results (Fig. 3) show unambiguous proliferative
activity by T cell populations in response to two of these peptides, a
clear indication that at least a portion of the T cell population in
prediabetic NOD mice activated during the course of onset of
spontaneous disease are reactive to the BDC2.5 ligand mimics.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
- and
-chains derived from the diabetogenic CD4+ T
cell clone, BDC2.5. The original BDC2.5 T cell clone was derived from a
NOD mouse, is highly diabetogenic when transferred to young
asymptomatic NOD mice, and is reactive to an unknown cell islet Ag
in association with H-2g7 (16). Deconvolution strategy
This search for the BDC2.5 ligand(s) was conducted with the use of
a synthetic combinatorial peptide library arrayed in a positional
scanning format, an approach that has been used previously to identify
a large series of peptide ligands specific for T cell clones reactive
to series of known and unknown Ags (38, 39, 40, 41, 42, 43, 44). The initial
scan with the PCL 97-4 library (Fig. 1) indicated that mixtures with P4
fixed and others with M9 fixed were most active in stimulating BDC2.5 T
cells; mixtures with other fixed amino acids at these two positions
were also active, but somewhat less so. In addition, numerous mixtures
with defined amino acids at the other positions were active but no one
mixture seemed to show unambiguous activity. This substantial number of
amino acid possibilities at each of the positions (Table I) precluded
synthesis of such a large array of candidate sequences and their
activity assessments on BDC2.5 T cells.
This situation required us to develop a different strategy using a biased positional scanning sublibrary for the identification of T cell ligands contained in the active mixtures. Although biased positional scanning libraries have been used by others for identification of peptide ligands for T cell clones (45, 46), to our knowledge, this is the first time a biased positional scanning sublibrary has been designed based on results obtained from scans made with an unbiased positional scanning library.
The first step was to synthesize a series of biased sublibrary mixtures
based on the library selections of Table I and to assess their activity
on BDC2.5 T cells. For example, mixture N37-103, containing
1000-fold fewer peptides than the PCL 97-4 mixtures, showed a
substantial degree of activity (Table II). This finding clearly
indicates that this less diverse biased sublibrary mixture contained a
subset of active peptide ligands. Moreover, the substantial increase in
activity of the biased mixture N37-103P4 M9 clearly indicates the
importance of P and M in positions 4 and 9 of the active peptides, and
permitted continuation of the deconvolution process with fewer possible
amino acids for each position.
The next step was to construct a positional scanning sublibrary based
on the N37-103P4 M9 mixture and to assess its activity on BDC2.5 T
cells. From these results (Table III), it was possible to eliminate
two-thirds of the amino acid possibilities of Table I and to focus on a
less diverse array of possible candidate residues important for
stimulating proliferative activity (Table IV). Of the 120 different
possible peptide sequences that were synthesized based on the library
predictions of Table IV, more than half proved to be highly active with
EC50 values in the 0.1–10 nM range (Tables V and VI). Having successfully identified peptide ligands with this
deconvolution strategy clearly demonstrates that when the activity of a
highly complex mixture (i.e., PCL 97-4, N37-103, or the N66 series,
Tables II and III) is reproducible, active individual compounds can
then be identified.
Analysis of BDC2.5 ligand sequences
During the different steps of the deconvolution process, in which
mixtures or individual compounds were assessed by their ability to
stimulate BDC2.5 T cells, it was found that the presence of P4, L5, W6,
R8, and M9 were important for activity. Also, it was possible to
determine optimal amino acid motifs present in the most active peptides
(R3 V7 E10, L3 V7 E10, and L3 G7 E10) based on activity assessments of
all of the individual peptides (1040 series, Tables IV and VII). In
addition, peptides identified in this study were aligned with the
proposed MHC binding motifs for H-2g7 recently
described with peptide binding competition assays (47) and
the recent x-ray structure of H-2g7 complexed
with peptide (48). This analysis revealed that the amino
acids found to be important at positions 5, 7, and 10 correspond to the
anchor residues for the MHC pockets 4, 6, and 9. More importantly, this
finding suggests that P4, W6, R8, and M9 are likely to be involved in
TCR recognition. Studies that address the importance of these residues
using alanine substitution analogs clearly demonstrate that peptides
having the H-2g7 binding motifs but with
substitutions at proposed TCR contact residues are not stimulatory
(data not shown). We are currently using the information obtained with
the screening of the peptide library and the activity of the individual
peptides in conjunction with a biometrical analysis (44)
to search protein databases to identify murine, viral, and bacterial
sequences that could be potential candidates for triggering activation
of this diabetogenic T cell clone. A number of sequences within known
proteins have already been identified, and the biological relevance of
these sequences is currently under study.
Immune properties of BDC2.5 ligands
The BDC2.5 ligands defined in this study are highly active in
stimulating BDC2.5 T cells (Table VI), these activated cells cause a
prompt onset of disease in NOD-scid animals (Table VIII),
and more importantly, they are capable of stimulating polyclonal T cell
populations from young, prediabetic NOD mice that have been
spontaneously activated in vivo (Fig. 3). This latter finding implies
that these peptides are related to a critical autoantigenic epitope
underlying the disease process in these animals. Further, assuming that
the spontaneously activated T cell subset in autoimmune diabetes in NOD
animals is polyclonal, it indicates that the specificity of the BDC2.5
T cell clone for this set of peptide mimics is shared by numerous
different cells in the NOD T cell repertoire that may be involved in
the disease process.
Does direct immunization of prediabetic NOD mice with these active peptides cause accelerated disease onset? These BDC2.5 ligands are effective immunogens; they provoke strong T cell responses in culture but, to date, we have not been able to cause accelerated diabetes in these animals in vivo (studies in progress). Nevertheless, the availability of an extensive array of peptide mimics that stimulate BDC2.5 T cells provides an important probe for the further exploration of molecular details of TCR/peptide/MHC interactions involved in autoimmune diabetes at the level of a single clone.
Similarity of BDC2.5 ligands to GAD65
Is GAD65 or one of its fragments causally involved in the etiology
of spontaneous diabetes? This enzyme has long been suspected as one of
the diabetogenic autoantigens in both man and mouse, although much of
the evidence for its involvement is indirect (25, 26, 27, 28, 29, 30, 31).
Stronger support for this possibility is provided by the recent
demonstration in NOD mice bearing anti-sense GAD transgenes that
selective suppression of GAD expression in pancreatic cells both
reduced insulitis and blocked development of disease in these animals
(32). In this study, one of the most active peptides in
stimulating pathogenic activity by BDC2.5 T cells, 1047-10
KVAPVWVRME with an EC50 = 4
nM, is remarkably similar to the sequence of GAD65 528-539
(KVAPVIKARMME) although the latter shows only minimal activity (Fig. 2). More telling is the finding that a longer GAD65 fragment, 526–541
(LSKVAPVIKARM MEYG), shows activity in the low micromolar range
(Table VIII and Fig. 2). This result adds further support for the
possibility that an epitope within this region of the GAD65 sequence is
the native autoantigen peptide involved in the spontaneous activation
of the original BDC2.5 T cell clone. If so, the BDC2.5 T cell clone
appears to have low avidity for this peptide/MHC complex, but
sufficient to cause spontaneous disease.
However, two problems remain with this notion. First, only rarely has deliberate delivery of GAD65 or one of its fragments to prediabetic NOD mice been associated with more rapid onset of disease (26, 31). Second, as indicated above, the mimic peptides (some of them very similar to the relevant GAD fragment) that are highly active on these cells in culture have not yet been shown to cause disease in vivo. These findings appear to indicate that the circumstances involving natural (auto)antigen presentation during onset of spontaneous disease in vivo and Ag presentation following deliberate immunization lead to quite different outcomes.
An alternative explanation derives from the fact that native GAD65 fragments fail to cause accelerated diabetes in NOD mice following direct immunization. This might reflect the presence of T cells in the NOD repertoire with low affinity for this peptide/MHC complex, and that the disease-causing epitope in NOD mice is a cross-reactive peptide derived from a different protein. The possibility that NOD mice express an allelic difference reflected in the GAD sequence was explored by direct sequence analysis of GAD65 cDNA and found not to be different from the previously reported sequence.
Recently completed studies by Quinn and colleagues provides an
interesting set of findings that should be considered in the context of
the current results described above and provide some insight that may
ultimately help to resolve this
conundrum.4 They show
that the GAD65 524–543 sequence contains two overlapping epitopes that
stimulate two functionally different sets of T cells with different TCR
-chains. T cells from NOD mice immunized with the 524–543 fragment
respond in culture to 527–538, use V12, and display regulatory
activity by protecting against disease onset. In contrast, T cells from
young prediabetic NOD mice spontaneously respond to 530–543, use
V4, and show no such regulatory activity. It is of some interest
that one of the GAD-like mimic peptides (1047-10) described here,
highly active on BDC2.5 T cells that use TCRV4 and capable of
inducing pathogenicity by these cells, contains 8 residues in common
with the intermediate GAD65 528–539 sequence. Studies currently in
progress are designed to determine whether this mimic peptide provokes
conflicting responses in vivo involving both of these T cell
subsets.
In summary, we report here for the first time a large series of peptide sequences that are highly specific, with EC50 values in the picomolar to low nanomolar range, in stimulating pathogenicity by clonotypic T cells of the BDC2.5 NOD mouse. Several of these peptides are structurally similar to a fragment of GAD65, and a 16-mer fragment of GAD65 stimulates these T cells with an EC50 value in the low micromolar range, findings that support the notion that this protein may be a key factor in autoimmune diabetes. BDC2.5 cells stimulated in culture with these highly active peptide analogs transferred adoptively prove to be highly diabetogenic in vivo. The mimic peptides also stimulate proliferative responses by spontaneously activated T cells derived from prediabetic NOD mice, a finding that indicates that one of the events involved in spontaneous onset of autoimmune diabetes in the NOD mouse involves activation of a significant fraction of the T cell repertoire having specificity for a peptide ligand structurally similar to these peptide mimics.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Darcy B. Wilson, Torrey Pines Institute for Molecular Studies, 3550 General Atomics Court, San Diego, CA 92121-1122.
3 Abbreviations used in this paper: NOD, nonobese diabetic; GAD65, glutamic acid decarboxylase; SI, stimulation index/indices.
4 A. Quinn, B. McInerney, E. P. Reich, O. Kim, K. Jensen, and E. E. Sercarz. 2000. Regulatory and effector CD4 T cells in NOD mice recognize overlapping determinants on GAD65 and use distinct V genes. Submitted for publication.
Received for publication June 16, 2000. Accepted for publication October 23, 3000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
T cell receptor monoclonal antibody provides an efficient therapy for autoimmune diabetes in nonobese diabetic NOD mice. Eur. J. Immunol. 21:1663.[Medline]
cell-reactive T cells is not restricted in non-obese diabetic mice. J. Exp. Med. 173:1091. granule antigen. Diabetes 43:197.[Abstract]
-cell antigens. Diabetes 46:1975.[Abstract]
cells. Science 284:1183.This article has been cited by other articles:
|
|
 |
|
  J. Wang, S. Cho, A. Ueno, L. Cheng, B.-Y. Xu, M. D. Desrosiers, Y. Shi, and Y. Yang Ligand-Dependent Induction of Noninflammatory Dendritic Cells by Anergic Invariant NKT Cells Minimizes Autoimmune Inflammation J. Immunol., August 15, 2008; 181(4): 2438 - 2445. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Godebu, D. Summers-Torres, M. M. Lin, B. J. G. Baaten, and L. M. Bradley Polyclonal Adaptive Regulatory CD4 Cells That Can Reverse Type I Diabetes Become Oligoclonal Long-Term Protective Memory Cells J. Immunol., August 1, 2008; 181(3): 1798 - 1805. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Li, B. Wang, J. A. Frelinger, and R. Tisch T-Cell Promiscuity in Autoimmune Diabetes Diabetes, August 1, 2008; 57(8): 2099 - 2106. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Feuerer, W. Jiang, P. D. Holler, A. Satpathy, C. Campbell, M. Bogue, D. Mathis, and C. Benoist Enhanced thymic selection of FoxP3+ regulatory T cells in the NOD mouse model of autoimmune diabetes PNAS, November 13, 2007; 104(46): 18181 - 18186. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Saxena, J. K. Ondr, A. F. Magnusen, D. H. Munn, and J. D. Katz The Countervailing Actions of Myeloid and Plasmacytoid Dendritic Cells Control Autoimmune Diabetes in the Nonobese Diabetic Mouse J. Immunol., October 15, 2007; 179(8): 5041 - 5053. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Li, N. Perez, S. Karumuthil-Melethil, and C. Vasu Bone Marrow Is a Preferential Homing Site for Autoreactive T-Cells in Type 1 Diabetes Diabetes, September 1, 2007; 56(9): 2251 - 2259. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Ueno, S. Cho, L. Cheng, J. Wang, S. Hou, H. Nakano, P. Santamaria, and Y. Yang Transient Upregulation of Indoleamine 2,3-Dioxygenase in Dendritic Cells by Human Chorionic Gonadotropin Downregulates Autoimmune Diabetes Diabetes, June 1, 2007; 56(6): 1686 - 1693. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. D. Holler, T. Yamagata, W. Jiang, M. Feuerer, C. Benoist, and D. Mathis The same genomic region conditions clonal deletion and clonal deviation to the CD8{alpha}{alpha} and regulatory T cell lineages in NOD versus C57BL/6 mice PNAS, April 24, 2007; 104(17): 7187 - 7192. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Luo, K. V. Tarbell, H. Yang, K. Pothoven, S. L. Bailey, R. Ding, R. M. Steinman, and M. Suthanthiran Dendritic cells with TGF-beta1 differentiate naive CD4+CD25- T cells into islet-protective Foxp3+ regulatory T cells PNAS, February 20, 2007; 104(8): 2821 - 2826. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. V. Tarbell, L. Petit, X. Zuo, P. Toy, X. Luo, A. Mqadmi, H. Yang, M. Suthanthiran, S. Mojsov, and R. M. Steinman Dendritic cell-expanded, islet-specific CD4+ CD25+ CD62L+ regulatory T cells restore normoglycemia in diabetic NOD mice J. Exp. Med., January 22, 2007; 204(1): 191 - 201. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. J. Mellanby, C. H. Koonce, A. Monti, J. M. Phillips, A. Cooke, and E. K. Bikoff Loss of Invariant Chain Protects Nonobese Diabetic Mice against T |
