| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Microbiology and Immunology, University of Arizona, Tucson, AZ 85724; and
La Jolla Institute for Allergy and Immunology, San Diego, CA 92121
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
islet cell destruction. Because
IDDM is due, in part, to one’s genetic background (1), a
cure may be affected by manipulation of the genome. However, this
approach is not heritable and may cause serious harm to the patient via
as yet unknown mechanisms. A potentially safer approach might be to
remove (or severely dampen) the capacity of a predisposed individual’s
T cells to respond to the pancreatic Ag(s) that are most directly
involved in the initiation and/or chronic activation that leads to the
destruction of islet cells. Several candidate islet Ags have already
been identified. Given more extensive research, IDDM-related immune
responses to a limited number of these Ags may be correlated with the
specific MHC backgrounds found in given individuals (2, 3). This information has been used to develop a strategy for
preventing the induction of IDDM by treating patients with diabetogenic
Ags before their disease had fully manifested itself
(4, 5, 6, 7, 8). However, in some initial clinical studies, IDDM
onset was not altered by diabetogenic Ag treatment, and evidence
suggested the possibility that such treatment may accelerate disease
induction (9). Because of their similarities to patients with human IDDM (10), nonobese diabetic (NOD) mice have been studied to determine what islet-associated Ags are recognized early in disease. In two studies, the earliest spontaneous responses in preclinical mice were directed against proteins or peptides of human glutamic acid decarboxylase (GAD65) (11, 12). The fact that mice spontaneously developed reactivity to these peptides before the development of reactivity to other islet-associated components led some researchers to hypothesize that GAD65 may play a role in disease initiation (11, 12) through the process of determinant spreading (13). More recently, additional immunogenic epitopes from GAD have been described (13); their role in disease pathogenesis remains unclear, but they appear to be responsible for a large part of the T cell response to GAD65. Neonatal mice treated with tolerizing doses of GAD65 protein were protected from spontaneous IDDM induction (11). Further, NOD mice were protected from IDDM by administration of GAD at 3 wk of age (12). Together, these studies suggest that autoimmune reactivity to GAD65 is an important step in the progression toward clinical disease in NOD mice.
One potential problem when using the NOD model is the ability of general immune stimulation to block efficient IDDM induction. Reports have demonstrated IDDM protection by treatment of NOD mice with LPS (14) or poly(I:C) (15). Indeed, NOD mice that are not kept in largely pathogen-free environments succumb to IDDM at significantly reduced rates (10). Thus, a second interpretation of GAD65-induced IDDM modulation could be that the high dose of Ag used to stimulate the immune system prevents animals from becoming clinically diabetic in a manner independent of the specific determinant. Some researchers, for example, have found that whereas GAD peptide immunization could not prevent IDDM in NOD mice, diphtheria-tetanus toxoid-acellular pertussis alone or insulin B chain peptide could mitigate IDDM (16). Furthermore, recent work also suggests that intrathymic injection of NOD mice with whole pancreas tissue or insulin B chain peptides could prevent IDDM but that certain GAD peptides accelerate disease (17).
To explore the potential role of GAD65 reactivity in the inhibition or acceleration of NOD IDDM, several experiments were performed using peptides from GAD65 to affect the generation of diabetogenic T cells in vivo as well as in the more controlled NOD fetal thymus organ cultures (FTOC) in vitro IDDM (ivIDDM) system.
NOD mice exposed only in utero to GAD peptides developed IDDM at a significantly lower rate. This limited protection is postulated to be due to increased intrathymic presentation of GAD peptides, leading to a reduction in the initial number of GAD-autoreactive T cells that would reach maturity. Thus, experiments to directly measure the effects on islet responsiveness after early T cell tolerance to GAD were performed. Previously, we had shown that FTOC that has been precultured for 14 days can inhibit the production of insulin in subsequently added fetal pancreas, as well as provide T cells which migrate into the cocultured fetal pancreas (FP) causing insulitis-like lesions (18). This system allows for the evaluation of the effects of immunomodulators on the development of diabetogenic T cells without the involvement of peripheral regulatory mechanisms. The ivIDDM activity of these cultures can be attenuated by coculturing the developing FTOC while in contact with NOD fetal pancreas organ cultures from the first day of culture. Coculturing presumably induces tolerance by exposing pre-T cells to high doses of islet Ags, leading to their inactivation or deletion. To determine whether GAD65 peptides could confer the same protective effects (and mimic those of in utero exposure), FTOC were treated with increasing doses of soluble peptide from the initiation of culture. Those T cells that were cultured in the presence of GAD65 peptides were compared with those from NOD FTOC that were cultured with control peptides or in the absence of high dose peptides to determine whether GAD65 treatment could alter the ability of NOD FTOC to inhibit insulin production by the subsequently added FP. Indeed, FTOC that are normally ivIDDM diabetogenic were rendered benign by early exposure to certain GAD peptides, suggesting that this specificity is critical to efficient T cell involvement in recognition and islet cell destruction.
A similar approach was taken to evaluate the acceleration of IDDM in NOD mice treated with GAD peptides as young adults. In this case, NOD FTOC were treated with GAD peptides after the production of mature T cells, increasing their ability to react with these Ags. When ivIDDM was subsequently performed, those FTOC that had been primed to GAD peptides demonstrated an increased induction of this in vitro correlate to diabetes.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
GAD65246–266 (NMYAMMIARFKMFPEVKEKG), GAD65509–528 (IPPSLRYLEDNEERMSRLSK) L for N at position 8 as reported in Ref. 2 ; GAD65524–543 (SRLSKVAPVIKARMMEYGTT), prostate-specific antigen (PSA) 95–109 (SFPHPLYDMSLLKNR), myelin basic protein (MBP) 123–137 (PRTPPPSQGKGRGLS), and hen egg lysozyme (HEL) 11–25 (AMKRHGLDNYRGYSL) peptides were synthesized by Biosyn Biotechnology and Research Genetics using standard and proprietary Fmoc chemistry (12), and a purity of >80% was established by HPLC and mass spectrometer measurements. All peptides were dissolved in HPLC grade water before addition to FTOC in sterile organ culture DME (in vitro). The GAD peptides and the control HEL peptide are recognized by the IAg7 MHC class II molecules expressed in NOD mice, whereas the PSA and MBP control peptides are not.
Mice
Breeding pairs of NOD/Lt mice were obtained as a gift from the laboratory of Dr. E. Leiter at The Jackson Laboratory (Bar Harbor, ME). Our colony was maintained in a specific pathogen-free vivarium at the University of Arizona Central Animal Facility and propagated by brother-sister mating. Mice are allowed free access to standard breeder chow (S-2335 irradiated breeder chow; Harlan Teklad, Madison WI) and autoclaved drinking water. The incidence of IDDM in NOD/Lt females in our colony at the University of Arizona is >80–90% by 40 wk of age. NOD/Lt mice were then bred to produce timed-pregnant females. The fetuses were removed from pregnant females at the indicated time points (plug date = day 0). We consistently found that our animals were variable with regard to their stage of development, even though they had been vaginally plugged on the same day. We therefore restaged the pups that we obtained based on their developmental characteristics (such as digit separation on the paws) as given in Ref. 19 .
Fetal thymus/pancreas organ culture
The organ culture methods used have been described in detail by our laboratory and others (20). Briefly, at least 6 thymus lobes and/or equal numbers of pancreata (usually 10) dissected from 13- to 16-day gestation fetal mice were placed on the surface of Millipore (25 mm thick, 0.45 mm pore size) filters supported on blocks of surgical Gelfoam (Upjohn, Kalamazoo, MI) in 3 ml medium in 10 x 35 mm plastic petri dishes. DMEM (4.5 g/L D-glucose), supplemented with 20% FBS (HyClone Laboratories, Logan, UT) was used. The medium also contained streptomycin (100 mg/ml), penicillin (250 mg/ml), gentamicin (10 mg/ml), nonessential amino acids (0.1 mM), sodium pyruvate (1 mM), 2-mercaptoethanol (2 x 10-5 M), as well as 3.4 g/L sodium bicarbonate. The cultures were grown in a fully humidified incubator in 5% CO2 in air at 37°C. Cells were harvested from FTOC and fetal pancreas organ cultures using collagenase digestion, as has been reported previously (20), or by manual dissociation by teasing tissue from the Millipore strips in HBSS plus 5% FBS solution. Viability was consistently high (>90%) for both thymic and pancreatic tissue, as determined by trypan blue exclusion.
Supernatant recovery and insulin RIA
At the specified days of culture, 300 µl culture supernatant were removed from each 3-ml culture dish for insulin RIA. MicroMedic insulin RIA kits were obtained from ICN MicroMedic Systems (Horsham, PA), used to measure the quantity of insulin in culture supernatants as directed by the manufacturer’s instructions, and standardized to bovine insulin. After the amount of insulin was measured in a 300-µl sample, the amount of insulin in the 3-ml culture was determined. This value was divided by the number of FP to give the amount of insulin secreted per pancreas (µIU per pancreas). The total amount of insulin in each culture was measured without regard to the degradation of insulin in culture and represents the total accumulation of insulin during the time interval reported.
Proliferation assays
NOD FTOC were cultured for 14 days in standard FTOC, as
previously reported (20), to produce phenotypically mature
populations of TCR-bearing T cells (18). Except for
studies of spontaneous responses, after the initial culture period of
14 days either in the presence from day = 0
("tolerance"-inducing) or absence (in vitro "priming") of
synthetic peptides (see Results), FTOC were treated with a
mixture of soluble GAD peptides (10 µg/ml), the control peptide (10
µg/ml), or fresh medium. After a final additional 48 h in organ
culture, cells were removed from FTOC, dissociated, counted, and plated
in triplicate at 2–4 x 105 cells/well in
96-well plates. Cells were then immediately challenged in 10 µg/ml
GAD65 peptides 246–266, 509–528, and 524–543 or the control peptides
(10 µg/ml) for 48 h. After this stimulation period, FTOC cells
were pulsed with 1 µCi [3H]thymidine for an
additional 24 h, mechanically harvested, and counted using a
Packard scintillation counter. In addition to GAD65 peptides and
the nonimmunogenic PSA peptide (10 µg/ml), identical FTOC populations
were also challenged with Con A (2.5 µg/ml) as a positive control.
Stimulation index (SI) was calculated as the test counts divided by the
background control standard (as defined in the text).
Peptide treatment of pregnant and young adult mice and assessment of IDDM in vivo
Pregnant NOD/Lt mice, at 14–15 days gestation, received a single injection (i.p.) consisting of 50 µg peptide in 250 µl sterile PBS. Treatment consisted of one of the following peptides or peptide combinations: 1) GAD65246–266; 2) GAD65509–528; 3) GAD65524–543; 4) GAD65509–528 and GAD65524–543 combined (50 µg of each peptide per mouse); 5) GAD65246–266, GAD65509–528, and GAD65524–543 combined (50 µg of each peptide per mouse), 6) HEL11–25; 7) MBP123–137; 8) MBP123–137 and PSA95–109 combined (50 µg of each peptide per mouse). Injection groups 1–5 consisted of autoreactive GAD65 IAg7-binding peptides. Injection group 6 consisted of HEL11–25, an IAg7-binding peptide known to produce proliferative responses in NOD mice. Injection groups 7 and 8 consisted of peptides unable to be bound by IAg7.
Age-matched 3- to 4-wk-old littermate prediabetic female NOD mice were injected once IP with a mixture of GAD65509–528 and GAD65524–543 at a low (0.5 µg of each peptide per mouse, n = 5) or high (50.0 µg of each peptide per mouse, n = 5) dose, or 50 µg/mouse of an irrelevant control peptide (PSA95–109, n = 5) in PBS.
Mice were assessed for the development of IDDM by testing every 10–14 days (after 10 wk of age for the offspring of pregnant mice and after 15 wk for young adults) for glycosuria. Those mice with a urine glucose reading of >200 mg/dl (GlucoStix) were removed from the colony and retested for hyperglycemia with a digital blood glucose monitor. Those mice with a confirmed blood glucose level >250 mg/dl were considered diabetic.
Statistical analysis
Data were analyzed using StatView 4.5 from Abacus Concepts (Berkeley, CA). Experimental differences in IDDM incidence in treatment groups were assessed by Kaplan-Meier life table analysis using the log rank (Mantel-Cox) test for significance. An unpaired Student t test was used for final determination of the significance of the effects of treatments given to organ cultures. p < 0.05 was considered statistically significant for all statistical analyses.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Early reports have established that GAD65 responsiveness is found in 3-wk-old NOD mice (11, 12) and that tolerance to GAD65 protein is protective (11). However, some reports suggest that GAD peptides can accelerate IDDM (17). Thus, to determine whether IDDM severity or onset in NOD mice could be increased or accelerated, GAD65 peptides were given in tolerogenic and stimulatory doses to prediabetic mice.
Age-matched 3- to 4-wk-old littermate female NOD mice were injected
once i.p. with either 50 µg each of a mixture of GAD65 peptides or 50
µg of the control peptide in saline. Mice were then followed for the
development of hyperglycemia (Fig. 1
).
Prediabetic female NOD mice treated with a 0.5-µg dose of GAD65
peptides (n = 5) developed IDDM at an accelerated rate
as compared with control peptide (n = 5)-treated groups
(p = 0.0466). The high dose treatment appeared
to have protective effects on treated NOD mice which, in stringent
analysis, was nearly statistically significant in Kaplan-Meier survival
analysis (p = 0.0845).
|
Because the initiation of IDDM seemed so critically dependent on the responsiveness to GAD65 initially, we examined the ability of individual and mixtures of GAD peptides 246–266, 509–528, and 524–543 to alter the clinical effects of IDDM in NOD mice at the earliest possible time point.
At 14–15 days gestation, three groups of pregnant NOD mice were injected once i.p. with 50 µg GAD65246–266, GAD65509–528, or GAD65524–543 (GAD65 p17, p34, or p35 single); 50 µg of a mixture of GAD65509–528 plus 50 µg GAD65524–543 (GAD65 p34 plus p35 mix); or 50 µg of a mixture of GAD65246–266, 50 µg GAD65509–528, and 50 µg GAD65524–543 (GAD65 p17 plus p34 plus p35 mix). Controls consisted of 50 µg HEL11–25 (control HEL single, a tight binding peptide to IAg7 that elicits a strong immune response), 50 µg MBP123–137 (control MBP single, which does not bind to IAg7) and a mixture of 50 µg MBP123–137 plus 50 µg PSA95–109 (control MBP plus PSA mix). After weaning, male NOD mice were culled, and female offspring mice were monitored for the development of hyperglycemia (see Materials and Methods). This protocol was chosen to examine the potential benefits of using a wider range of GAD determinants for tolerance induction and disease prevention as compared with single peptides alone.
As compared with offspring of NOD mothers treated midgestation
with either a single control peptide (the
IAg7-binding HEL11–25 or
the nonbinding MBP123–137) or a mixture of
control peptides (PSA95–109 plus
MBP123–137), offspring of mice given some GAD65
peptide treatment regimens clearly had a reduced incidence of IDDM
(Fig. 2). NOD offspring treated with
GAD65246–266 singly (p =
0.035, n = 14) developed IDDM at a reduced rate when
compared with HEL control peptide. Offspring treated with the
GAD65509–528-GAD65524–543
mixture (p = 0.003, n = 10)
developed IDDM to a reduced overall level, and at a significantly
reduced rate of onset, as compared with age-matched control (PSA plus
MBP) mix-treated animals. However, offspring of mice treated with
GAD65509–528 (p = 0.202)
or GAD65524–543 singly
(p = 0.388) were not significantly protected
from disease when compared with animals treated with HEL. Offspring of
mice treated with a mixture of GAD65246–266,
GAD65509–528, and
GAD65524–543 were also not significantly
protected (p = 0.226, n = 9)
when compared with animals from dams treated with MBP plus PSA. Of
interest, however, is that the initial incidence of IDDM in the
three-peptide-treated group is similar to the controls at 18 wk, but
this value failed to climb until week 37, and then only slightly.
Animals given the mix with two peptides or single peptides developed
disease more slowly, but after 27 wk the incidence increased rapidly to
a value similar to that of the three-peptide-treated animals. Clearly,
the rapid rise of disease incidence in the three-peptide-treated
animals is responsible for the lack of significance of disease
protection in this group. However, some animals in all GAD
peptide-treated groups never developed disease. The
GAD509–528-GAD65524–543
mixture group has been conducted for >45 wk with 40% of the animals
still disease free. Some of the
GAD65246–266-treated mice have also been
followed for 45 wk, with a similar IDDM incidence.
|
The response in NOD FTOC is not specifically or spontaneously enhanced to GAD65 peptides
To determine whether NOD FTOC was spontaneously primed to GAD65 peptides without any exogenous stimuli, NOD FTOC were cultured to produce phenotypically mature populations of TCR-bearing T cells. After this initial culture period, cells were removed from FTOC and placed in 96-well plates. Cells were then challenged to proliferate in response to GAD65 peptides 246–266, 509–528, and 524–543 (10 µg/ml). In addition to GAD65 peptides, identical FTOC populations were also challenged with a nonimmunogenic peptide (PSA95–109) and Con A.
NOD FTOC-derived cells did not proliferate in response to GAD65246–266, GAD65509–528, or GAD65524–543, significantly (p > 0.3) above the level of the nonimmunogenic control peptide PSA (SI = 1.0) or FTOC cells left unchallenged (not shown). The cultures did respond to Con A stimulation (SI = 4.5 ± 0.5), and were thus considered competent to respond in a receptor-mediated manner. The response to Con A, which is lower than that of mature splenic T cells, is typical of the immature cells produced by organ cultures (21), and the response is roughly equivalent to those made by thymocytes during the first month after birth (22).
T cells in NOD FTOC can be primed to GAD65 peptides
Based on data described above, it appeared that a significantly enhanced spontaneous response to GAD65 peptides was lacking in FTOC. We wished to determine whether mature phenotype NOD FTOC indeed produced self-reactive cells in low abundance and could be enriched or activated/primed to produce GAD65-responsive T cells, given that this might normally occur in the NOD periphery.
NOD FTOC were cultured for 14 days in standard FTOC, as described previously (18) to produce phenotypically mature populations of TCR-bearing T cells. Peptide solutions of either a mixture of GAD65 peptides 246–266, 509–528, and 524–543; fresh medium; or the control PSA peptide were then added to FTOC to produce a final concentration of 10 µg/ml. After an additional 48 h in organ culture, cells were removed from FTOC and challenged vs each peptide individually in proliferation assays.
As shown in Fig. 3, NOD FTOCs that were
primed using the mixture of GAD65 peptides were significantly more
responsive to GAD65524–543
(p < 0.02) than were identical cultures that
were primed with the control peptide (p <
0.01) or left untreated (p < 0.01). Those
cultures primed using GAD65 peptides were also significantly more
responsive to GAD65509–528 than cultures
that were not primed (p = 0.045). In contrast,
the response to GAD65246–266 was not
significantly enhanced compared with control peptide
(p = 0.51) or untreated cultures
(p = 0.19). Cultures primed with PSA and then
challenged remained unresponsive to PSA (p =
0.10), whereas the Con A response was positive.
|
To determine the functional impact of specific autoantigen
stimulation on the ivIDDM model system, FTOC were primed with synthetic
GAD65 peptides after the development of phenotypically mature T cells,
at the initiation of coculture with FP (usually 14 days of culture).
Peptide solutions of either a mixture of the three GAD65 peptides at 10
or 100 µg/ml or 100 µg/ml control PSA peptide were then added to
FTOC, as was done previously to prime FTOC. This treatment increases
FTOC-proliferative responses to GAD65 peptides 509–528 and 524–543
(Fig. 3
). At this time, freshly procured NOD FP were added to each of
the GAD65-treated and control FTOC. At 7, 14, and 21 days
poststimulation/coculture, supernatants were removed from the cultures
and used to determine the effect of stimulated and unstimulated FTOC on
insulin production in the FP in coculture.
As shown in Table I, at the peak of
ivIDDM (day 21), priming NOD FTOC to GAD65 peptides 524–543 increased
the ivIDDM activity (bold type). Interestingly, this "priming"
protocol did not produce an enhanced ivIDDM effect when using
GAD65246–266 or
GAD65509–528, even though the latter was clearly
an antigenic peptide for the FTOC-derived cells (Fig. 2). Indeed, in
one experiment, both 10 and 100 µg/ml of
GAD65246–266 and 100 µg/ml
GAD65509–528 prevented ivIDDM at 21 days of FP
coculture (italics). This effect, however, was modest and not
reproducible. These data suggest that there can be a parallel between
the ivIDDM system and the diabetogenic response in vivo, because both
can show an increase in disease if T cells are activated to
diabetogenic Ags before exposure to pancreatic tissue.
|
In parallel with the priming experiments detailed above and in
Fig. 3, we wished to determine whether priming and recall to GAD65
peptides could be specifically inhibited by treatment with GAD65
peptides during the initial development of NOD FTOC. These studies are
of particular interest in light of recent reports that propose the use
of peptides to induce protective tolerance to early autoantigens by
treatment of prediabetic individuals (23, 24).
NOD fetal thymi (FT) were placed in standard FTOC, and some cultures
were immediately treated (day = 0) with peptide solutions of
either a mixture of GAD65509–528 and
GAD65524–543, the control PSA peptide at 10
µg/ml, or standard organ culture medium. FTOC was then cultured for
14 days to produce phenotypically mature populations of TCR-bearing T
cells. During the final 48 h of FTOC, as with primed cultures
above (Fig. 3), FTOC were treated with a mixture of GAD65 peptides (10
µg/ml), control peptide (10 µg/ml), or fresh medium. At the end of
this period, the cells were removed from FTOC and assayed for specific
proliferative capacity in response to GAD65 and control peptide
challenge.
A representative experiment is shown in Fig. 4. NOD FTOC that develops in the presence
of GAD65 peptides is no longer reactive to peptides 524–543 and
509–528 of GAD65. In summary of these experiments, the response to
GAD65524–543 (SI = 1.11 ± 0.11,
n = 3) was much less in GAD65 peptide-precultured FTOC
after priming than in cultures that had not been precultured with GAD65
peptide mixtures (SI = 2.64 ± 0.20, n = 5),
or as compared with FTOC challenged with control peptide (SI =
1.1 ± 0.04, p > 0.3, n = 3 for
all peptides). As a positive control of responsiveness
(25), Con A stimulation of identically treated cultures
remained significantly higher than all peptide-challenged FTOC (SI
= 2.23 ± 0.42, p < 0.05, n
= 3).
|
Reports from in vivo studies have indicated that selective tolerance to GAD65 protein prevents the onset of IDDM in NOD mice (11, 12). We wished to determine whether rendering FTOC unresponsive by peptide treatment during the development of mature T cells would ablate ivIDDM.
NOD FT were placed in standard FTOC and then immediately treated with peptide solutions of either a mixture of GAD65 peptides (see Materials and Methods) at a total of 0.1–100 µg/ml or 100 µg/ml control PSA peptide. FTOC was then cultured for 14 days in standard FTOC to produce phenotypically mature populations of TCR-bearing T cells (in the presence of GAD65 and control peptides). At this time, freshly procured NOD FP were added to each of the treated and untreated control FTOC. At 21 days of coculture, supernatants were removed from the cultures and used to determine the effect stimulated and unstimulated FTOC had on insulin production in the FP in coculture.
Fig. 5 shows two experiments in which
there was a dose-dependent reduction in ivIDDM caused by early
"tolerogenic" treatment of FTOC with GAD65 peptides.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We have previously reported on the use of a modified in vitro FTOC system for the study of the development of diabetogenic T cells in NOD mice (18). Here, we wished to determine whether the ability of our in vitro system to decipher how the response to GAD peptides might alter IDDM in NOD mice treated in utero as well as those in vivo systems in which GAD protein was therapeutic, as reported by others. If substantial correlations exist between in vitro activity and disease outcome in vivo, the FTOC system could provide new insights and predictions about therapies to prevent IDDM, and possibly be used for rapid analysis of the efficacy of various T cell-directed Ag treatments.
Our ability to alter the induction of IDDM in vivo with GAD peptides is in agreement with some studies, but not with others. The reason for these different results is unclear, but it is possible that the induction of IDDM resistance with GAD peptides may be more difficult to achieve in adult NOD mice with the dose and route of administration used in some experiments. The ability of the same peptides that delay IDDM in vivo to prevent IDDM onset when given during early thymic development in vitro suggests that these peptides may be used to mitigate disease. From our in utero data and those from other studies in which GAD65 peptides accelerated IDDM in young adult NOD mice (17), we would postulate that protection and acceleration of IDDM are both age and dose dependent. Thus, a fetal NOD mouse given a high dose of GAD65 peptide may very well be protected during the early part of its life when it is using the repertoire established during prenatal development. As an adult, the same NOD mouse that is now more immunologically advanced and that has continued to produce T cells that are no longer under specific tolerogenic pressures may be more resistant to disease prevention with the same peptide regimen. Indeed, such a mouse may become diabetic at an increased rate when given the same peptide therapy, especially if a relatively low dose of peptide were delivered. Dose-dependent tolerance induction to islet Ags has already been reported (29). In the experiments reported here, we have found that injection of single peptide preparations or a mixture of GAD509–528 and GAD524–543 into 15-day gestation pregnant NOD mice can significantly delay IDDM onset in the pups. The mixture of GAD509–528 plus GAD524–543 was considerably better in preventing IDDM than the single peptides given alone, suggesting that treatment with a broader range of epitopes of GAD could alter the course of IDDM, although the mixture of all three GAD peptides did not seem to do any better. In contrast, the same treatment given to young adult NOD mice can either have a marginal inhibitory effect or accelerate disease, depending on the dose of GAD peptides used. The latter results confirm earlier reports suggesting that some GAD treatment regimens can, indeed, induce disease (30). This result is consistent with the clinical data mentioned above.
It is of interest that in the in utero studies the peptides apparently were effectively retained in the pups long enough to alter the production of diabetogenic T cells. Presumably, the delay in the induction and overall incidence of IDDM in mice treated in utero with GAD peptides, rather than complete inhibition of disease, is due to the continued production of new T cells by the thymus after the introduced high dose of peptides have been degraded and lost. It would seem likely that treatment with higher (or multiple) doses of peptide over the entire late gestation and early neonatal period will improve protection against disease. These experiments are currently under way.
We found that NOD FTOC could be induced to respond to a nominal Ag
(GAD65524–543) at a level 
3 times greater
than background proliferation levels. This is of interest because
reports have demonstrated the presence of allogeneic CTL precursors in
FTOC (31, 32) and the ability of FTOC to respond in MLC
and IL-2 production assays (31, 33) but not the ability of
these cells to be specifically reactive to nominal Ag. As compared with
spleen cell preparations from adult mice, a 3-fold proliferation index,
as shown in Fig. 3, is low. However, our results compare quite
favorably with FTOC studies that showed a 2- to 28-fold (usually 3–5
times control) response in allogeneic MLC (33, 34). The
finding that cells recovered from FTOC (that are roughly equivalent in
function and gestational age to 1 wk after birth) are able to respond
to nominal Ag is in agreement with a report that suggests that neonatal
immune populations are responsive to antigenic stimulation
(35). In fact, in our proliferation assay system, 10
µg/ml were used to challenge and restimulate NOD FTOC; a level
one-half to one-tenth that of other reported proliferation assays.
Using greater amounts of peptide Ag (or >2.5 µg/ml Con A) caused
uniformly depressed proliferation index values (not shown).
We found that NOD FTOC could not be effectively primed to the
immunogenic GAD65246–266 vs a control
nonimmunogenic peptide (Fig. 3). The response to
GAD509–528 was also weak. Other work showing
early reactivity to GAD246–266,
GAD509–528, and
GAD524–543 from 3-wk-old NOD mice (11, 12) suggests that these mice have been primed to these peptides
at an early age. The lack of a broad response to GAD peptides in NOD
FTOC may indicate an inherent limitation in the response diversity or
precursor frequency of the cultures during early fetal development. We
chose the 14-day time point because the maximum number of cells in FTOC
are found at roughly 14 days of culture and because longer cultures, to
our present knowledge, do not produce cells with a more mature
phenotype. It is of obvious importance to determine whether responses
to other nominal Ags at similar levels can also be generated by FTOC
cells, as well as to determine which Ag FTOC is capable of responding
to, as compared with similarly aged NOD mice (1 wk old).
However, this report now indicates that 14-day gestation fetal thymus
lobes cultured in FTOC for an additional 14 days are capable of
responding to at least one early GAD Ag (524–543) to a
level of statistical significance and that doses of Ag that are
normally benign to adult T cell populations are inhibitory to
FTOC-derived responder populations.
The need for priming to obtain a proliferative response to GAD peptides by organ culture-derived cells requires consideration as to how the ivIDDM response is achieved. It may be that extrathymic events are critical in the induction of IDDM in NOD mice. Thus, in vivo, an event that causes the damage of islet cells with release of islet cell Ags may allow for the stimulation of GAD-specific T cells that have escaped deletion in the thymus due to poor activation machinery in NOD mice (36). This idea gains support from data that show that the GAD65 peptides that we used have been reported to be presented in a class II-restricted manner (24), downstream of a more primary (possibly class I-restricted) event (37). Such events may be pancreas damage due to occult pancreas infections, or infections with viruses such as coxsackie B4 which may possess epitopes that are identical with some GAD peptides; thus inducing immunoreactivity to GAD by cross-reaction (molecular mimicry) (27, 38, 39, 40). This endogenous priming may also be a result of a generalized inflammatory defect which is adjunct to viral infections (41). In either case, an early event (preceding the generation of a response to GAD65) may lead to up-regulation of processing and presentation of otherwise cryptic peptides. This process of up-regulation of presentation leading to exposure of previously cryptic peptide determinants on a protein has been extensively characterized in other autoimmune systems such as the experimental autoimmune encephalomyelitis model of multiple sclerosis (42, 43, 44). It is of some interest, however, that we were able to generate an ivIDDM response in our system in the present study and in earlier work (18) without overt islet damage in the organ cultured pancreas. The necessary priming of diabetogenic T cells in our system must occur by some process other than acute viral infection or inflammation, possibly by culture-induced release of pancreatic Ags.
We found that the priming of FTOC with GAD peptides in the manner required to induce a response to GAD524–543 resulted in an increase in ivIDDM activity, but an increased response to GAD509–528 did not increase ivIDDM. This result suggests that a peptide that is capable of eliciting a T cell response can cause diabetes in our system. Thus, peptides that have been shown to be antigenic in NOD FTOC (GAD509–528 and GAD524–543) were differentially capable of increasing ivIDDM. This result is important because it shows that peptides that are antigenic for NOD FTOC are not necessarily capable of affecting ivIDDM. These peptides are internal controls for one another in this experiment, which also suggests that the mere activation of T cells by a peptide Ag in this system is not responsible for the increase in ivIDDM activity. In the present study, we used a mixture of GAD peptides to accelerate IDDM in young adult mice, but GAD524–543 alone can accelerate IDDM in these animals (17), suggesting that the ivIDDM system can parallel results obtained in vivo.
However, our in utero data suggest that GAD246–266 was capable of preventing IDDM, even though it was not recognized by FTOC-derived T cells. This peptide could not exacerbate ivIDDM either. Therefore, FTOC is not a completely accurate indicator of the range of responsiveness of NOD mice in vivo. Perhaps, as discussed above, the response to GAD246–266 matures later in the development of the thymus than can be determined in FTOC, preventing its involvement in ivIDDM. However, the apparent persistence of peptide in the pups of mice treated in utero may allow for the induction of unresponsiveness by GAD246–266 in these experiments.
We also found that a relatively low dose (10 µg/ml in vitro) of a
mixture of GAD peptides could mitigate ivIDDM if they were given to NOD
FTOC at the beginning of the culture period, before the development of
mature T cells. Thus, peptides that could induce ivIDDM could also
prevent ivIDDM if they were given under conditions that could prevent
their ability to induce a response (immune tolerance (Figs. 4 and 5)).
This result would be expected if GAD peptides are responsible for some
of the immune recognition of pancreas tissue, given that preculture of
NOD FTOC with the whole pancreas is also capable of preventing ivIDDM
against another challenge with pancreas tissue in organ culture
(18).
The in utero data suggest that mixtures of GAD peptides were the same
as single peptides in preventing IDDM, if the final percentage of
diseased mice is measured. It is interesting, however, that the
induction of disease in mice treated with different combinations of
peptides was different. For the GAD246–266
single peptide-treated mice and the mice treated with all three
peptides, a relatively large number of animals became diabetic at the
same time as the controls, but after 20 wk few mice treated with the
three-peptide mixture became diabetic. The
GAD246–266-treated mice and the
GAD509–528 plus
GAD524–543-treated mice eventually reached the
same overall disease incidence, although the induction of disease in
the latter was delayed. It is possible that a later induction of
GAD246–266 responsiveness in the developing pups
may be responsible for late disease induction in adult progeny (e.g.,
>25 wk in Fig. 2). Animals treated with this peptide may only be
partially tolerant for this response, and those treated with
GAD509–528 and GAD524–543
would not be tolerant at all. These mice would develop disease later,
and this response may account for the delayed disease induction seen in
some of the treated mice. Presumably, the response to
GAD509–528 and GAD524–543
matures earlier in the developing pups, and these responses could be
more easily prevented by treatment with these peptides in utero with
the protocol used in our studies, especially if the peptides were used
together. Apparently, however, mixing all three peptides diluted out
the ability of GAD509–528 and
GAD524–543 to induce tolerance to the early IDDM
response, because this response occurred in the groups of mice treated
with all the peptides. These data are an indication of the complexity
of peptide treatment to prevent IDDM, even with a genetically identical
population of animals. These results suggest that caution should be
exercised in the use of these treatments in a clinical setting.
Other researchers, using in vivo systems, have shown that protection from IDDM by injection of insulin (16, 45, 46) or intranasal administration of GAD (24) involves "immune diversion" from a Th1 to a Th2 response. A similar mechanism has been proposed for the protection of NOD mice by intrathymic injection of islet cells, whole insulin B chain, and whole GAD65 (17). Although our data using the ivIDDM model do not preclude Th1 to Th2 diversion as the mechanism of protection by pretreatment of the FTOC with GAD peptides, the decrease in the ability of the cells from treated FTOC to respond to these peptides in proliferation assays is also consistent with clonal deletion or efficient regulation/anergy of GAD-responsive cells. Regardless, the loss of these cells through deletion caused by therapeutic high peptide concentrations during T cell development may prove an effective method for prevention of IDDM, should the critical initiating Ags be found.
Overall, our data suggest that GAD is an important target Ag in IDDM and that it may be a trigger or required component of the T cell response cascade that results in IDDM. Work with other islet-associated Ags such as insulin (45, 46) or insulin peptides (16), and heat shock protein 60 peptides (47) to prevent IDDM are also promising. Recent clinical trials (9), however, support work shown here and elsewhere (17, 48) that these Ags administered to adult NOD later in development, when peripheral regulatory responses have matured, can exacerbate disease. The present data showing the efficacy of inducing tolerance to GAD peptides during fetal development in FTOC or in utero, suggest that IDDM may be prevented with the appropriate immunotherapy given early in T cell development. Our results also suggest that the ivIDDM model may be useful in the rapid screening of candidate Ags in various combinations for peptide immunotherapy to prevent IDDM. Promising combinations of peptides could then be tested in the longer term in vivo models to establish the best time to treat with these Ags before use in clinical trials.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Dominick DeLuca, Department of Microbiology and Immunology, Life Sciences North, Room 648, University of Arizona, Tucson, AZ 85724. E-mail address: deluca{at}u.arizona.edu
3 Abbreviations used in this paper: IDDM, insulin-dependent diabetes mellitus; ivIDDM, in vitro IDDM; FP, fetal pancreas; FT, fetal thymus; FTOC, fetal thymus organ culture; GAD, glutamic acid decarboxylase; NOD, nonobese diabetic; SI, stimulation index; PSA, prostate-specific antigen; HEL, hen egg lysozyme; MBP, myelin basic protein.
Received for publication December 22, 2000. Accepted for publication April 24, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-cell function in recent-onset Type I diabetes (the IMDIAB VII). Diabetologia 43:1000.[Medline]
-cell antigens. Diabetes 46:1975.[Abstract]
and prostaglandin-E. Am. J. Pathol. 147:1193.[Abstract]
transcription. J. Clin. Invest. 95:628.
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
 |
|
 R. K. Gregg, R. Jain, S. J. Schoenleber, R. Divekar, J. J. Bell, H.-H. Lee, P. Yu, and H. Zaghouani A Sudden Decline in Active Membrane-Bound TGF-{beta} Impairs Both T Regulatory Cell Function and Protection against Autoimmune Diabetes J. Immunol., December 15, 2004; 173(12): 7308 - 7316. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Y. Arnold, A. R. Burton, and D. A. A. Vignali Diabetes Incidence Is Unaltered in Glutamate Decarboxylase 65-Specific TCR Retrogenic Nonobese Diabetic Mice: Generation by Retroviral-Mediated Stem Cell Gene Transfer J. Immunol., September 1, 2004; 173(5): 3103 - 3111. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||
) or high (50.0 µg/mouse of each
peptide, 
) dose or 50 µg/mouse of an irrelevant control peptide
(PSA95–109, 
). Those mice treated with a low dose of
GAD peptides (n = 5) developed IDDM at an
accelerated rate as compared with control mice (p =
0.0466). Despite the appearance that high dose treatment had protective
effects on treated NOD mice (n = 5), the curve was
not found to be statistically significant (p =
0.0845) when compared with the PSA peptide-treated control.
), or a single control peptide (HEL (
) or MBP (
)).
Offspring female mice (the number in each treatment group is shown in
parentheses in the key inset) were monitored for the development of
diabetes. IDDM in offspring of mice treated with
GAD65246–266 (n = 14) was
significantly inhibited as compared with mice treated with a single
control peptide (HEL, p = 0.035) or the control
peptide mixture (MBP plus PSA, p = 0.014). IDDM
induction in offspring of mice treated with a mixture of
GAD509–528 and GAD524–543
(n = 10) were also significantly inhibited as well
as delayed as compared with mice treated with control peptides (HEL,
p = 0.011 and MBP plus PSA,
p = 0.003). However, offspring from females treated
with GAD509–528(p = 0.203)
and GAD524–543 (p = 0.388) were not
significantly inhibited in their IDDM induction compared with the HEL
single peptide control. Offspring from mice treated with a mixture of
GAD65 peptides 246–266, 509–528, and 524–543 (n
= 9) were also not significantly different from the MBP plus PSA
(p = 0.226) control peptide mixture or the HEL
single peptide control (p = 0.448).
