| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Microbiology and Infectious Diseases, Faculty of Medicine, University of Calgary, Alberta, Canada;
Department of Pathology and Laboratory Medicine, University of British Columbia and British Columbia’s Children’s Hospital, Vancouver, British Columbia, Canada; and
Department of Medical Microbiology and Immunology, University of Alberta, Edmonton, Alberta, Canada
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The putative cross-reactive T cells capable of causing autoimmune disease via molecular mimicry must somehow evade normal immune tolerance mechanisms. Engagement of self-peptide/MHC complexes by autoreactive thymocytes on bone marrow-derived APCs usually leads to thymocyte death or to functional unresponsiveness to subsequent encounters with Ag (31). Autoreactive T cells, particularly those recognizing autoantigens that are not ferried to (or that are not expressed in) the thymus can escape central tolerance, but these cells are usually dealt with by mechanisms of peripheral tolerance or ignorance (31). The objective of these processes is that the peripheral immune system is almost exclusively populated by T cells that can recognize foreign, but not self-Ags in the context of self-MHC molecules. Although intrinsic defects in T cell tolerance may underlie the genetic susceptibility to certain spontaneous autoimmune diseases (32, 33), there is no evidence to indicate that autoimmunity arising from molecular mimicry requires faulty tolerogenic mechanisms. Because the mechanisms of tolerance preferentially target T cells recognizing self-peptide/MHC with high affinity/avidity (31), it would be logical to assume that the foreign Ag-reactive T cells that trigger autoimmunity recognize self-peptide/MHC complexes with an avidity that falls below the threshold for tolerance. However, because we have shown that the pathogenic activity of autoreactive T cells parallels their avidity for peptide/MHC (34), it is difficult to imagine how T cells with low avidity for self might drive an autoimmune disease.
Studies of autoimmunity based on molecular mimicry between self and foreign Ags as a potential mechanism in the initiation of autoimmunity have almost exclusively used activated T cell clones (or hybridomas) or resting TCR-transgenic T cells, but not both (30, 35, 36). Although activated T cells are more susceptible to TCR triggering by low-affinity ligands (37, 38, 39) and display higher functional avidities for agonists than naive T cells (40), naive and activated T cells are thought to recognize the same peptide ligands. Here we have tested the validity of this assumption by investigating whether peptide ligands that cannot trigger functional responses on naive T cells (i.e., nonagonistic peptides) can nevertheless do so on their activated progeny. We reasoned that, if this idea were true, then foreign Ag-specific T cells expressing potentially self-reactive TCRs might be able to escape tolerance by ignoring self-peptide/MHC during development. Our experimental approach was to examine the degree of degeneracy in peptide recognition by naive vs activated CD8+ T cells isolated from recombination activating gene-2-deficient nonobese diabetic (NOD)3 mice expressing a transgenic, highly diabetogenic, H-2Kd-restricted TCR (8.3-TCR) (8.3-NOD.RAG-2-/-) (41). This TCR recognizes the peptides NRP (KYNKANWFL) and NRP-A7 (KYNKANAFL) in the context of H-2Kd (42), and is representative of TCRs expressed by the CD8+ T cells that, upon avidity maturation, drive the progression of islet inflammation to diabetes in NOD mice (34). In this work we found that 1) in vitro activation of 8.3-TCR-transgenic (monoreceptor) T cells resulted in a dramatic expansion of the T cells’ antigenic repertoire; 2) the NRP-reactive T cell population arising spontaneously in the pancreatic islets of wild-type NOD mice displayed a similar degree of antigenic promiscuity as did in vitro activated 8.3-CD8+ T cells; and 3) the antigenic promiscuity of the wild-type, intraislet NRP-reactive CD8+ T cell population expanded as the avidity of the population for NRP-A7/MHC increased with age. These observations demonstrate that, upon activation and avidity maturation, T cell populations can mount functional responses against antigenic peptides that they previously ignored.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Mice (8.3-NOD.RAG-2-/-; 6- to 8-wk old)
expressing the TCR
rearrangements of the
H-2Kd-restricted cell-reactive
CD8+ T cell clone NY8.3 have been described
(41). NOD mice were purchased from Taconic Farms
(Germantown, NY). RMA-SKd cells were obtained
from B. Wipke and M. Bevan (University of Washington, Seattle, WA). The
GK1.5 (anti-CD4) hybridoma was obtained from the American Type
Culture Collection (ATCC, Manassas, VA). Anti-Lyt-2 (CD8) (53-6.7),
anti-L3T4 (IM7), anti-V8.1/8.2 (MR5-2), anti-CD44-biotin
(IM7), anti-CD69-biotin (H1.2F3),
anti-H-2Kd (SF1-1.1), and
anti-H-2Db (KH95) mAbs were obtained from
PharMingen (San Diego, CA).
Peptides/database searches
The peptide libraries were prepared using multipin synthesis
technology and standard F-moc chemistry (Chiron Technologies,
San Diego, CA). Specific single custom peptides were purified through
reversed phase HPLC to >80% purity and sequenced by ion spray mass
spectrometry. Peptides were resuspended in 0.1 M HEPES (Sigma, St.
Louis, MO) in 40% acetonitrile (Fisher Scientific, Fair Lawn, NJ) at
pH 7.4. The first screen was performed using the following NRP-based
dipeptide libraries: DE1 ((D or E)YNKANWFL)), NQ1, GS1, FY1, VT1,
AL1, WI1, DE3, KR3, GS3, FY3, VT3, AL3, WI3, DE4, NQ4, GS4, FY4, VT4,
AL4, WI4, DE5, KR5, NQ5, GS5, FY5, VT5, WI5, DE6, KR6, GS6, FY6, VT6,
AL6, WI6, DE7, KR7, NQ7, GS7, FY7, VT7, AL7, DE8, KR8, NQ8, GS8, VT8,
AL8, and WI8. The single amino acid variants of NRP that were tested in
subsequent experiments are listed in Fig. 2
. Database searches were performed using the PIR database and a
psearch algorithm (version 1.3) using the search string (supertope):
(KANQGSFYVTLWIMC)(YF)(NQ
SFYVLIMC)(KIMC)(ANGTMC)(NAGSFLMC)(WNQGSFYVTALMC)(FMC)(LIV).
|
Naive or NRP-differentiated (see below) splenic
CD8+ T cells from
8.3-NOD.RAG-2-/- mice (2 x
104/well) were incubated in duplicate with
peptide-pulsed (0.01, 0.1, and 1 µM), 
-irradiated (3000 rad) NOD
splenocytes (105/well) for 3 days at 37°C in
5% CO2. Cultures were pulsed with 1 µCi of
[3H]thymidine during the last 18 h of
culture and harvested.
Cytokine secretion
Naive or NRP-differentiated splenic CD8+ T
cells from 8.3-NOD.RAG-2-/- mice (2 x
104/well) were incubated with peptide-pulsed
(0.001, 0.01, 0.1, 1, and/or 10 µM), -irradiated NOD splenocytes
(105/well) in 96-well plates for 48 h at
37°C. Islet-derived T cell lines or FACS-sorted tetramer-positive and
-negative CD8+ T cells (adjusted at 2
x 104 CD8+ T cells/well)
were tested the same way, but using 1 µM of peptide. The supernatants
(100 µl/well) were assayed for IL-2, IL-4, and/or IFN- content by
ELISA using commercially available kits (Genzyme, Cambridge,
MA).
Generation of islet-derived and NRP-differentiated splenic CD8+ T cells
Naive spleen cells from 8.3-NOD.RAG-2-/- mice were adjusted to 104 CD8+ T cells/100 µl of complete medium (CM; RPMI 1640 medium containing 10% FBS), stimulated with NRP-pulsed (1 µM) NOD splenocytes for 3 days, and expanded in the presence of rIL-2 for 4–6 days. Islet-derived CD8+ T cells from 8.3-NOD.RAG-2-/- and nontransgenic NOD mice were generated by culturing pancreatic islets in CM containing 0.5 U/ml Takeda recombinant human IL-2 (rhIL-2; 10–50 islets/well in 24-well plates). The in vivo activated, IL-2R+ lymphocytes migrating from islets into the culture medium were further expanded in rhIL-2-containing medium for an additional 3–6 days. T cells were used in functional assays within 6–9 days of islet isolation.
Tetramer staining and cell sorting
Tetramers were prepared as described by Altman et al. (43). Hand-picked pancreatic islets pooled from 15–40 mice/age group/experiment (9- or 20-wk-old NOD mice) were cultured in CM supplemented with 0.5 U/ml Takeda rhIL-2 for 7–9 days. T cells (106 per 20 µl) were then stained for 45 min at room temperature in 20 µl of wash medium (0.2% sodium bicarbonate, 0.1% sodium azide, and 2% FBS in RPMI 1640) containing anti-CD8-FITC (clone YTS169.4; 0.5 µg) and NRP-A7/Kd tetramer (85.5 nM). After washing, the cells were resuspended in wash medium and sorted with a FACStar flow cytometer (BD Biosciences, Mountain View, CA). Purity of sorting was 87–93% for tetramer-positive cells and 95–97% for tetramer-negative cells.
51Cr-release assays
RMA-SKd cells (preincubated at 26°C overnight) were labeled with [51Cr]sodium chromate (DuPont-NEN, Boston, MA) for 2 h, washed, resuspended at 105 cells/ml in RPMI 1640 containing 0.25% BSA, seeded at 104 cells/100 µl/well, pulsed with peptides (0.1–1 µM) for 1 h at 37°C, and used as target cells in 51Cr-release assays. Effector cells (NRP-differentiated splenic CD8+ T cells; 100 µl) were added to each well in duplicate at a 1:10 target-effector ratio. Cultures of RMA-SKd cells with peptides but no T cells were used as controls to confirm that the peptides were not cytotoxic. Plain medium or 1% Triton X-100 was added to sets of target cells for examination of spontaneous and total cell lysis, respectively. The plates were incubated at 37°C for 8 h, and the supernatants were collected for determination of specific 51Cr release: (% lysis = 100 x (test cpm - spontaneous cpm)/(total cpm - spontaneous cpm)). Values >20% release were considered significant.
H-2Kd-stabilization assay and H-2Kd/peptide association rates
RMA-SKd cells that had been cultured overnight at 26°C were seeded, in quadruplicate, at 104 cells/well in 96-well plates, pulsed with peptides in RPMI 1640, 0.25% BSA for 1 h at 26°C, incubated at 37°C for 3 h, washed, and stained with anti-H-2Kd-FITC or anti-H-2Db-FITC, and the mean fluorescence intensity (mfi) for MHC class I expression was analyzed by flow cytometry (42). Controls used included tum (H-2Kd-binder), lymphocytic choriomeningitis virus-GP33 (H-2Db-binder), and no peptide. The Kd was measured by repeating the experiments described above except using different concentrations of peptides (10, 1, 0.1, 0.01, and 0.001 µM). The Kd values were calculated as the concentration of peptide required to rescue 50% of the H-2Kd molecules on RMA-SKd cells (100% at 10 µM).
Statistical analyses
Data were compared by Mann-Whitney U test or
2.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To determine which amino acid substitutions could be introduced at
each position of the NRP sequence without inducing a major loss of
functional activity, we first tested the agonistic activity of
NRP-based combinatorial dipeptide libraries prepared in the positional
scanning format. To generate the libraries, each of the seven nonanchor
positions of NRP (P1 and P3-P8) was substituted with groups of two
related amino acids (DE, KR, NQ, GS, FY, VT, AL, and WI). The libraries
were tested for their ability to induce proliferation of, and IFN-,
IL-2, and IL-4 secretion by, naive splenic CD8+ T
cells from 6- to 8-wk-old 8.3-NOD.RAG-2-/-,
which bear a monoclonal TCR repertoire (41). Negative and
positive controls included the H-2Kd-binding
peptide Tum (KYQAVTTTL) and NRP or NRP-A7, respectively.
8.3-NOD.RAG-/- mice do not usually develop
islet inflammation until after 10 wk of age, and most, if not all, of
their peripheral CD8+ T cells have a naive
phenotype (CD11alow,
CD44low, CD69-,
CD62Lhigh) (Ref. 41 and data not
shown). In general, there was a good correlation between the magnitude
of the proliferative and cytokine (IL-2 and IFN-) secretory
responses induced by each library (Fig. 1; none of the peptide libraries induced IL-4 secretion; data not shown).
Interestingly, naive 8.3-CD8+ T cells did not
secrete IL-2 in response to 12 of the 24 libraries that elicited
IFN- secretion (Fig. 1), and each of the remaining 12 libraries
induced significantly more IFN- than IL-2 (Fig. 1). The sequences of
dipeptide libraries that elicited functional responses from
8.3-CD8+ T cells (
10% of the proliferation
and/or cytokine secretion induced by NRP but not TUM in at least two of
three experiments; see Fig. 1) were selected for deconvolution. This
was performed by determining which of the 52 possible peptides deduced
from the dipeptide library screen could induce the
8.3-CD8+ T cell responses mentioned above, over a
range of concentrations (1, 0.1, and 0.01 µM) (Fig. 2
and data not
shown). This set of experiments also included single amino acid
variants of NRP that carried substitutions (other than cysteine and
methionine) that were not tested in the dipeptide library screen
(NRP-R1, NRP-H1, NRP-Q3, NRP-R4, NRP-H4, NRP-H5, NRP-A6, NRP-H6,
NRP-H7, NRP-Y8, and NRP-H8) as well as two negative control peptides
deduced from a negative dipeptide library (NRP-W8 and NRP-L8) (Fig. 2).
Peptides NRP-L5, NRP-S5, NRP-Q6, and NRP-I7 were largely insoluble and
could not be tested. As shown in Fig. 2, 42 of the 65 single amino acid
variants of NRP that were tested in these experiments triggered
proliferation of, and/or cytokine (IL-2 and/or IFN-) secretion by,
naive 8.3-CD8+ T cells. As expected on the basis
of previous studies in other systems (14), substitutions
of the two primary TCR contact residues K4 and F8 for any other
residue, except I at position 4, resulted in a complete loss of
activity (Fig. 2). Thirty-six of these 42 single amino acid variants of
NRP were only efficient at inducing some, but not all, of the tested
responses (25 peptides elicited weak proliferative and/or cytokine
responses, whereas 11 peptides were unable to elicit proliferation of,
or IL-2 and/or IFN- secretion by, 8.3-CD8+ T
cells) (Fig. 2). The remaining 23 of the 65 NRP analogs listed in Fig. 2 did not elicit any responses from naive
8.3-CD8+ T cells at 1, 0.1, or 0.01 µM (Fig. 2), and only one of these (NRP-D6) was able to induce the proliferation
of naive 8.3-CD8+ T cells at 10 µM (data not
shown). Therefore, naive 8.3-CD8+ T cells can
only mount full agonistic responses against a relatively small number
of NRP homologues.
|
We next investigated whether any of 20 NRP homologues that lacked
agonistic activity on naive 8.3-CD8+ T cells
(herein referred to as "nonagonists" for naive
8.3-CD8+ T cells) could trigger functional
responses of in vitro activated 8.3-CD8+ T cells
(peptides highlighted in gray in Fig. 2). This was performed by testing
the ability of NRP-differentiated 8.3-CD8+ T
cells to secrete IFN- in response to peptide-pulsed NOD splenocytes.
Activated CD8+ T cells turn off IFN-
transcription and translation immediately upon Ag withdrawal
(44). Therefore, measurement of IFN- secretion allows a
comparison of the inherent ability of naive vs activated
CD8+ T cells to respond to the same peptides
without the potentially confounding contribution of assay sensitivity
to outcome. As shown in Fig. 3, in vitro differentiated 8.3-CD8+ T cells
secreted high levels of IFN- in response to 10 of the 20
nonagonistic peptides that were tested. These 10 peptides could also
efficiently reconstitute H-2Kd cDNA-transfected
RMA-S cells (RMA-SKd) for lysis by differentiated
8.3-CD8+ T cells, in some cases even at low
concentrations (0.1 µM) and even more efficiently than agonistic
peptides (data not shown). Therefore, these results indicated that
activated 8.3-CD8+ T cells exhibit a greater
degree of peptide promiscuity than their naive progenitors.
|
To investigate whether the increased antigenic promiscuity
of the 8.3-TCR on activated vs naive T cells also applied to NATs, we
searched protein databases for NRP mimics. This was performed using a
supertope that included residues that were tolerated by NRP (Fig. 1)
plus M or C at any nonanchor position, F or Y at the first anchor
position (P2), and L, I, or V at the second anchor position (P9). This
analysis resulted in the identification of 92 potential mimics of NRP
(data not shown). All these peptides were tested for their ability to
bind to H-2Kd molecules on
RMA-SKd cells, to elicit cytokine secretion by,
and proliferation of, naive 8.3-CD8+ T cells, and
to enable the cytotoxic activity of NRP-differentiated
8.3-CD8+ T cells. Thirty-two of these 92 peptides
did not bind to H-2Kd molecules and had no
functional activity on 8.3-CD8+ T cells. The
remaining 60 peptides bound H-2Kd molecules or
were recognized by NRP-differentiated 8.3-CD8+ T
cells in cytotoxicity assays (>20% 51Cr
release; data not shown), and a few of them elicited a functional
response from naive 8.3-CD8+ T cells (Fig. 4). These 60 peptides were classified into three broad groups based on
their ability to elicit functional responses from naive
8.3-CD8+ T cells: 1) peptides that did not elicit
any responses (designated as "nonagonists"; n =
54); 2) peptides that elicited weak responses (designated as "partial
agonists"; n = 3); and 3) peptides that elicited
strong responses (designated as "full agonists"; n
= 3) (labeled -, +/-, or +, respectively, in Fig. 4). It should be
noted that the designations agonistic and nonagonistic are arbitrary
and refer exclusively to the ability of the different peptides to
induce functional T cell responses from naive
8.3-CD8+ T cells; as shown below, some
nonagonistic mimics can induce functional responses from
differentiated 8.3-CD8+ T cells. As expected on
the basis of the data shown in Fig. 1, presence of the two TCR contact
residues of NRP (K4 and F8) was necessary for full agonism (see Fig. 4). However, their presence was insufficient because only 6 of 24
peptides carrying these two residues had agonistic activity on naive
8.3-CD8+ T cells (Fig. 4). Interestingly, the
number of agonistic residues carried by individual peptides was not a
good marker of the degree of agonism because poorly agonistic
substitutions (in NRP) were present in mimics with full agonistic
activity. However, the number of nonanchor residues within a peptide
that were shared with NRP was a significant indicator of agonism
(1.9 ± 0.1 for nonagonistic peptides vs 3.5 ± 0.3 for
agonistic peptides, p < 0.0005). When taken together,
these data indicated that 1) the agonistic activity of NATs was a
function of their ability to contact the 8.3-TCR, rather than of their
affinity for MHC; and 2) only a very small fraction of all the NATs
carrying the two primary TCR contact residues of NRP could trigger
naive 8.3-CD8+ T cells.
|
We next compared the ability of NRP-A7 (an alanine mutant of NRP
with superior agonistic properties), the six naturally occurring
agonistic mimics, and the nonagonistic mimics described above, to
trigger IFN- and IL-2 secretion by naive and NRP-differentiated
8.3-CD8+ T cells. In general, agonistic mimics
induced more IFN- and IL-2 secretion from differentiated
8.3-CD8+ T cells than from naive
8.3-CD8+ T cells (Fig. 5
A). Nonagonistic mimics that lacked both K4 and F8 were
unable to trigger IFN- and IL-2 secretion from NRP-differentiated
8.3-CD8+ T cells (data not shown), as was the
case for naive 8.3-CD8+ T cells (see above).
However, 12 of the 31 nonagonistic NRP mimics that carried TCR-contact
residues K4 and F8 of NRP (comprising the 11 naturally occurring
homologues and the 20 single amino acid variants that are highlighted
in gray in Figs. 1 and 4) were able to trigger IFN- secretion by
NRP-activated 8.3-CD8+ T cells (Fig. 5A,
upper panel). Nine of these 12 nonagonistic mimics were also able
to elicit IL-2 secretion from differentiated
8.3-CD8+ T cells (Fig. 5A, lower
panel). Surprisingly, some of these peptides were at least as
efficient as NRP-A7 or some of the six naturally occurring agonists at
eliciting IFN- and/or IL-2 secretion by these T cells (Fig. 5A). Note that most of these peptides were unable to trigger
IFN- secretion by naive 8.3-CD8+ T cells, even
at high concentrations (see below and Fig. 6). Kinetic studies indicated that, as seen with single amino acid
variants of NRP, there was no correlation between the on-rates
(Kd) of peptide binding to MHC (Fig. 4) and the degree or type of response elicited by the different
peptides on differentiated 8.3-CD8+ T cells (Fig. 5A). This indicated that the ability of nonagonistic NRP
mimics to elicit cytokine secretion by differentiated
CD8+ T cells is stringently dependent on their
ability to contact the TCR. Taken together, these results confirmed
what the results with single amino acid variants of NRP had indicated,
namely, that differentiated 8.3-CD8+ T cells
respond efficiently to many more peptides than their naive
precursors.
|
|
B,
these in vivo activated CD8+ T cells responded to
even more nonagonistic mimics than NRP-differentiated
8.3-CD8+ T cells, suggesting that the quality
and/or degree of activation modulates the degree of antigenic
promiscuity exhibited by an activated T cell. All the nonagonistic
peptides capable of triggering IFN- secretion from differentiated
8.3-CD8+ T cells could also efficiently elicit
their cytotoxic activity (data not shown).
To ascertain whether the increased antigenic promiscuity of
differentiated 8.3-CD8+ T cells was the result of
a reduction in the threshold of TCR triggering upon T cell activation,
we performed titration experiments with different concentrations of
peptides. Naive 8.3-CD8+ T cells did not secrete,
or secreted only very low levels of, IFN- when challenged with high
concentrations of nonagonistic NRP mimics (Fig. 6, upper left
panel). Surprisingly, differentiated
8.3-CD8+ T cells secreted as much or more IFN-
in response to low concentrations of some nonagonistic peptides (i.e.,
NRP-W3) than in response to similar concentrations of NRP-A7 or
agonistic mimics (i.e., NAT-47) (Fig. 6). This indicated that the
ability of nonagonistic peptides to trigger differentiated
8.3-CD8+ T cells did not result from a
concentration-dependent increase in the sensitivity of differentiated
vs naive T cells to TCR ligation, but rather from an ability of
differentiated T cells to respond efficiently to low concentrations of
poor ligands of their TCRs.
Responsiveness of NRP/NRP-A7-reactive CD8+ T cells propagated from islets of prediabetic NOD mice
It was possible that the reactivity of
8.3-CD8+ T cells for nonagonistic mimics of NRP
was a peculiarity of the 8.3-TCR. To rule this out, we investigated
whether these peptides could also induce functional responses from the
in vivo activated NRP/NRP-A7-reactive CD8+ T
cells that are spontaneously recruited into pancreatic islets in
nontransgenic NOD mice (34). Islets from nine different
nondiabetic 9-wk-old NOD mice were cultured in the presence of rIL-2
for 6–9 days to selectively expand IL-2R-positive T cells, which
contain a significant percentage of NRP/NRP-A7-reactive
CD8+ T cells (34). These nine T cell
lines were tested for their ability to secrete IFN- in response to
12 naturally occurring nonagonistic NRP mimics carrying K4 and F8 and
the six naturally occurring agonistic NRP mimics (Fig. 7). Controls included Tum (negative control), NRP-A7 (positive control),
and INS (an insulin-derived peptide recognized by islet-associated T
cells from young NOD mice; Ref. 45). We used NRP-A7
instead of NRP as positive control peptide because NRP-reactive
CD8+ T cells from NOD mice bind
NRP-A7/Kd tetramers with higher
avidity than NRP/Kd tetramers
(34). As shown in Fig. 7, there was a high degree of
variability among lines in terms of the number of peptides they could
respond to (positive responses highlighted). However, six of the nine
lines responded efficiently to one or more nonagonists, and in some
cases the response against nonagonistic mimics was more vigorous than
the response against agonistic mimics. This demonstrated that, like in
vitro activated 8.3-CD8+ T cells, in vivo
activated NRP-reactive CD8+ T cells propagated
from islets of nontransgenic NOD mice can also respond to nonagonistic
mimics of NRP.
|
Progression of pancreatic islet inflammation to overt diabetes in
the NOD mouse overlaps with the "avidity maturation" of the
NRP/NRP-A7-reactive CD8+ T cell population; as
prediabetic NOD mice age (from 9 to 20 wk), their in vivo activated,
islet-associated CD8+ T cells contain increasing
numbers of NRP-A7-reactive cells, and these cells bind
NRP-A7/Kd tetramers with increased avidity and
longer half-lives (34). We reasoned that this increase in
the avidity of in vivo activated NRP/NRP-A7-reactive
CD8+ T cells for peptide/MHC might allow the T
cells to respond to lower affinity ligands of their TCRs. If that were
true, the avidity maturation of the NRP/NRP-A7-reactive
CD8+ T cell population would be paralleled by an
increase in the antigenic promiscuity of the T cells. This hypothesis
was tested by investigating whether activated
CD8+ T cells propagated from islets of 13
different 20-wk-old nondiabetic NOD mice could respond to more
nonagonistic NRP mimics than the activated CD8+ T
cells propagated from islets of 9-wk-old NOD mice. Note that although
lines derived from 20-wk-old mice have an overall higher avidity for
NRP-A7/Kd than cells from younger animals, they
contain percentages of NRP-A7/Kd
tetramer-positive T cells that are similar to those seen in lines
derived from 9-wk-old mice (34). As shown in Fig. 7, CD8+ T cells propagated from 20-wk-old mice
responded to significantly more nonagonistic (but not agonistic) NRP
mimics than CD8+ T cells propagated from 9-wk-old
mice (approximately seven vs three mimics; p < 0.019).
Furthermore, cells from 20-wk-old mice secreted significantly higher
levels of IFN- in response to nonagonistic (n = 6 of
12) and agonistic (n = 1 of 6) mimics of NRP (but not
NRP-A7, as expected) than cells from 9-wk-old mice (Fig. 7). Staining
of islet-derived T cells from 9- and 20-wk-old NOD mice with mAbs
specific for activation markers confirmed that these differences in the
ability of cells from 9- and 20-wk-old NOD mice to respond to
nonagonistic mimics of NRP were not the result of differences in the
percentage of tetramer-positive cells that had undergone activation in
vivo. As shown in Fig. 8, the NRP-A7/Kd tetramer-positive
CD8+ T cells that were contained within these
lines expressed levels of CD44 and CD69 similar to those expressed by
NRP-differentiated 8.3-CD8+ T cells, and
significantly higher than those expressed by naive
8.3-CD8+ T cells.
|
in response to these peptides. Unlike the
tetramer-positive cells of 9- and 20-wk-old NOD mice (
87–93%
purity), the tetramer-negative cells from 20-wk-old NOD mice (97%
purity) did not mount significant responses against NRP-A7, the 12
nonagonistic mimics (data not shown), or the six agonistic mimics (Fig. 9A). This indicated that nearly all the NRP-A7-reactive T
cells of 20-wk-old NOD mice were contained within the
NRP-A7/Kd tetramer-positive subpopulation. In
contrast, the tetramer-negative cells from 9-wk-old NOD mice (96%
purity) secreted significant levels of IFN- in response to NRP-A7
and three of the six agonistic mimics of NRP (Fig. 9A). This
indicated that at least some of the NRP-A7-reactive cells of 9-wk-old
(but not 20-wk-old) mice do not bind NRP-A7/Kd
tetramers, or do so weakly (i.e., with low avidity), as we had
previously shown (34). Finally, we compared the ability of
NRP-A7/Kd tetramer-positive
CD8+ T cells from 9- vs 20-wk-old NOD mice to
secrete IFN- in response to NRP-A7 and the seven mimics of NRP that
elicited quantitatively different responses from nonsorted lines (i.e.,
the peptides highlighted in dark in Fig. 7). As shown in Fig. 9B, the tetramer-positive CD8+ T cells
of 20-wk-old NOD mice produced higher levels of IFN- than the
tetramer-positive CD8+ T cells of 9-wk-old NOD
mice upon stimulation with six of these mimics. When taken together,
these results demonstrate that the ability of polyclonal, monospecific
T cell populations to mount functional responses against nonagonistic
mimics of a cognate autoantigen is a function not only of their
activation state, but also of their overall avidity for
peptide/MHC.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The systematic replacement of every residue of NRP by any other amino acid revealed that the TCR contact residues K4 and F8 could essentially not be replaced (with one exception) without abrogating recognition by 8.3-CD8+ T cells, as seen with other TCRs (14). The degree of tolerance for amino acid replacements was greatest at positions 1 and 7 and decreased toward the TCR contact residue K at position 4. Because amino acid substitutions at positions apparently not involved in MHC-binding and TCR contact can have dramatic consequences on T cell activation (5), we suspect that substitutions near K4 affect the ability of K4 to contact the 8.3-TCR. Also worth noting is the fact that replacement of the tryptophan (W) at position 7 by a number of small, neutral, or polar residues increased the agonistic properties of the resulting NRP analogs, suggesting enhanced availability of the phenylalanine at P8 for binding to the TCR.
A search of protein databases using a supertope resulted in the
identification of 92 potentially cross-reactive peptides, but only six
of these peptides were able to elicit significant proliferative and/or
cytokine responses by naive 8.3-CD8+ T cells.
This indicated that combinations of amino acids that were tolerated
individually within the supertope did not have the same functional
activity when combined in a single peptide, even when some of the
individual residues had superagonistic activity in the context of the
NRP sequence. This observation is at odds with the results of most,
albeit not all, studies with CD4+ T cell clones
or hybridomas, where recognition of hypermutated peptides could be
predicted on the basis of the effects of single amino acid
substitutions (12, 23, 46, 47). Surprisingly, studies with
activated 8.3-CD8+ T cells revealed that the
8.3-TCR has a much more extensive degree of degeneracy in peptide
recognition when expressed on activated cells than on their naive
counterparts. Ex vivo and in vivo differentiated
8.3-CD8+ T cells mounted cytokine responses
(IFN- and IL-2 secretion) against as many as 21 peptides that could
not elicit any responses from unstimulated splenic
8.3-CD8+ T cells. Because the supertope that we
used to search for naturally occurring mimics was based on naive
8.3-CD8+ T cell responses, the promiscuity of
activated 8.3-CD8+ T cells is likely to be
greater than what this study suggests. In fact, experiments with a
mouse mitochondrial peptide that was excluded by the database searches
as a possible mimic (AFFKAMLFM) have indicated that this peptide can
also trigger activated 8.3-CD8+ T cells (J.
F. Elliott and P. Santamaria, unpublished observations).
These peptides that can activate differentiated but not naive CD8+ T cells ("secondary agonists") are probably bound with too low an affinity by the 8.3-TCR to elicit the activation of naive T cells. Importantly, however, some of these hypothetical low affinity interactions can trigger functional responses on activated T cells at concentrations at which some primary agonists have no or very low activity. This suggests that the reduction in the threshold for triggering that a T cell undergoes upon activation (48) renders the T cell partially unable to distinguish between low and high affinity ligands of its TCR without necessarily increasing its responsiveness to the latter. This discovery suggests that a foreign Ag-reactive T cell may become self-reactive upon T cell activation. In this respect, this observation provides an explanation for the results of a recent study showing the existence of molecular mimicry between human cytochrome P450 and a hepatitis C virus epitope at the effector cell level. In this system, the self Ag did not have agonistic properties on naive peripheral T cells, but efficiently elicited the cytotoxicity of differentiated cytotoxic T cells in a 51Cr-release assay (28).
What could possibly be the mechanisms that underlie this phenomenon? The stimulatory potency of peptides for a single TCR is related to a number of factors, including their affinity for the MHC restricting element, the affinity of the corresponding peptide/MHC complexes for the cognate TCR, and the half lives (or Kd) of the interactions (49). Our data has shown that primary and secondary agonists do not differ in their affinity for MHC binding and that the only mimics of NRP that can function as secondary agonists carry two TCR-contact residues of NRP (K4 and F8). In addition, NATs with primary and secondary agonistic activity differed in the number of nonanchor residues that they shared with NRP (3.5 ± 0.3 for primary agonists vs 2.6 ± 0.3 for secondary agonists). Because the ability of NOD islet-derived T cell lines to respond to secondary agonists increases with their avidity for NRP/MHC, these data suggest that the ability of activated T cells to respond vigorously to secondary agonists is related to the formation of the immunological synapse upon T cell activation. T cell activation induces the oligomerization of TCRs and associated coreceptors into supramolecular clusters by a serial triggering mechanism (50, 51, 52). It has been shown that the number of peptide/MHC complexes in the clusters correlates with the half-life of the TCR-peptide/MHC interaction, such that low concentrations of agonists and low affinity ligands are inefficient at inducing cluster formation and stable signaling (51). The efficiency of TCR signaling in the clusters also depends on the ability of peptide/MHC complexes to ligate the TCR long enough to trigger the recruitment and activation of ZAP-70 by p56lck, the activation of the lipid raft-associated adaptor linker for activation of T cells, and downstream signaling (48). Because activated T cells express significantly higher levels of rafts and p56lck on the surface than naive T cells (53), the exquisite antigenic promiscuity of activated vs naive 8.3-CD8+ T cells can be explained by a model in which short-lived interactions between TCRs and low affinity ligands (i.e., NRP mimics bearing the two major TCR contact residues of NRP but differing from NRP at minor TCR contact residues) would be able to elicit efficient signaling in activated T cells, but not in their naive progenitors.
The increase in the antigenic promiscuity of the NRP-A7-reactive CD8+ T cell population with its avidity for peptide/MHC is likely due to preferential expansion of T cells bearing high affinity TCRs for NRP-A7 during the spontaneous anti-islet autoimmune response (34). Available structural data suggest that the Ag-binding site of TCR is flexible, that the low affinity and slow kinetics of TCR-peptide/MHC-binding is a consequence of this flexibility, and that this flexibility accounts for the cross-reactivity of TCRs for different peptide/MHC complexes (54, 55, 56). Because increases in the affinity of Abs arising by somatic hypermutation are associated with reductions in their conformational flexibility and antigenic promiscuity (57), it is somewhat surprising that the avidity maturation of the NRP-A7-reactive CD8+ T cell population leads to an increase in its peptide promiscuity. This suggests that the peptide/MHC-binding site of high affinity TCRs is more flexible than that of low affinity TCRs. Alternatively, the increased antigenic promiscuity of the high avidity NRP-A7-reactive CD8+ T cell population is due to an increased activation state.
Whatever the mechanisms, this amplification of the antigenic repertoire of single TCRs upon T cell activation may have evolved as a means for the immune system to be able to fight mutants arising during microbial or viral infections (58, 59, 60). Some of these mutant viruses encode lower affinity ligands of reactive TCRs that compete with variants carrying high affinity epitopes for TCR recognition, or encode TCR antagonists that can assist in the survival of variants carrying wild-type epitopes (59, 60). Owing to their activation-induced promiscuity, activated and memory T cells would be able to mount immune responses against escape mutants that cannot elicit the activation of naive T cells bearing TCRs specific for the wild-type epitopes. This interpretation is consistent with the observation that viral infections expand polyclonal populations of T cells containing T cells capable of providing protection against a range of antigenic variants (61). However, this process is a double-edged sword that compromises the natural resistance of the hosts to autoimmunity afforded by the processes of immunological tolerance, which preferentially target TCRs with high affinity for self-peptide/MHC. Activation of a nonautoreactive T cell population by a foreign Ag during an infection would allow the differentiated T cell progeny to react against self-derived, low affinity ligands of their TCRs (autoantigens).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
cell apoptosis network. 
2 Address correspondence and reprint requests to Dr. Pere Santamaria, Department of Microbiology and Infectious Diseases, Faculty of Medicine, University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta, Canada T2N 4N1. E-mail address: psantama{at}ucalgary.ca
3 Abbreviations used in this paper: NOD, nonobese diabetic; CM, complete medium; NAT, naturally occurring mimic of NRP; rhIL-2, recombinant human IL-2; mfi, mean fluorescence intensity.
Received for publication February 23, 2001. Accepted for publication May 1, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
T cell receptors. Annu. Rev. Immunol. 16:523.[Medline]
T cell receptor structure at 2.5 A and its orientation in the TCR-MHC complex. Science 274:209.This article has been cited by other articles:
|
|
 |
|
  K. L. Graham, N. Sanders, Y. Tan, J. Allison, T. W. H. Kay, and B. S. Coulson Rotavirus Infection Accelerates Type 1 Diabetes in Mice with Established Insulitis J. Virol., July 1, 2008; 82(13): 6139 - 6149. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Tailor, S. Tsai, A. Shameli, P. Serra, J. Wang, S. Robbins, M. Nagata, A. L. Szymczak-Workman, D. A. A. Vignali, and P. Santamaria The Proline-Rich Sequence of CD3{epsilon} as an Amplifier of Low-Avidity TCR Signaling J. Immunol., July 1, 2008; 181(1): 243 - 255. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Desrosiers, K. M. Cembrola, M. J. Fakir, L. A. Stephens, F. M. Jama, A. Shameli, W. Z. Mehal, P. Santamaria, and Y. Shi Adenosine Deamination Sustains Dendritic Cell Activation in Inflammation J. Immunol., August 1, 2007; 179(3): 1884 - 1892. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Kao, M. A. Daniels, and S. C. Jameson Loss of CD8 and TCR binding to Class I MHC ligands following T cell activation Int. Immunol., December 1, 2005; 17(12): 1607 - 1617. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Lieberman, T. Takaki, B. Han, P. Santamaria, D. V. Serreze, and T. P. DiLorenzo Individual Nonobese Diabetic Mice Exhibit Unique Patterns of CD8+ T Cell Reactivity to Three Islet Antigens, Including the Newly Identified Widely Expressed Dystrophia Myotonica Kinase J. Immunol., December 1, 2004; 173(11): 6727 - 6734. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Reijonen, R. Mallone, A.-K. Heninger, E. M. Laughlin, S. A. Kochik, B. Falk, W. W. Kwok, C. Greenbaum, and G. T. Nepom GAD65-Specific CD4+ T-Cells with High Antigen Avidity Are Prevalent in Peripheral Blood of Patients With Type 1 Diabetes Diabetes, August 1, 2004; 53(8): 1987 - 1994. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Moore, J. Grimm, B. Han, and P. Santamaria Tracking the Recruitment of Diabetogenic CD8+ T-Cells to the Pancreas in Real Time Diabetes, June 1, 2004; 53(6): 1459 - 1466. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. V. Serreze, T. M. Holl, M. P. Marron, R. T. Graser, E. A. Johnson, C. Choisy-Rossi, R. M. Slattery, S. M. Lieberman, and T. P. DiLorenzo MHC Class II Molecules Play a Role in the Selection of Autoreactive Class I-Restricted CD8 T Cells That Are Essential Contributors to Type 1 Diabetes Development in Nonobese Diabetic Mice J. Immunol., January 15, 2004; 172(2): 871 - 879. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Yamanouchi, J. Verdaguer, B. Han, A. Amrani, P. Serra, and P. Santamaria Cross-Priming of Diabetogenic T Cells Dissociated from CTL-Induced Shedding of {beta} Cell Autoantigens J. Immunol., December 15, 2003; 171(12): 6900 - 6909. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Hutton and G. S. Eisenbarth A pancreatic {beta}-cell-specific homolog of glucose-6-phosphatase emerges as a major target of cell-mediated autoimmunity in diabetes PNAS, July 22, 2003; 100(15): 8626 - 8628. [Full Text] [PDF]
|
|
|
|
|
|
S. M. Lieberman, A. M. Evans, B. Han, T. Takaki, Y. Vinnitskaya, J. A. Caldwell, D. V. Serreze, J. Shabanowitz, D. F. Hunt, S. G. Nathenson, et al. Identification of the {beta} cell antigen targeted by a prevalent population of pathogenic CD8+ T cells in autoimmune diabetes PNAS, July 8, 2003; 100(14): 8384 - 8388. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Kimachi, K. Sugie, and H. M. Grey Effector T cells have a lower ligand affinity threshold for activation than naive T cells Int. Immunol., July 1, 2003; 15(7): 885 - 892. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
