International Immunology, Vol. 15, No. 5, pp. 623±632
doi: 10.1093/intimm/dxg056
ã 2003 The Japanese Society for Immunology
Mutually antagonistic signals regulate
selection of the T cell repertoire
Geoffrey L. Stephens1,3, Jonathan D. Ashwell2 and Leszek Ignatowicz1
1Institute of Molecular Medicine and Genetics, Medical College of Georgia, Augusta, GA 30912-2600,
USA
2Laboratory of Immune Cell Biology, National Cancer Institute, National Institutes of Health, Bethesda,
MD 20892, USA
3Present
address: Laboratory of Immunology, National Institute of Allergy and Infectious Diseases,
National Institutes of Health, Bethesda, MD 20892, USA
Keywords: knockout, lymphocyte development, T lymphocyte, thymus, transgenic
Abstract
The sensitivity of T cells to agonist-induced death during development contrasts with their
proliferative responses after agonist challenge in the periphery. The means by which TCR
engagement results in these distinct outcomes is incompletely understood. It has been previously
hypothesized that glucocorticoids (GC) modulate the threshold for thymocyte activation by
blunting the consequences of TCR engagement. In support of this possibility, inhibition of GC
production in fetal thymic organ culture was shown to result in CD4+CD8+ thymocyte apoptosis.
This was dependent upon MHC diversity, implying that endogenous GC might regulate antigenspeci®c selection. Similarly, mice expressing reduced GC receptor (GR) levels due to the presence
of an antisense transgene have fewer CD4+CD8+ thymocytes, which was due to an impaired
transition from CD4±CD8± precursors and increased apoptosis. Here we ask how manipulating
peptide diversity in the context of reduced GC signaling might affect T cell development and
function. In mice with impaired GR expression there was a rescue of thymocyte cellularity and
proportions as the diversity of peptides presented by self-MHC was reduced. Furthermore, whereas
more CD4+ T cells survived the selection process in mice expressing single-peptide±MHC class II
complexes and reduced GR levels, these cells largely failed to recognize the same MHC molecules
bound with foreign peptides. Together, these results support a role for endogenous GC in
balancing TCR-mediated signals during thymic selection.
Introduction
Selection processes in the thymus ensure that peripheral T
cells ful®ll two essential prerequisites: activation by foreign
peptides bound to host MHC molecules, but tolerance to selfderived peptides presented in the same context. Ful®llment of
the former occurs passively in the thymus, by tying thymocyte
survival and subsequent differentiation to self-restriction (1,2).
In contrast, the latter requires the active elimination of
thymocytes that react strongly with self-peptide±MHC molecules (1,3,4). It is thought that the strength of the TCRmediated signal received during selection ultimately distinguishes T cells that survive from those that do not. Selection
occurs only if signals received through the TCR fall between
two thresholds; signals below one are incompatible with
survival (neglected death), while signals above another result
in deletion (5,6).
By in¯uencing the strength of TCR engagement, coreceptors and adhesion molecules can modify the outcome
of thymocyte selection (7±10). Soluble factors (e.g. IL-7, IL-4)
present in the thymic milieu provide additional cues for
thymocytes that may in¯uence differentiation, division or
death (11,12). In addition, glucocorticoids (GC) produced
by thymic epithelial cells can modify the biological consequences of TCR-mediated signaling (13). While signaling
through either the TCR or GC receptor (GR) individually results
in apoptosis, simultaneous exposure to both antagonizes the
death pathway (13,14). Thus, GC, by modulating the signaling
threshold during thymic selection, may in¯uence the range of
selectable avidities for a particular TCR, thereby directly
in¯uencing the antigen-speci®c T cell repertoire (15±18). This
phenomenon of `mutual antagonism' potentially provides an
Correspondence to: L. Ignatowicz; E-mail: [email protected].
Transmitting editor: T. Sasazuki
Received 18 October 2002, accepted 20 February 2003
624 Glucocorticoids in¯uence thymic selection
Glucocorticoids in¯uence thymic selection 625
explanation to a long-standing paradoxÐhow do thymocytes
distinguish between death and survival signals originating
from the same receptor (14)? Implicit in the mutual antagonism
model is the idea that a reduction in GR activity should
increase the thymocyte response depending on the degree of
TCR occupancy (19). In this model, thymocytes bearing TCR
with avidities appropriate for selection undergo deletion, while
those bearing TCR with inadequate avidities undergo selection (20,21).
The role of endogenous GC in thymocyte selection has been
examined previously in transgenic mice that express an
antisense GR transcript speci®cally in the thymus (homozygous transgenic mice are referred to as TKO), but not in the
periphery. These mice displayed a hypoplastic thymus,
hyporesponsiveness to GC and increased sensitivity to
apoptosis induced by anti-TCR antibodies (22). Other studies
using GR `knockout' mice have cast doubt on the inferences
made using pharmacologic inhibitors of GC synthesis, GC
antagonists and TKO mice (23).
In light of this, we have re-examined the possibility that GC
in¯uence the consequences of TCR-mediated signaling during thymocyte development. To do so, we bred TKO mice with
strains differing in their ability to present peptides in the
context of MHC-encoded molecules. We have tested the
prediction that an altered antigen-speci®c repertoire is a
consequence of reduced thymic GC responsiveness by
comparing the repertoires and reactivities of CD4+ T cells
selected on a single-peptide±MHC complex. Our data indicates that GC affect T cell repertoire formation by restricting
the development of T cells to those that recognize host MHC.
Thus, by setting a threshold for thymocyte selection, GC inhibit
the development of useless T cells and consequently enforce
MHC restriction.
Methods
Animals
Mice engineered to express transgenic AbEp complexes in the
absence of endogenous Abb0 and invariant chain (Ii) were
generated at the National Jewish Medical and Research
Center (Denver, CO) as previously described (24). TKO mice
that express a fragment of the 3¢-untranslated region of the rat
GR under the control of the proximal lck promoter were
described elsewhere (22). TKO mice were backcrossed with
the AbEpIi0 strain to generate mice homozygous for the TKO
transgene, which expressed the AbEp transgene, but were
devoid of the Ii and endogenous Abb (TKO.AbEpIi0). Some
TKO.AbEpIi0 mice were further crossed with AbEpIi0b2m0 mice.
TKO mice de®cient in both Abb and b2m were generated by
backcrosses with doubly de®cient Abb0b2m0 mice. TKO mice
singly de®cient for Ii, Abb or b2m resulted from the crosses
mentioned above. All mice used in experiments were 6±8
weeks old and were on the H-2b (C57BL/6) genetic background. Mice de®cient in b2m were originally purchased from
the Jackson Laboratory (Bar Harbor, ME).
AutoMACS sorting of cells
Cells were then incubated in the presence of biotinylated
monoclonal antibody for 15 min at 4°C and rinsed with MACS
running buffer (PBS containing 1 mM EDTA). Cells were then
incubated for an additional 15 min at 4°C in the presence of
MACS streptavidin beads, rinsed and sorted with an
AutoMACS cell separator (Miltenyi Biotec, Bergisch
Gladbach, Germany) according to the manufacturer's protocol.
Preparation of cells and ¯ow cytometric analysis
Tissues were harvested and dissociated into single-cell
suspensions. The following antibodies were used for ¯ow
cytometric analysis: allophycocyanin±anti-CD4 (clone GK1.5),
perCP±anti-CD8a (clone 53-6.7), PE±anti-CD69 (clone
H1.2F3) and FITC±anti-CD5 (clone 53-7.3) (purchased from
PharMingen, San Diego, CA). Samples were stained in BSS
wash buffer (BSS containing 2% FCS and 0.1%NaN3) after
pre-incubation for 5 min with anti-FcgII/III (clone 2.4G2)
antibody. Samples were acquired using a dual-laser
FACSCalibur (Becton Dickinson Immunocytometry Systems,
Mountain View, CA) with CellQuest software (Becton
Dickinson). Analysis of ¯ow cytometry data was carried out
using WinMDI (Joseph Trotter, La Jolla, CA) or WinList (Verity
Software House, Topsham, ME) software. Dead cells were
excluded from analysis by gating appropriate forward and
side scatter.
Analysis of apoptosis
Samples were ®rst stained with antibodies as described
above. To measure apoptotic cell death, following staining,
cells were rinsed and resuspended in 1 3 Annexin binding
buffer (0.01 HEPES, pH 7.4; 0.14 M NaCl; 2.5 mM CaCl2).
Annexin V (PharMingen) was added to samples according to
the manufacturer's recommendations and subsequently analyzed by ¯ow cytometry.
Generation of T cell hybridomas
Plastic 75-cm2 tissue culture ¯asks were coated with puri®ed
anti-CD3 (clone 145-2C11) at a concentration of 25 mg/ml
overnight at 4°C and vigorously rinsed at least 5 times with
PBS before use. Pooled mesenteric, axillary and inguinal
lymph nodes were isolated and prepared for MACS sorting as
described above. Puri®ed CD4+ T cells from control and TKO
Fig. 1. The rescue of CD4+CD8+ thymocytes is inversely proportional to the diversity of peptides presented by MHC-encoded molecules. (A)
Representative two-color ¯ow cytometry analysis after staining cells with allophycocyanin-labeled anti-CD4 and perCP-labeled anti-CD8
antibodies. The phenotype of each pair of mice is shown above and below the middle table. Proportion of positive cells present in the thymus
(TKO+ shown in upper panels; A±F and M±R) and lymph nodes (TKO± shown in lower panels; G±L and S±X) of indicated mice. Numbers in
quadrants represent the corresponding percentages of each respective population. Bar graphs representing absolute (B) thymocyte and (C)
CD4+ lymph node cell numbers of the indicated populations were calculated from the percentage of cells determined by ¯ow cytometry
shown above. Results, represented as means 6 SEM, were from four to seven independent experiments. Mice were sacri®ced between 4 and
6 weeks of age. Signi®cant differences between TKO mice of different genotypes were determined by the Mann±Whitney U-test. *P < 0.05.
626 Glucocorticoids in¯uence thymic selection
Glucocorticoids in¯uence thymic selection 627
mice were then added to the culture ¯asks and incubated for 6
days. On the third day of incubation, IL-2 was added to the
culture ¯asks. At the end of the culture period, the resulting T
cell blasts were fused with the thymoma cell line BW5147 a±b±
to generate T cell hybridomas as described previously (24).
Analysis of the reactivities of T cell hybridomas
Uncloned T cell hybridomas from AbEpIi0 and TKO.AbEpIi0
mice were brie¯y expanded on individual wells of 24-well
plates, and assessed for expression of CD4 and TCR by ¯ow
cytometry. Hybridomas expressing high levels of CD4 were
tested for IL-2 production in response to spleen cells from
b2m0 and Ii0 mice. In parallel, hybridomas were checked for
their ability to produce IL-2 after overnight incubation on antiCD3-coated 96-well plates. To assess the frequency of
responders from each type of mouse, only hybridomas
responding to anti-CD3 cross-linking were included in the
analysis. Measurement of IL-2 production was accomplished
by overnight incubation of the IL-2-dependent cell line HT-2
with supernatants harvested from the responders. HT-2 cells
were incubated for 5 h at 37°C in the presence of 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT;
Sigma, St. Louis, MO). Viability was determined by measuring
absorbance at 590 nm in each well containing HT-2 cells using
an ELISA plate reader.
Results
Rescue of thymocytes in TKO mice correlates with the
diversity of MHC-bound peptides
If the increased death of double-positive thymocytes in TKO
mice were due to unopposed TCR occupancy, one would
predict that it would vary as a function of antigen/MHC
diversity. To test this, we examined the CD4/CD8 pro®le in
thymocytes and lymph nodes from the indicated strains of
TKO mice that differ in their abilities to present self-peptides
(Fig. 1A). In the presence of a diverse array of peptides, TKO
mice displayed a sharp reduction in thymic cellularity and
relative decrease in the proportion of CD4+CD8+ thymocytes
(Fig. 1Aa versus Am), as previously reported (22). In the lymph
nodes, there was a 1:1 ratio of CD4 to CD8 T cells [Fig. 1Ag
and (22)], whereas in non-transgenic animals this ratio was
closer to 1.5:1 (Fig. 1As). As the Ii plays an essential role in the
endosomal traf®cking and loading of peptides into the binding
groove of MHC class II, in its absence only a fraction of the
normal spectrum of self-peptides is loaded into Ab (24). The
effect of limited peptide diversity during thymic selection was
noticeable when TKO mice were crossed onto the Ii-de®cient
background (Fig. 1Ab), where an increase in the proportion of
CD4+CD8+ thymocytes was apparent. Consistent with impaired positive selection in the absence of a full complement
of self-peptides, lymph node CD4+ T cells were proportionally
less in both TKO and non-transgenic animals (cf. Fig. 1At to
Ah). Occupancy of all MHC class II molecules with a single
peptide (AbEpIi0) in TKO mice resulted in a further increase in
the proportion of CD4+CD8+ cells (Fig. 1Ac). This trend
continued in TKO.MHC II0 mice that expressed only MHC
class I-encoded molecules (Fig. 1Ad versus Ap) and in mice
whose only MHC-encoded molecule is bound with a single
peptide (TKO.AbEpIi0b2m0) (Fig. 1Ae), where the relative
proportion of CD4+CD8+ cells approached control levels
(Fig. 1Aq). Finally, TKO mice lacking expression of all MHCencoded molecules [TKO.MHC (I 3 II)0] had a normal
proportion of double-positive thymocytes (Fig. 1Af).
The absolute number of CD4+CD8+ thymocytes in TKO mice
also correlated inversely with the degree of peptide diversity.
This was most evident in mice that expressed only singlepeptide±MHC (Fig. 1B, left graph). Conversely, the number of
CD4 single-positive (SP) thymocytes decreased as the
peptide diversity was reduced (Fig. 1B, middle graph), while
the numbers of CD8 SP thymocytes did not signi®cantly vary
from mouse to mouse. The rescue of CD4+CD8+ thymocyte
numbers was incomplete, however (~40 or ~33% of in same
genotype TKO± animals respectively), consistent with the
observation that TKO mice have a decreased proportion of
cells undergoing the CD4±CD8± to CD4+CD8+ transition, an
effect that one would not expect TCR ligand diversity to alter.
The number of lymph node CD4+ T cells in TKO.Abwt animals
was reduced, similar to previous studies (Fig. 1C) (22). TKO
mice expressing fewer Ab-bound peptides during selection
had similar numbers of lymph CD4+ T cells compared to nontransgenic animals, which is consistent with the limited
capability of a few or single peptide(s) to impose negative
selection (25).
Rescue of thymocyte proportions and cellularity depends on
the diversity of peptides presented by Ab, not its expression
level
In addition to reducing the number of peptides presented
during selection, the absence of Ii results in a reduction in the
surface expression of Ab by bone marrow-derived antigenpresenting cells (APC) and thymic epithelial cells. Therefore,
the rescue of thymocyte numbers observed in TKO mice may
be a consequence of decreased expression of Ab, as all of the
`low peptide diversity' mice tested were Ii de®cient. Mice
presenting a diverse array of peptides in the context of
reduced numbers of surface Ab (designated AbEpIi+) have
been described (26). The AbEpIi+ strain expresses Ab
covalently linked to Ep52±68 as a transgene and in the absence
of endogenous Ab. Due to the presence of the Ii, the covalently
attached Ep52±68 peptide is cleaved from the binding groove
of transgenic Ab and subsequently replaced by a diverse array
of self-peptides (27,28). Due to low expression of the
Fig. 2. Rescue of thymocytes in TKO mice is dependent on reduced peptide diversity and is not related to expression level of Ab. (A)
Expression of CD4 and CD8 in TKO.AbEpIi+ and AbEpIi+ mice was determined by ¯ow cytometry after staining single-cell thymocyte (top
panels) and lymph node (lower panels) suspensions with allophycocyanin-labeled anti-CD4 and perCP-labeled anti-CD8 antibodies. Numbers
in quadrants represent the corresponding percentages of each respective population. (B and C) Absolute cell numbers were calculated from
the percentage of cells recovered as determined by the ¯ow cytometric analysis. Results are represented as means 6 SEM calculated from
three independent experiments (TKO.AbEpIi+, n = 5; AbEpIi+, n = 4) from mice sacri®ced between 4 and 6 weeks of age.
628 Glucocorticoids in¯uence thymic selection
Fig. 3. Reducing thymocyte exposure to GC increases sensitivity to TCR engagement. (A) Expression of CD5 and (C) CD69 on gated
CD4+CD8+ thymocytes correlates with increased sensitivity to TCR stimuli and is inversely related to the availability of TCR ligands.
Histograms show analysis of CD5 expression by ¯ow cytometry after surface staining with anti-CD5±FITC antibodies. Histograms are shown
after gating on the populations indicated at the top of each column. (B) Bar graphs of CD5 expression represent the mean ¯uorescence
intensities 6 SEM. (C) Flow cytometric analyses of CD69 expression was accomplished by staining single-cell thymocyte suspensions.
Histograms shown are gated on CD4+CD8+ thymocytes in mice with reduced (TKO; solid line) or normal (Non-TKO; broken line) levels of GR
on thymocytes. The genotype of each mouse is indicated to the left of the panel. Results are represented as means 6 SEM and signi®cant
differences between TKO and non-Tg mice were determined by the Mann±Whitney U-test. Results were obtained from four independent
experiments (TKO.MHC II0, n = 5; MHC II0, n = 4; TKO.Ii0, n = 9; Ii0, n = 9; TKO.Abwt, n = 5; Abwt, n = 4). *P < 0.010, **P < 0.005
transgene-encoded Ab, thymic epithelial cells express 2-fold
fewer Ab complexes while APC express ~10-fold less. These
mice have ~2 times more CD4+ T cells and CD4+ SP
thymocytes compared to wild-type mice due to inef®cient
negative selection, while the ef®ciency of positive selection is
unperturbed (26). The inef®cient clonal deletion of CD4+ T
cells in these mice results in an autoimmune syndrome
characterized by an activation of T and B cells, which leads
to progressive lymphoadenopathy and minor hypergammaglobulinemia (26).
We crossed TKO transgenics onto the AbEpIi+ background
to distinguish between the effects of decreased peptide
diversity and reduced Ab expression. Flow cytometric analysis
of total splenocytes from both AbEpIi+ and Ii0 mice previously
demonstrated a similar 10-fold reduction in the surface levels
of Ab in each strain, consistent with previous ®ndings (26). The
reduced expression of Ab, however, had little impact on
restoring the cellularity and proportions of thymocytes in TKO
mice, which remained well below non-transgenic levels
(Fig. 2A, top left). Contrary to rescue of thymocytes observed
in mice devoid of MHC, the proportion of CD4+CD8+
thymocytes was decreased in TKO.AbEpIi+ mice in comparison to TKO.Abwt mice (Fig. 2B). No signi®cant difference was
observed in either the absolute number of CD4+CD8+
thymocytes (Fig. 2B, right graph), suggesting that the
decrease occurred passively due to an increase in CD4+ SP
thymocytes in TKO.AbEpIi+ mice. Furthermore, an increase in
the CD4/CD8 ratio in the lymph nodes of TKO.AbEpIi+ mice to
~1.4:1 from the 1:1 ratio observed in TKO.Abwt mice further
validates this point (Fig. 2C). Thus, the diversity of TCR ligands
available for selection ultimately determines the extent of
thymocyte deletion in TKO mice.
Thymocyte resistance to GC leads to MHC-dependent
increases in the expression of activation markers
Expression of CD5 and CD69 on immature thymocytes is
regulated by the avidity of the selecting interaction (29±31).
Previous work demonstrated the level of CD5 on CD4+CD8+
thymocytes was increased in TKO mice (16). To determine if
reduced thymocyte exposure to GC increased their sensitivity
to TCR engagement in vivo, we quantitated the expression of
the activation markers CD5 and CD69 in a panel of TKO mice
that differed in the diversity of peptides presented during
selection. The level of CD5 on CD4+CD8+ thymocytes was
increased over the control levels in all mice expressing the
TKO transgene (Fig. 3A, left column). This difference was
signi®cant as calculated from the mean ¯uorescence intensity
values (Fig. 3B, left graph; *P < 0.005, **P < 0.010). A
corresponding increase in the level of CD5 on CD4+CD8+
thymocytes accompanied an increase in peptide diversity in
both control and TKO mice, suggesting that CD5 expression
re¯ects the availability of TCR ligands (Fig. 3A, left most
column). SP CD4 thymocytes from TKO.MHC II0 and nontransgenic MHC II0 mice expressed indistinguishable levels of
Glucocorticoids in¯uence thymic selection 629
Table 1. Absence of T cells from TKO mice selected on a
single AbEp complex that respond to other peptides bound
to Ab
T cell hybridoma
sourcea
TKO.AbEpIi0
AbEpIi0
Fig. 4. The extent of thymocyte apoptosis in TKO mice is related to
the diversity of MHC-bound peptides. (A) Flow cytometric analysis of
apoptosis by Annexin V±FITC was accomplished by staining freshly
prepared thymocyte single-cell suspensions. Histograms shown are
gated on CD4+CD8+ thymocytes in mice with reduced (TKO; solid
line) or normal (non-Tg; broken line) thymic levels of the GR. Text at
the right indicates the genotype of the respective mice. (B)
Histograms represent CD4+CD8+ thymocytes stained with Annexin
V+-FITC from mice of the indicated genotypes. Results are
represented as means 6 SEM, and signi®cant differences between
TKO and non-Tg mice were determined by the Mann±Whitney Utest. Results were obtained from four independent experiments
(TKO.MHC II0, n = 5; MHC II0, n = 4; TKO.Ii0, n = 9; Ii0, n = 9;
TKO.Abwt, n = 5; Abwt, n = 4). *P < 0.005.
CD5 (Fig. 3A and B, middle columns), indicating that the
increased level seen in TKO mice depends on TCR engagement and not an intrinsic property of the transgene. The
expression of CD5 on CD4 SP thymocytes on both Abwt and Ii0
backgrounds was also higher in TKO mice than in nontransgenics, and again its level correlated with the diversity of
self-peptides. Consistent with the unvarying level of MHC
class I on all of the mice tested, the CD5 expression on CD8 SP
thymocytes mice was unchanged among TKO mice crossed
on MHC II0, Ii0 and Abwt genetic backgrounds. However, the
level of CD5 on CD8+ T cells was on average increased
compared to the corresponding non-transgenic control population, although less dramatically than in CD4+ T cells (Fig. 3A
and B, last column). The expression of CD69 was also higher
on the bulk of CD4+CD8+ thymocytes in mice bearing the TKO
transgene (Fig. 3C). Together, these data further support the
notion that a reduction in GC blunting of TCR-mediated signals
alters the perceived avidity of TCR±MHC engagement.
Increased thymocyte apoptosis in TKO mice depends on
recognition of self-peptide±MHC
Thymocytes bearing TCR with high avidity for self-peptide±
MHC undergo apoptosis during thymic selection (1). In
previous reports, inhibition of GC synthesis with metyrapone
in fetal thymic organ culture enhanced the TCR-dependent
apoptosis of thymocytes (21). To examine this issue in a more
physiologic setting, we asked if decreasing the diversity of
selecting peptides during thymic selection correlated with a
reduction in the apoptotic index of thymocytes. Apoptosis was
measured in fresh thymocytes by assaying for phosphatidylserine exposure using Annexin V (Fig. 4A). As the proportion of
Frequency responding to APC from the
followingb
Abwtb2m0c
Ii0
10
49
0
28
aT cell blasts from each type of mouse were fused with the
thymoma cell line BW5147a-b- as described in Methods. Individual
hybridoma clones were screened for expression of CD4 and TCR
before further use. In addition, hybridomas were incubated on antiCD3 were excluded from the above analysis.
bResults from 21 and 39 individual hybridomas from AbEpIi0 and
TKO AbEpIi0, respectively.
cSigni®cant difference between groups with Fisher's exact
probability, P = 0.036.
MHC class II-bound peptides decreased, a signi®cant
decrease in CD4+CD8+ thymocyte apoptosis was apparent
(Fig. 4B). Therefore, the extent of apoptosis due to reduced
GC exposure during thymocyte development is dependent
upon cognate interactions with a diverse array of selfpeptides.
Thymocyte development in an environment of reduced GC
signaling impacts the repertoire of selected T cells
We hypothesized that if GC shift the avidity `window' for thymic
selection, the peripheral repertoire of T cells selected in TKO
mice might have a different capacity to recognize foreign
peptides bound to host MHC. Previously, mice have been
engineered that express a single (32) peptide tethered to
MHC class II molecules. In these mice, the b chain of Ab is
covalently linked to a peptide derived from the a chain of the
Eb MHC class II protein (Ep52±68) which, in the absence of the Ii
and endogenous Abb, covalently bound Ep is the only
detectable peptide presented to MHC class II-restricted T
cells (33). As a single peptide bound to MHC class II mediates
both positive and negative selection in this model, many CD4+
T cells are selected to mature that would be deleted in wildtype mice because of their strong cross-reactivity with other
self-peptides (32,34).
To examine how reduced GC signaling during thymocyte
development affects the peripheral T cell repertoire, we took
advantage of this cross-reactivity by breeding TKO mice onto
the AbEpIi0 background. Hybridomas prepared from lymph
node CD4+ T cells isolated from TKO.AbEpIi0 and AbEpIi0 mice
were compared for their responses to APC from b2m0 mice (to
exclude any responses of CD4+ T cells selected on nonclassical MHC molecules). As a control, the same hybridomas
were incubated in parallel on 96-well plates coated with antiCD3 antibody and only CD3-responsive hybridomas were
included in the study (supplemental data). Consistent with
previous results (32), a large fraction of hybridomas from
AbEpIi0 responded to splenocytes from Abwt mice and to a
lesser extent splenocytes from Ii-de®cient mice (Table 1).
Strikingly, however, few or no hybridomas from TKO.AbEpIi0
mice responded to APC from b2m0 or Ii0 mice respectively.
Thus, consistent with the possibility that endogenous GC
630 Glucocorticoids in¯uence thymic selection
affect selection by inhibiting the development of thymocytes
with low avidity for self-peptide±MHC, the repertoire of CD4+ T
cells selected from TKO.AbEpIi0 mice failed to respond upon
encounter with foreign peptides.
Discussion
Several observations led to the hypothesis that locally
produced GC regulate the signaling threshold for thymocytes undergoing positive and negative selection. Initial
studies of GC function using T cell hybridomas demonstrated that stimulation through either the GR or TCR
individually led to apoptosis, while simultaneous occupancy
of both receptors antagonized the death pathway (14,35).
In addition, treatment of fetal thymic organ culture with
reagents that inhibit either GC production or GR binding
enhanced TCR-induced apoptosis (13). The observation
that GC antagonize TCR-triggered signaling events formed
the foundation of the mutual antagonism model of
thymocyte development (14). A testable prediction stemming from this model is that altering the responses of preselection thymocytes to GC will result in a corresponding
change in the mature T cell repertoire.
By varying the level of available TCR ligands and, consequently, the probability of TCR occupancy, a direct correlation
between the restoration of the CD4+CD8+ thymocyte numbers
and the occupancy of GC and TCR was most apparent in TKO
mice doubly de®cient in the expression of MHC class I and II
molecules. The restoration of the CD4+CD8+ thymocyte
proportions was accompanied by a substantial, albeit incomplete, restoration of CD4+CD8+ thymocyte numbers. The lack
of a complete rescue of thymic cellularity indicates that the
observed reduction in thymus size cannot be due entirely to
enhanced TCR-mediated deletion of CD4+CD8+ thymocytes,
consistent with the observation that there is a partial block in
the double-negative to double-positive transition in these
animals (22).
The low frequency of T cell hybridomas generated from
TKO.AbEpIi0 mice that recognize syngeneic Abwt cells might
re¯ect an enhanced deletion of the cohort of T cells that
normally respond to other self-peptides bound to Ab. This
would be consistent with an overall lowered threshold required
for negative selection in the absence of the appropriate
amount of GC blunting of TCR-mediated signals. In this
scenario, the cells that survive selection might be skewed
towards those with low functional avidities, having been
primarily selected on conserved regions of MHC molecules
independent of their ability to scan peptides. It is plausible that
such `framework' selected cells could develop if one considers a two-step mechanism for TCR recognition (36). In this
model, an initial TCR contact with MHC is established with little
contribution from the peptide. Once settled on the MHC, the
TCR senses the peptide with its two CDR3 loops, only
becoming activated if a stable interaction is established.
Therefore, within the context of this model it is conceivable that
when GC-mediated signaling is reduced, a subsequent
increase in the perceived TCR avidity may allow MHC binding
in the ®rst step to support positive selection, independent of
recognition of MHC-bound peptides.
It is also possible that many CD4+ T cells that normally
respond to foreign peptides in AbEpIi0 mice are present in
an anergized state in TKO.AbEpIi0 mice and are consequently not likely to be included among the hybridomas
generated, as fusion strongly favors those cells that are
actively dividing. Thus, the hybridomas from these mice
might simply re¯ect an over-representation of those CD4+
T cells selected on non-classical MHC molecules, which
are passively over-represented among total CD4+ T cells in
AbEpIi0 mice (25). However, this seems unlikely given the
paucity of these cells in lymph nodes (37) and the failure
of hybridomas from TKO.AbEpIi0 mice to respond to Ii0
splenocytes (Table 1), which were b2m suf®cient. We are
currently investigating the presence of anergized T cells in
TKO mice, in particular those with reportedly suppressive
capabilities.
Alternatively, a reduced GC environment may sensitize
thymocytes to TCR engagement revealing an otherwise
undetectable leak of low-abundance self-peptides in AbEpIi0
mice, upon which TKO thymocytes undergo negative selection. In fact, a predominant involvement of low-abundance
peptides in the selection of CD4+ T cells was proposed by
another group using a different `single-peptide' model (38).
They suggested that the predominately expressed AbEp
complexes likely had little impact on the appearance of
CD4+ T cells and therefore, an undetectable leak of other
peptides must be mediating the selection process. However,
the same group more recently reported that many T cells are
indeed
selected
on
abundant
peptide(s)
(39).
Notwithstanding, while a leak of endogenous peptides in
other single peptide systems is readily detectable, several
reports have failed to detect any such leak in AbEpIi0 mice
(33,38,40).
A considerable body of evidence suggests that GC affect
thymocyte development (13,15±18,21,22,41). One study has
reported that thymocyte development is grossly normal in
mice de®cient in GR (23). The extrapolation of that observation
to the present study is problematic, however. First, neither
antigen-speci®c selection nor TCR repertoire was analyzed in
those mice and therefore those data do not bear upon the
present analysis of antigen-speci®c thymocyte development.
Second, the same authors have recently reported that the
animals are not true knockout mice, but express a truncated
GR lacking exon 2-encoded residues (the N-terminal half of
the molecule) but containing the C-terminal half of the
molecule, which includes DNA-binding, ligand-binding and
transactivation domains (42). These truncated GR are able to
bind GC with the same af®nity as wild-type mice and cDNA
microarray analysis of GC-treated thymocytes has shown that
this receptor is capable of modulating the expression of a
large number of genes (J. D. Ashwell, manuscript submitted).
Therefore, one cannot conclude that the GR is dispensable for
any aspects of thymocyte development based on the analysis
of these animals.
In conclusion, an assessment of the impact of thymic GC
hyporesponsiveness was carried out in mice, which present
many, few, one or no peptides in the context of MHC class II
molecules. Our results are consistent with the hypothesis that
endogenous GC bias the selection of TCR towards those with
a re®ned capacity to scan various peptides presented by
Glucocorticoids in¯uence thymic selection 631
MHC-encoded molecules. The partial rescue of CD4+CD8+
thymocytes observed in TKO mice, when the probability of
TCR occupancy was reduced, further supports the idea that
signals derived from the GR oppose TCR-mediated signaling
during thymocyte development.
18
19
20
Abbreviations
b2m
APC
GC
GR
Ii
SP
b2-microglobulin
antigen-presenting cell
glucocorticoid
glucocorticoid receptor
invariant chain
single positive
21
22
23
References
1 Surh, C. D. and Sprent, J. 1994. T-cell apoptosis detected in situ
during positive and negative selection in the thymus [see
Comments]. Nature 372:100.
2 Nikolic-Zugic, J. and Bevan, M. J. 1990. Role of self-peptides in
positively selecting the T-cell repertoire. Nature 344:65.
3 Kisielow, P., Bluthmann, H., Staerz, U. D., Steinmetz, M. and von
Boehmer, H. 1988. Tolerance in T-cell-receptor transgenic mice
involves deletion of nonmature CD4+8+ thymocytes. Nature
333:742.
4 Kappler, J. W., Roehm, N. and Marrack, P. 1987. T cell tolerance
by clonal elimination in the thymus. Cell 49:273.
5 Ashton-Rickardt, P. G. and Tonegawa, S. 1994. A differentialavidity model for T-cell selection. Immunol. Today 15:362.
6 Jameson, S. C., Hogquist, K. A. and Bevan, M. J. 1995. Positive
selection of thymocytes. Annu. Rev. Immunol. 13:93.
7 Zuniga-P¯ucker, J. C., McCarthy, S. A., Weston, M., Longo, D. L.,
Singer, A. and Kruisbeek, A. M. 1989. Role of CD4 in thymocyte
selection and maturation. J. Exp. Med. 169:2085.
8 Fine, J. S. and Kruisbeek, A. M. 1991. The role of LFA-1/ICAM-1
interactions during murine T lymphocyte development. J.
Immunol. 147:2852.
9 Chidgey, A. and Boyd, R. 1997. Agonist peptide modulates T cell
selection thresholds through qualitative and quantitative shifts in
CD8 co-receptor expression. Int. Immunol. 9:1527.
10 Amsen, D. and Kruisbeek, A. M. 1998. Thymocyte selection: not
by TCR alone. Immunol. Rev. 165:209.
11 Kishimoto, H. and Sprent, J. 1999. Several different cell surface
molecules control negative selection of medullary thymocytes.
J. Exp. Med. 190:65.
12 Brugnera, E., Bhandoola, A., Cibotti, R., Yu, Q., Guinter, T. I.,
Yamashita, Y., Sharrow, S. O. and Singer, A. 2000. Coreceptor
reversal in the thymus: signaled CD4+8+ thymocytes initially
terminate CD8 transcription even when differentiating into CD8+
T cells. Immunity 13:59.
13 Vacchio, M. S., Papadopoulos, V. and Ashwell, J. D. 1994. Steroid
production in the thymus: implications for thymocyte selection.
J. Exp. Med. 179:1835.
14 Zacharchuk, C. M., Mercep, M., Chakraborti, P. K., Simons, S. S.,
Jr and Ashwell, J. D. 1990. Programmed T lymphocyte death. Cell
activation- and steroid-induced pathways are mutually
antagonistic. J. Immunol. 145:4037.
15 Tolosa, E., King, L. B. and Ashwell, J. D. 1998. Thymocyte
glucocorticoid resistance alters positive selection and inhibits
autoimmunity and lymphoproliferative disease in MRL-lpr/lpr mice.
Immunity 8:67.
16 Vacchio, M. S., Lee, J. Y. and Ashwell, J. D. 1999. Thymusderived glucocorticoids set the thresholds for thymocyte selection
by inhibiting TCR-mediated thymocyte activation. J. Immunol.
163:1327.
17 Lu, F. W., Yasutomo, K., Goodman, G. B., McHeyzer-Williams,
L. J., McHeyzer-Williams, M. G., Germain, R. N. and Ashwell, J. D.
2000. Thymocyte resistance to glucocorticoids leads to antigen-
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
speci®c unresponsiveness due to `holes' in the T cell repertoire.
Immunity 12:183.
Van Laethem, F., Baus, E., Smyth, L. A., Andris, F., Bex, F.,
Urbain, J., Kioussis, D. and Leo, O. 2001. Glucocorticoids
attenuate T cell receptor signaling. J. Exp. Med. 193:803.
Ashwell, J. D., Lu, F. W. and Vacchio, M. S. 2000. Glucocorticoids
in T cell development and function. Annu. Rev. Immunol. 18:309.
Zacharchuk, C. M., Mercep, M. and Ashwell, J. D. 1991.
Thymocyte activation and death: a mechanism for molding the T
cell repertoire. Ann. NY Acad. Sci. 636:52.
Vacchio, M. S. and Ashwell, J. D. 1997. Thymus-derived
glucocorticoids regulate antigen-speci®c positive selection. J.
Exp. Med. 185:2033.
King, L. B., Vacchio, M. S., Dixon, K., Hunziker, R., Margulies,
D. H. and Ashwell, J. D. 1995. A targeted glucocorticoid receptor
antisense transgene increases thymocyte apoptosis and alters
thymocyte development. Immunity 3:647.
Purton, J. F., Boyd, R. L., Cole, T. J. and Godfrey, D. I. 2000.
Intrathymic T cell development and selection proceeds normally in
the absence of glucocorticoid receptor signaling. Immunity
13:179.
Bikoff, E. K., Huang, L. Y., Episkopou, V., van Meerwijk, J.,
Germain, R. N. and Robertson, E. J. 1993. Defective major
histocompatibility complex class II assembly, transport, peptide
acquisition, and CD4+ T cell selection in mice lacking invariant
chain expression. J. Exp. Med. 177:1699.
Chmielowski, B., Muranski, P. and Ignatowicz, L. 1999. In the
normal repertoire of CD4+ T cells, a single class II MHC/peptide
complex positively selects TCR with various antigen speci®cities.
J. Immunol. 162:95.
Gaszewska-Mastalarz, A., Muranski, P., Chmielowski, B., Kraj, P.
and Ignatowicz, L. 2000. Altered selection of CD4+ T cells by class
II MHC bound with dominant and low abundance self-peptides.
J. Immunol. 165:6099.
Ignatowicz, L., Winslow, G., Bill, J., Kappler, J. and Marrack, P.
1995. Cell surface expression of class II MHC proteins bound by a
single peptide. J. Immunol. 154:3852.
Wilson, N. A., Wolf, P., Ploegh, H., Ignatowicz, L., Kappler, J.
and Marrack, P. 1998. Invariant chain can bind MHC class II at
a site other than the peptide binding groove. J. Immunol.
161:4777.
Azzam, H. S., Grinberg, A., Lui, K., Shen, H., Shores, E. W. and
Love, P. E. 1998. CD5 expression is developmentally regulated by
T cell receptor (TCR) signals and TCR avidity. J. Exp. Med.
188:2301.
Azzam, H. S., DeJarnette, J. B., Huang, K., Emmons, R., Park,
C. S., Sommers, C. L., El-Khoury, D., Shores, E. W. and Love, P. E.
2001. Fine tuning of TCR signaling by CD5. J. Immunol. 166:5464.
Wong, P., Barton, G. M., Forbush, K. A. and Rudensky, A. Y. 2001.
Dynamic tuning of T cell reactivity by self-peptide±major
histocompatibility complex ligands. J. Exp. Med. 193:1179.
Ignatowicz, L., Kappler, J. and Marrack, P. 1996. The repertoire of
T cells shaped by a single MHC/peptide ligand. Cell 84:521.
Grubin, C. E., Kovats, S., deRoos, P. and Rudensky, A. Y. 1997.
De®cient positive selection of CD4 T cells in mice displaying
altered repertoires of MHC class II-bound self-peptides. Immunity
7:197.
Chmielowski, B., Muranski, P., Kisielow, P. and Ignatowicz, L.
2000. On the role of high- and low-abundance class II MHC±
peptide complexes in the thymic positive selection of CD4+ T
cells. Int. Immunol. 12:67.
Iwata, M., Hanaoka, S. and Sato, K. 1991. Rescue of thymocytes
and T cell hybridomas from glucocorticoid-induced apoptosis by
stimulation via the T cell receptor/CD3 complex: a possible in vitro
model for positive selection of the T cell repertoire. Eur. J.
Immunol. 21:643.
Wu, L. C., Tuot, D. S., Lyons, D. S., Garcia, K. C. and Davis, M. M.
2002. Two-step binding mechanism for T-cell receptor recognition
of peptide MHC. Nature 418:552.
Bendelac, A., Rivera, M. N., Park, S. H. and Roark, J. H. 1997.
Mouse CD1-speci®c NK1 T cells: development, speci®city, and
function. Annu. Rev. Immunol. 15:535.
Barton, G. M. and Rudensky, A. Y. 1999. Requirement for diverse,
632 Glucocorticoids in¯uence thymic selection
low-abundance peptides in positive selection of T cells. Science
283:67.
39 Barton, G. M., Beers, C., deRoos, P., Eastman, S. R., Gomez,
M. E., Forbush, K. A. and Rudensky, A. Y. 2002. Positive selection
of self-MHC-reactive T cells by individual peptide±MHC class II
complexes. Proc. Natl Acad. Sci. USA 99:6937.
40 Pacholczyk, R., Kraj, P. and Ignatowicz, L. 2001. An incremental
increase in the complexity of peptides bound to class II MHC
changes the diversity of positively selected alpha beta TCR. J.
Immunol. 166:2357.
41 Pazirandeh, A., Xue, Y., Rafter, I., Sjovall, J., Jondal, M. and Okret,
S. 1999. Paracrine glucocorticoid activity produced by mouse
thymic epithelial cells. FASEB J. 13:893.
42 Cole, T. J., Myles, K., Purton, J. F., Brereton, P. S., Solomon, N. M.,
Godfrey, D. I. and Funder, J. W. 2001. GRKO mice express an
aberrant dexamethasone-binding glucocorticoid receptor, but are
profoundly glucocorticoid resistant. Mol. Cell. Endocrinol.
173:193.