| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Molecular Biology, The Scripps Research Institute, La Jolla, CA 92037; and
Digital Gene Technologies, La Jolla, CA 92037
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
How the lymphocyte population recognizes this homeostatic set point is an unanswered question, though the persistence of naive CD4 T cells in the absence of Ag stimulation appears to depend at least in part on low-level stimulation by APCs (8, 9, 10, 11, 12). We hypothesized that any mechanism for sensing the steady state numbers of lymphocytes would occur primarily within lymphoid tissues and would depend on the detection of local concentrations of a stromal cell factor produced at a fixed rate. Excess amounts of this factor (representing lymphopenia) would be permissive for homeostatic expansion, whereas low levels (representing absorption by a full lymphocyte complement) would limit expansion (13). Thus, homeostatic proliferation of T cells would be driven not by a global ability to detect "space" in the immune system (14, 15), but instead by the ability to measure the local concentrations of this survival or nurturing factor. Recent evidence suggests that CD4 and CD8 T cells may depend on different factors (16). For instance, production of IL-7 and IL-15 by stromal cells has been suggested to regulate the numbers of CD8 T cells but not CD4 T cells (17).
In the context of our hypothesis, we considered the ligands for the
chemokine receptor CCR7 (18, 19, 20, 21, 22, 23, 24) to be strong candidates
for factors regulating homeostatic proliferation of CD4 T cells for the
following reasons. Naive CD4 T cells express the CCR7 receptor
(22, 23). CCR7 has two ligands, CCL21 (also known as TCA4,
SLC, or 6Ckine) and CCL19 (also known as ELC), and these
chemokines are expressed in sites relevant to the regulation of
homeostasis. Specifically, CCL21 and CCL19 are coexpressed at high
constitutive levels by the stromal reticular cells located in the
lymphoid T cell compartments. CCL21 is also expressed by the high
endothelial venules in lymphoid tissue (18, 19, 20, 21). In
addition, transgenic expression of the CCR7 ligand CCL21 within
pancreatic islet 
cells (Insulin promotor-TCA4 transgenic
mice) was sufficient to induce the formation of
lymphoid-like structures within the islets (24). This
accumulation of T lymphocytes around CCL21-expressing islets was not
associated with a decrease in T lymphocyte numbers in other peripheral
lymph nodes or in spleen, suggesting that transgenic expression of
excess CCL21 was able to support the survival of a larger total T
lymphocyte pool (nearly 20% larger). In this study, we manipulate the
levels of CCL21 expression and directly test whether CCL21 availability
can regulate homeostatic proliferation and progression toward
autoimmunity.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Insulin-influenza hemagglutinin (Ins-HA),3 TCR-SFE, and Ins-TCA4 transgenic mice were generated and characterized as previously described (24, 25). All mice were housed in the vivarium at The Scripps Research Institute (La Jolla, CA) in accordance with institutional and National Institutes of Health guidelines.
CFSE labeling and T cell transfer
T cells were isolated from murine spleens as previously described (26). The resulting cell populations comprised >90% Thy-1.2+ cells, as determined by flow cytometry. For CFSE labeling, T cells were resuspended at 1–5 x 107/ml and incubated for 10 min at 37°C with 10 µM CFSE (Molecular Probes, Eugene, OR). Cells were then washed three times with 10 ml of ice-cold PBS. An aliquot from the labeled population was stored for later flow cytometric analysis to define the parameters of the nondividing cell population. Ten million T cells were injected i.v. into recipient mice (8–10 wk old) via the retro-orbital vein. Irradiated recipients were exposed to 600 rad. Proliferation of CFSE-positive cells was analyzed in T cells recovered from spleen, peripheral, and pancreatic lymph nodes using FACSCalibur and CellQuest software (BD Biosciences, San Jose, CA) 1.5, 3, 5, and 7 days post-transfer into recipient mice. In all experiments, recovered T cells were also labeled with anti-CD4 or anti-CD8 Abs (either PE- or allophycocyanin-conjugated). Therefore, selective gating on each T cell population during flow cytometric analysis allowed CD4 T cell proliferation to be examined separately from CD8 T cell proliferation.
Histology
Immunohistochemistry was performed on cryosections (6–10 µm) fixed in ice-cold acetone as previously described (22). Sections were stained with the purified rat anti-mouse primary Abs listed below (BD PharMingen, San Diego, CA), followed by biotin F(ab')2 mouse anti-rat IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) and streptavidin-HRP (The Jackson Laboratory, Bar Harbor, ME). Positive cells were visualized by 3-amino 9-ethylcarbazole substrate (Sigma-Aldrich, St. Louis, MO), and tissue was counterstained with hematoxylin. The following Abs were used: anti-CD4, anti-B220, anti-mucosal addressin cell adhesion molecule-1, anti-peripheral lymph node addressin, and anti-ER-TR7 (Accurate Chemical and Scientific, Westbury, NY).
Flow cytometry
PE-conjugated anti-CD4, allophycocyanin-conjugated
anti-CD8, biotinylated anti-V8.1/8.2, and
streptavidin-CyChrome Abs (BD PharMingen) were used to stain cell
suspensions from spleen and lymph nodes. The activation state of
CFSE-labeled CD4 and CD8 T cells was analyzed by flow cytometry using
PE- and allophycocyanin-conjugated anti-mouse CD4, anti-mouse
CD8, anti-mouse CD62L, and anti-mouse CD44 Abs. CFSE-positive
and immunostained cells were analyzed using FACSCalibur and CellQuest
software. In all experiments, recovered T cells were also labeled with
anti-CD4 or anti-CD8 Abs (either PE- or
allophycocyanin-conjugated). Therefore, selective gating on each T cell
population during flow cytometric analysis allowed CD4 T cell
activation to be examined separately from CD8 T cell activation.
Serological analysis of diabetes
Diabetes was characterized by polyuria/polydipsia, weight loss,
glycosuria as assessed by urine chemstrips (Bayer, Wuppertal, Germany),
and persistent hyperglycemia (>250 mg/dl), as measured with blood
glucose chemstrips (Boehringer Mannheim, Indianapolis, IN) using an
Accu-ChekII monitor (Boehringer Mannheim). Diabetes incidence
curves were compared between animal groups using 2 x 2
contingency tables and 
2 analysis.
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Similarly, within 7 days after transfer into BALB/c mice, expressing
wild-type levels of CCL21 and made lymphocyte deficient by sublethal
irradiation, cell division could be detected in 70% of the CD4
T cells and 80% of the CD8 T cells (Fig. 1
, A and B).
However, the number of cell divisions undergone within 7 days by the
proliferating population of transferred CD8 T cells was greater than
that of the proliferating population of transferred CD4 T cells (Fig. 1C and D). In nonirradiated mice with nondepleted
lymphocyte populations, <20% of the transferred CD4 T cells and
<40% of the transferred CD8 T cells had undergone proliferation
within the same 7-day period (Fig. 1, A and
B).
|
). The
plt mutation is a deletion of both the primary
CCL21 gene expressed in secondary lymphoid organs and
CCL19, the only other known ligand for CCR7 (22, 23, 28, 29, 30). Yet, in stark contrast to that seen in wild-type
irradiation-depleted BALB/c mice, <20% of the transferred
CFSE-labeled CD4 T cells underwent proliferation when transferred into
lymphocyte-deficient plt/plt mice (Fig. 1, A and
C). This low level of proliferation was equivalent to that
seen after transfer into nonirradiated recipients. Because CFSE-labeled
CD8 T cells showed normal homeostatic proliferation in the
plt/plt recipients (Fig. 1B; 1 day), CD4 but not
CD8 T cells appear to be strongly dependent on CCR7 ligand expression
to support their homeostatic proliferation.
Two types of observations suggest that mechanisms regulating
homeostatic expansion of T cells may play a significant role in the
development of autoimmune disease. First, induction of lymphopenia in
rodents is often used to precipitate autoimmunity; second, homeostatic
expansion of CD8 T cells has been associated with the acquisition of a
transient "preactivated" phenotype (31, 32, 33, 34). Thus, to
determine the relative contributions of Ag-specific and -nonspecific
mechanisms toward T cell proliferation, we studied the response of
normal (nontransgenic B10.D2) and TCR-transgenic CD4 T cells in
lymphocyte-depleted and -nondepleted recipients that expressed the
target Ag for the transgenic T cells. In these mice, the HA Ag is
expressed in pancreatic islet cells (Ins-HA mice), and the
TCR-transgenic T cells (TCR-SFE) are specific for the HA peptide, amino
acids 110–119, presented on I-Ed (25, 26, 35, 36).
Upon transfer into normal Ins-HA recipients, nontransgenic B10.D2 T
cells showed minimal homeostatic proliferation compared with those
transferred into T cell-deficient
Ins-HA/RAG-1-/- recipients (Fig. 1, E and F). In contrast, when HA-specific TCR-SFE
cells were transferred to normal Ins-HA mice, a significant Ag-specific
proliferation was evident within 5 days, especially among T cells in
pancreatic lymph nodes (Fig. 1, E and F). Despite
the observed T cell proliferation, diabetes did not develop in these
mice. However, when HA-specific TCR-SFE cells were transferred into
Ins-HA/RAG-1-/- recipients, the percentage of
proliferating cells dramatically increased. Indeed, the number of
divisions per cell was also greatly increased, as evidenced by the
strong shift of CFSE peaks to the left, and these mice developed
diabetes. Homeostatic mechanisms thus appeared to have a major
influence on Ag-specific responses. Although Ag-specific T cells could
proliferate in normal Ag-expressing recipients, the removal of
bystander T cells released the Ag-specific T cells from significant
inhibition.
We also tested whether, in the absence of other tissue sources,
pancreatic expression of CCL21 by itself was sufficient to support
homeostatic proliferation. For these studies, transgenic mice that
expressed high levels of CCL21 (Ins-TCA4) in pancreatic islets were
bred to plt/plt mice (24). Both the transgene
and the plt/plt deletion were maintained on a BALB/c genetic
background. Therefore, homeostatic proliferation of BALB/c T cells was
examined 7 days post-transfer in mice carrying the Ins-TCA4 transgene
and homozygous for the plt/plt mutation
(plt x Ins-TCA4). Although CD4 T cells from BALB/c
mice failed to undergo homeostatic proliferation in
plt/plt mice even after irradiation (Fig. 1, A and B), pancreatic expression of CCL21 was
sufficient to induce homeostatic proliferation of these same cells
within the pancreatic lymph nodes of both lymphopenic and
nonlymphopenic plt x Ins-TCA4 recipients (Fig. 2A). In contrast, CFSE-labeled
CD4 T cells isolated from the other non-CCL21-expressing peripheral
lymph nodes of these same irradiated plt x Ins-TCA4 mice failed
to undergo homeostatic proliferation (data not shown).
|
).
Interestingly, the rate of proliferation was much more rapid in
nonirradiated plt x Ins-TCA4 mice than in irradiated
plt x Ins-TCA4. It is possible that, because
irradiation depletes all lymph node compartments, adoptively
transferred CD4 T cells more readily homed and were more rapidly
exposed to CCL21 in the pancreatic lymph nodes in nonirradiated
plt x TCA4 recipients than in the irradiated
mice.
Because homeostatic proliferation of CD8 T cells has been associated
with the transient acquisition of a preactivated/memory phenotype, we
examined the expression of CD62L as a function of homeostatic CD4 T
cell proliferation (15, 32, 33, 34). As expected, before
transfer the lymph node donor CD4 T cells displayed a naive
CD62Lhigh phenotype (Fig. 2B). Of the
lymphocytes recovered from the pancreatic lymph nodes 7 days
posttransfer (Fig. 2B), both the very small population of
CD4 cells that failed to undergo a single round of division and the
much larger population that underwent a single round of division
displayed a similar naive phenotype (CD62Lhigh).
Of the cells that had undergone between three and five rounds of
division, a variable number displayed a CD62Llow
phenotype (Fig. 2B). However, of the cells that had
undergone seven rounds of division within 7 days, nearly all displayed
a CD62Llow phenotype.
With similar kinetics, CD44 levels were found to shift from low to high
as a function of cell division (data not shown), but this change was
neither as dramatic nor as consistent as that of CD62L. Significantly,
homeostatic proliferation induced by irradiation and/or CCL21 led to
identical T cell phenotypes (Fig. 2B). Although our models
and results differ slightly from those of Clarke and Rudensky
(37), our results are largely consistent. In their C57BL/6
models, they detected two patterns of homeostatic CD4 T cell
proliferation: a large population of CD4 T cells that proliferated very
slowly and a small population that proliferated and acquired a
preactivated/memory phenotype (CD44high,
CD62Llow) with similar kinetics as those examined
in our studies. In both their studies and ours, the preactivated/memory
phenotype induced by homeostatic proliferation was distinct from and
developed more slowly than Ag-induced activation (36, 37).
Altogether our studies indicate that, depending on the degree of lymphopenia and/or the local expression of CCL21, a large population of preactivated cells potentially can be generated by homeostatic mechanisms. Because the physiological relevance of this lymphocyte population was ill defined, we investigated whether CCL21-induced homeostatic proliferation could facilitate the onset of autoimmune disease even in nonlymphopenic mice.
Because the absence of CCR7 ligands limited CD4 T cell homeostatic
proliferation even in a lymphopenic environment, we first tested
whether an excess of CCL21 could overcome the inhibition in nondepleted
recipients. For these experiments, transgenic mice were used that
expressed high levels of either HA (Ins-HA) or CCL21 (Ins-TCA4) in
pancreatic islets (24). As expected, irradiation-depleted
Ins-HA recipients supported strong proliferation of transferred
HA-specific TCR-SFE T cells (Fig. 3),
whereas nondepleted recipients showed a lower Ag-specific response
(Fig. 3). Strikingly, expression of CCL21 in the Ins-HA/Ins-TCA4
double-transgenic mice was sufficient to eliminate the need to induce
lymphopenia. Strong proliferative responses (in both the percentage of
dividing cells and number of cell divisions) were observed in
HA-specific TCR-SFE T cells transferred into nonirradiated
double-transgenic mice (Fig. 3). This was most evident in the spleen
and in pancreatic lymph nodes, where responses were nearly equivalent
to those found after transfer into irradiation-depleted Ins-HA
mice.
|
). However, progression toward end-stage
autoimmune diabetes could only be induced in lymphopenic Ins-HA mice
(Fig. 1, E and F). Therefore, we tested whether,
in the presence of CCL21 overexpression, the host lymphocyte population
still needed to be depleted, perhaps to deplete CD4 regulatory cells.
We find that transfer of HA-specific TCR-SFE T cells into
Ins-HA/Ins-TCA4 recipients caused autoimmune diabetes with the same
kinetics (Fig. 4) as that seen in
lymphopenic recipients expressing wild-type levels of CCL21. Together,
these data suggest that the excess CCL21 was able to replace the
requirement for T cell depletion of the recipient. Thus, the
augmentation of HA-specific responses by the transgene CCL21 appeared
to include the boosting of effector responses.
|
, A–D). As described in
models of spontaneous autoimmune diabetes such as the nonobese diabetic
(NOD) mouse (26, 38) and the Ins-HA/TCR-transgenic mouse
model (25, 26, 35, 36), islet lymphoid infiltrates show a
characteristic organization strikingly similar to normal lymphoid
tissue, and the Ins-HA/Ins-TCA4 mice fit this pattern. Thus, not only
did T cells segregate spontaneously from B cells, but the development
of lymphoid-like stroma (identified by induced expression of ER-TR7 on
islet stromal cells) was induced to support these compartments (Fig. 4, A–D). These studies link two important phenomena, CD4 T cell homeostasis and autoimmunity, by suggesting that a CCR7 chemokine ligand can have a major influence on CD4 T cell homeostatic proliferation and induction of T cell effector responses in vivo. Homeostatic proliferation failed to occur in the absence of both CCR7 ligands (CCL21 and CCL19), even in lymphopenic mice. However, transgenic expression of one CCR7 ligand (CCL21) was sufficient to restore homeostatic proliferation within lymph nodes in close proximity to CCL21 expression. Similarly, in nonlymphopenic mice with undisrupted CCL21 and CCL19 genes, transgenic overexpression of CCL21 was sufficient to induce a 20% increase in the T lymphocyte pool and to promote autoimmunity in the proximity of transgene expression (24). Because CCL19 and CCL21 share the same receptor (CCR7), we would predict that similar effects would have been observed if a CCL19 transgene and not a CCL21 transgene had been expressed.
These effects may help explain a number of well-established observations on the relationship between lymphopenia and susceptibility to autoimmune disease. For example, it has been found that in models of adoptively transferred disease, including autoimmune diabetes in NOD mice (39, 40), inflammatory bowel disease (31), and a number of transgenic mouse models (25, 39), recipients had to be depleted of T cells by irradiation or genetic deficiency (e.g., SCID, RAG-1-/-) (31). Similarly, susceptibility to spontaneous autoimmune disease can be dependent on lymphopenia induced by neonatal thymectomy (41, 42) or a genetic mutation (e.g., the lyp gene in the BB rat model of autoimmune diabetes) (43). Strikingly, spontaneous development of autoimmune diabetes in nonlymphopenic NOD mice is associated with expression of CCL21 in prediabetic mice (38). In most of these cases, CD4 T cells mediate and/or regulate the onset of autoimmune disease. Thus, although lymphoid stromal cell-derived chemokines such as CCL21 may have a primary function in T cell homeostasis, perturbations in their expression can also alter susceptibility to autoimmunity.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Monica J. Carson, Department of Molecular Biology, The Scripps Research Institute, MB10, 10550 North Torrey Pines Road, La Jolla, CA 92037. E-mail address: carson{at}scripps.edu
3 Abbreviations used in this paper: Ins-HA, insulin-influenza hemagglutinin; NOD, nonobese diabetic.
Received for publication June 1, 2001. Accepted for publication October 5, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
-chain locus. Curr. Biol. 8:701.[Medline]
This article has been cited by other articles:
|
|
 |
|
  A. P. Martin, T. Marinkovic, C. Canasto-Chibuque, R. Latif, J. C. Unkeless, T. F. Davies, Y. Takahama, G. C. Furtado, and S. A. Lira CCR7 Deficiency in NOD Mice Leads to Thyroiditis and Primary Hypothyroidism J. Immunol., September 1, 2009; 183(5): 3073 - 3080. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. F. Moxham, J. Karegli, R. E. Phillips, K. L. Brown, T. T. Tapmeier, R. Hangartner, S. H. Sacks, and W. Wong Homeostatic Proliferation of Lymphocytes Results in Augmented Memory-Like Function and Accelerated Allograft Rejection J. Immunol., March 15, 2008; 180(6): 3910 - 3918. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. P. Martin, M. G. Grisotto, C. Canasto-Chibuque, S. L. Kunkel, J. S. Bromberg, G. C. Furtado, and S. A. Lira Islet Expression of M3 Uncovers a Key Role for Chemokines in the Development and Recruitment of Diabetogenic Cells in NOD Mice Diabetes, February 1, 2008; 57(2): 387 - 394. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. P. Martin, J. M. Alexander-Brett, C. Canasto-Chibuque, A. Garin, J. S. Bromberg, D. H. Fremont, and S. A. Lira The Chemokine Binding Protein M3 Prevents Diabetes Induced by Multiple Low Doses of Streptozotocin J. Immunol., April 1, 2007; 178(7): 4623 - 4631. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Sancho, J. M. Vieira, C. Casalou, M. Mesquita, T. Pereira, B. M. Cavaco, S. Dias, and V. Leite Expression and function of the chemokine receptor CCR7 in thyroid carcinomas. J. Endocrinol., October 1, 2006; 191(1): 229 - 238. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. K. Taylor, P. T. Walsh, D. F. LaRosa, J. Zhang, M. A. Burchill, M. A. Farrar, and L. A. Turka Constitutive Activation of STAT5 Supersedes the Requirement for Cytokine and TCR Engagement of CD4+ T Cells in Steady-State Homeostasis J. Immunol., August 15, 2006; 177(4): 2216 - 2223. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Sanchez-Sanchez, L. Riol-Blanco, and J. L. Rodriguez-Fernandez The Multiple Personalities of the Chemokine Receptor CCR7 in Dendritic Cells J. Immunol., May 1, 2006; 176(9): 5153 - 5159. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. Obhrai, M. H. Oberbarnscheidt, T. W. Hand, L. Diggs, G. Chalasani, and F. G. Lakkis Effector T Cell Differentiation and Memory T Cell Maintenance Outside Secondary Lymphoid Organs J. Immunol., April 1, 2006; 176(7): 4051 - 4058. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Su, D. S. Chang, S. S. Gambhir, and J. Braun Monitoring the Antitumor Response of Naive and Memory CD8 T Cells in RAG1-/- Mice by Positron-Emission Tomography J. Immunol., April 1, 2006; 176(7): 4459 - 4467. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Zaft, A. Sapoznikov, R. Krauthgamer, D. R. Littman, and S. Jung CD11chigh Dendritic Cell Ablation Impairs Lymphopenia-Driven Proliferation of Naive and Memory CD8+ T Cells J. Immunol., November 15, 2005; 175(10): 6428 - 6435. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. P Hickman and L. A Turka Homeostatic T cell proliferation as a barrier to T cell tolerance Phil Trans R Soc B, September 29, 2005; 360(1461): 1713 - 1721. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Marleau and N. Sarvetnick T cell homeostasis in tolerance and immunity J. Leukoc. Biol., September 1, 2005; 78(3): 575 - 584. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Huan, S. Subramanian, R. Jones, C. Rich, J. Link, J. Mooney, D. N. Bourdette, A. A. Vandenbark, G. G. Burrows, and H. Offner Monomeric Recombinant TCR Ligand Reduces Relapse Rate and Severity of Experimental Autoimmune Encephalomyelitis in SJL/J Mice through Cytokine Switch J. Immunol., April 1, 2004; 172(7): 4556 - 4566. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Ito, A. Matejuk, C. Hopke, H. Drought, J. Dwyer, A. Zamora, S. Subramanian, A. A. Vandenbark, and H. Offner Transfer of Severe Experimental Autoimmune Encephalomyelitis by IL-12- and IL-18-Potentiated T Cells Is Estrogen Sensitive J. Immunol., May 1, 2003; 170(9): 4802 - 4809. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. W. Christopherson II, A. F. Hood, J. B. Travers, H. Ramsey, and R. A. Hromas Endothelial induction of the T-cell chemokine CCL21 in T-cell autoimmune diseases Blood, February 1, 2003; 101(3): 801 - 806. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), and nonirradiated Ins-HA x Ins-TCA4
(
) recipients after transfer of 1 x 107 splenic T
lymphocytes from TCR-SFE transgenic mice (E). Six mice
were used in each condition.