| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Molecular and Cellular Biology Program,
Department of Immunology, and
Department of Comparative Medicine, University of Washington, Seattle, WA 98195
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
1 in response to activated, but not nonactivated, APCs presenting MBP. This tolerant response depends on continuous activity of regulatory T cells because, in their absence, these uniquely tolerized MBP-specific T cells can again induce autoimmunity. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
activity (7). In contrast to CD4+CD25+Foxp3+ T cells, Tr1 and Th3 regulatory cells differentiate from mature Foxp3– T cells and mediate dominant suppression via production of TGF- and IL-10 (Tr1) or TGF- alone (Th3) (8, 9, 10). Whether regulatory T cells act directly on responding effector T cells or indirectly via inhibitory effects on APCs has been a subject of intense investigation. Most evidence indicates that a direct effect of regulatory T cells on dendritic cells (DCs) is critical to their suppressive activity. Regulatory T cells can dominantly suppress DC maturation (11) and their suppressive activity can be overcome by APC activation in vitro and in vivo (11, 12, 13). Accordingly, stimuli that trigger DC maturation in vivo can initiate autoimmune disease (14, 15). Recently, two-photon laser-scanning microscopy studies have shown that regulatory T cells interact directly with DCs presenting self-Ags in vivo in the absence of effector T cells and that these interactions inhibit subsequent stable interactions between effector T cells and DCs (16, 17). These studies suggest that the ability to prevent DC maturation is central to regulatory T cell-mediated tolerance. However, it is not known how regulatory T cells suppress autoreactive T cell responses when the self-Ag is presented by activated DCs.
Multiple sclerosis (MS) is believed to be an autoimmune disease reflecting a loss of tolerance in myelin-specific T cells (18). To study the tolerance mechanisms that regulate myelin-specific T cells, we previously generated TCR-transgenic mice specific for an MHC class II-restricted epitope of myelin basic protein (MBP), MBP121–140 (19). MBP is a major protein component of CNS myelin and a minor component of peripheral myelin (20). Our studies showed that most MBP121–140-specific thymocytes generated in adult animals are eliminated by bone marrow-derived cells presenting MBP acquired from peripheral myelin (19). However, MBP121–140-specific T cells in young (<3 wk old) mice escape central tolerance (19), reflecting the developmentally regulated expression of MBP that generates only low levels of the protein during the first 2–3 wk of life (21). Transgenic MBP-specific T cells found in the periphery of 4-wk-old mice are capable of responding to endogenous MBP, as administration of pertussis toxin alone induces autoimmune disease (19). However, no spontaneous autoimmunity develops in the transgenic mice, indicating that peripheral tolerance mechanisms suppress MBP121–140-specific T cell responses.
We investigated peripheral tolerance mechanisms that regulate MBP-specific T cells using an experimental system in which TCR-transgenic MBP121–140-specific T cells from MBP-deficient (MBP–/–) mice were adoptively transferred into wild-type MBP+/+ recipients. This approach allowed us to analyze the regulation of naive MBP-specific T cells that were not subjected to central tolerance as they encounter MBP in the periphery. Our results show that CD4+CD25+ regulatory T cells are required to suppress Th1 cytokine production by the transferred MBP121–140-specific T cells and protect recipient mice from autoimmunity. However, activation of APCs in vivo 5 days after transfer of naive T cells abrogated protection from disease. Surprisingly, APC activation 30 days after T cell transfer did not induce disease and instead triggered an increase in transcription of IL-10 and TGF-1 by MBP-specific T cells. Our studies indicate that regulatory T cells maintain long-term peripheral tolerance by inducing naive self-reactive T cells responding to nonactivated APCs to differentiate into a novel state of tolerance in which production of Th1 cytokines is decreased and transcription of suppressive cytokines is increased in response to MBP presented by activated APCs. Both the initial suppression and the subsequent tolerant state of self-reactive T cells require the presence of regulatory T cells, underscoring their importance in maintaining self-tolerance, even under inflammatory conditions.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Wild-type B10.PL(73 NS)/Sn mice purchased from The Jackson Laboratory were maintained in our breeding colony. The TCR
–/– mutation was backcrossed onto the B10.PL background for 10 generations before use. MBP121–140 TCR-transgenic, MBP-deficient shiverer (MBP–/–), RAG-1–/–, and Thy1.1 B10.PL mice have been described previously (19, 22). All mice were bred and maintained in a specific pathogen-free facility and all procedures involving animals were approved by the Institutional Animal Care and Use Committee at the University of Washington (Seattle, WA). Mice in disease induction experiments were euthanized upon losing >20% of their starting weight.
Antibodies
Abs specific for CD4, V2, V8, Thy1.1, Thy1.2, IL-2, IFN-
, TNF-, and IL-4 were obtained from BD Biosciences. Anti-CD40 (clone 1c10) was obtained from R&D Systems. Anti-IL-10R (clone 1B.1.3), anti-TGF- (clone 1D11), and isotype control Ab (clone GL113) were gifts from Dr. S. Stohlman, University of Southern California (Los Angeles, CA). FITC-conjugated annexin-V was purchased from Molecular Probes.
Cytokine neutralization
Neutralizing anti-IL-10R and anti-TGF- (500 µg each) were administered i.p. on days –6, –1, and 4 relative to the day of transfer of MBP-specific T cells. In mice that received the combination of both Abs, or an equal quantity of control IgG, Abs were administered at days 1 and 4 after transgenic T cell transfer.
Magnetic bead cell separation
Magnetic cell separation was performed on an autoMACS (Miltenyi Biotec) using biotin-conjugated Abs specific for CD3, CD25, V2, or Thy1.1 (BD Biosciences) and either streptavidin- or anti-biotin-conjugated microbeads (Miltenyi Biotec).
Histology
CNS sections were fixed and stained with H&E as previously described (23). For H&E staining of peripheral tissue, formalin-fixed mice were decalcified in formic acid (Cal-Rite). Sections were cut through the pelvis, perpendicular to the spine.
Transgenic T cell transfers
Spleen and lymph nodes were harvested from MBP–/– TCR-transgenic mice and an aliquot was stained with Abs specific for V2, V8, and CD4 to determine the percentage of transgenic T cells. Lymphocytes containing 3 x 106 transgenic T cells were injected i.v. into recipient mice, except for Fig. 1B, where lymphocytes were first V2 enriched by magnetic bead selection and then 3 x 105 transgenic T cells were injected per recipient. For retransfer experiments, transgenic T cells were purified from recipients by magnetic bead separation using anti-Thy1.1; 3 x 106 purified T cells were injected i.v. into the second recipients.
|
Mice were perfused and single-cell suspensions were prepared from the CNS and spleen as previously described (24). The percentage of transgenic MBP121–140-specific T cells in the tissues was determined by flow cytometry using Abs against CD4, V2, and V8.
In vivo T cell proliferation
Proliferation in response to endogenous MBP was measured by maintaining mice on drinking water containing 0.8 mg/ml BrdU (Sigma-Aldrich) for 2 wk beginning either 1 wk before transfer of MBP-specific T cells (early recipients) or 3–4 wk posttransfer (late recipients). Splenocytes were stained with Abs specific for CD4 and Thy1 genetic markers, and the FITC BrdU Flow kit (BD Biosciences) was used to determine BrdU incorporation. Transgenic T cells were differentiated from nontransgenic T cells by expression of Thy1.1 and Thy1.2 genetic markers. Proliferation in response to an in vivo MBP peptide pulse was measured following i.v. injection of 0.4 µmoles MBP121–140 or MBPAc1–11 (control) peptide (Genemed Synthesis). A total of 1 mg of BrdU was given i.p. at 0, 24, and 48 h later. Mice were sacrificed 24 h after the final BrdU injection and splenocytes were analyzed as described above.
Intracellular cytokine staining
Cytokine production was measured in response to an in vivo peptide pulse according to a protocol modified from Pape et al. (25). Mice were administered 0.4 µmoles MBP121–140 peptide i.v. and sacrificed 2 h later. Splenocytes were prepared in medium supplemented with GolgiStop (4 µl/6 ml) and GolgiPlug (1 µl/ml) (BD Biosciences) and stained with anti-CD4 and anti-Thy1.1 Abs. Cells were then permeabilized with the BD Cytofix/Cytoperm kit, stained with PE-conjugated Abs specific for IL-2, IFN-, TNF-, IL-4, or isotype controls and analyzed by flow cytometry gating on CD4+Thy1.1+ cells. Positive staining with isotype control Abs was subtracted as background. "N" indicates the number of mice analyzed per group.
Real-time RT-PCR
Genetically marked transgenic T cells were sorted to >90% purity on a FACSAria (BD Biosciences). RNA extraction was performed using either the Absolutely RNA RT-PCR Miniprep kit (Stratagene) or RNAqueous reagents (Ambion). cDNA was prepared using the SuperScript III First-Strand Synthesis kit (Stratagene) with oligo(dT) primers. Real-time RT-PCR was performed with Brilliant SYBR Green QPCR reagents (Stratagene) on a DNA Engine Opticon system (MJ Research). Relative quantities of target genes were determined using relative standard curves as described in ABI User Bulletin No. 2 and are presented as unit quantity normalized to -actin expression. cDNA from Th1- and Th2-skewed splenocyte controls were a gift from Dr. A. Weinmann (University of Washington). The following primers were used: -actin, 5'-GATCTG GCACCACACCTTCT-3', 5'-GGGGTGTTGAAGGTCTCAAA-3'; T-bet (Tbx21), 5'-CAACAACCCCTTTGCCAAAG-3', 5'-TCCCCCAAGCAGTTGACAGT-3'; Foxp3, 5'-TCCTTCCCAGAGTTCTTCCA-3', 5'-AA GTAGGCGAACATGCGAGT-3'; IL-10, 5'-CCTCAGGATGCGGCTGAG-3', 5'-CCACTGCCTTGCTCTTATTTT-3'; TGF-1, 5'-TTGCTTCAGCTCCACAGAGA-3', 5'-TGGTTGTAGAGGGCAAGGAC-3'.
Anti-CD40/LPS administration
Agonistic anti-CD40 (30 µg) and LPS (50 µg, Sigma-Aldrich) were administered i.v. in 200 µl of sterile PBS.
CFSE analysis
Thy1 genetically marked cells from TCR-transgenic MBP–/– mice (naive transgenic T cells), and transgenic T cells purified from day 30 wild-type recipients by magnetic bead separation, were labeled with 4.2 µM CFSE (Molecular Probes) for 20 min. A total of 1 x 106 to 5 x 106 labeled transgenic T cells were transferred to either wild-type, MBP–/–, or day 30 wild-type recipients of transgenic T cells. Splenocytes were analyzed 5 days later gating on CD4+Thy1.1+ or CD4+Thy1.2+ cells.
Statistical analyses
All p values are derived from two-tailed Student’s t tests, with the following exceptions. To compare BrdU incorporation in nontransgenic vs transgenic CD4+ T cells in the same mice, a paired Student’s t test was used (see Fig. 3A). For all real-time RT-PCR data, groups were compared by random permutation analysis.
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
To investigate how naive MBP-specific T cells that were not subjected to central tolerance respond to MBP presented in the periphery, CD4+ T cells isolated from MBP–/– TCR-transgenic mice were transferred into either T cell-deficient or wild-type mice. In both RAG–/– and TCR–/– recipients, the MBP-specific T cells induced severe autoimmune disease characterized by acute weight loss and inflammatory cell infiltrates in the CNS, peripheral nerves, and tissues surrounding peripheral nerves (Fig. 1), consistent with constitutive presentation of MBP in the periphery (19). No signs of disease were observed in wild-type recipients (data not shown). Reconstitution of T cell-deficient mice with bulk splenocytes 7 days before transfer of the MBP-specific T cells protected mice from both histological (Fig. 1A) and clinical signs (data not shown) of disease.
The phenotype of cells that prevented MBP121–140-targeted autoimmunity was determined by transferring different subsets of splenocytes into RAG–/– recipients before transfer of naive MBP121–140-specific T cells. Transfer of purified CD3+ T cells was sufficient to protect mice from disease, however, depletion of CD25+ cells abrogated all protection (Fig. 1B). These data, in conjunction with observed lack of protection by CD4-depleted splenocytes and successful protection by CD8-depleted splenocytes (data not shown), indicate that CD4+CD25+ T cells are required to prevent MBP121–140-specific T cell-mediated autoimmune disease. Interestingly, splenocytes from MBP–/– mice completely protected RAG–/– recipients from disease (Fig. 1B). Thus, regulatory T cells that did not mature in an animal synthesizing MBP can suppress pathogenic MBP-specific T cells.
We investigated whether IL-10 and/or TGF- activity were required in vivo for regulatory T cell-mediated protection. Neither administration of anti-IL-10R nor anti-TGF- Ab alone abrogated protection from disease (Fig. 1C). However, wild-type mice that received anti-IL-10R and anti-TGF- together lost >20% of their body weight 
1 wk after transfer of MBP-specific cells. These mice showed focal lesions in the brain, spinal cord, and periphery, comparable to unprotected recipients. In contrast, control mice that received an equal quantity of isotype-matched Ab did not lose weight and had few to no lesions in the periphery. One control had some lesions in the brain but minimal involvement of the spinal cord.
Regulatory T cells do not prevent expansion of MBP121–140-specific T cells in the spleen or CNS
We previously showed that MBP121–140-specific T cells proliferate in response to endogenous MBP in lymphoid tissues immediately after transfer into wild-type mice (19). To determine whether regulatory T cells prevent the accumulation or trafficking of MBP121–140-specific T cells, we analyzed the number of TCR-transgenic T cells in the spleen and CNS after transfer into either TCR–/– or wild-type mice. The CD4+V2+V8+ T cell number increased significantly in the spleens of wild-type recipients 4 days after transfer compared with the endogenous CD4+ V2+V8+ T cell number in uninjected wild-type mice (p = 0.006, Fig. 2A). The MBP-specific T cell number was higher in T cell-deficient recipients than in wild-type recipients (p = 0.03), although there was more variation in T cell-deficient mice. By 7 days posttransfer, the number of transgenic T cells in the spleen had decreased in both wild-type and T cell-deficient recipients relative to day 4, but still remained above background. Between 7 and 10 days posttransfer, T cell-deficient recipients typically exhibited severe weight loss requiring euthanasia. In contrast, wild-type recipients remained healthy, with no signs of autoimmunity, even though the MBP-specific T cell number in the spleen remained 4-fold above background at 28 days posttransfer.
|
2+V8+ T cells also increased in the CNS of wild-type and T cell-deficient recipients by 4 days posttransfer, indicating that regulatory T cells do not prevent trafficking of MBP-specific T cells to the CNS (Fig. 2B). The number of transgenic T cells in the CNS of wild-type mice was similar at days 4 and 7 posttransfer. At 28 days posttransfer, the number of CD4+V2+V8+ T cells in the CNS of healthy, wild-type recipients was still 10-fold higher than the number in uninjected mice. Thus, elevated numbers of MBP-specific T cells can be sustained in the CNS of mice containing regulatory T cells without inducing disease. MBP-specific T cells persisting in the periphery of MBP+/+ mice are not anergic
The proliferation rate of MBP-specific T cells was measured after transfer into wild-type recipients by maintaining mice on drinking water containing BrdU for 2 wk beginning either 1 wk before ("early recipients") or 3–4 wk after ("long-term recipients") transgenic T cell transfer. Mice were sacrificed after the 2-wk period and BrdU incorporation in transgenic and nontransgenic splenocytes was analyzed. Virtually 100% of the transgenic T cells incorporated BrdU in the early recipients, reflecting the rapid rate of proliferation triggered by their initial exposure to endogenous MBP (Fig. 3A). In contrast, only 20% of the transgenic T cells in long-term recipients were BrdU+, comparable to the percentage of BrdU+ nontransgenic CD4+ splenocytes in the same mice (p = 0.1). The transgenic T cells in long-term recipients exhibited an activated/memory phenotype (CD45RblowCD44high, data not shown), indicating that they had initially responded to self-Ag. Thus, the proliferation of MBP-specific T cells slows dramatically over time in long-term recipients, even though the T cells are exposed to the same amount of endogenous MBP.
To determine whether MBP121–140-specific T cells become unresponsive to Ag, long-term recipients were injected with either MBP121–140 or irrelevant (MBPAc1-11) peptide 3–4 wk after transfer into wild-type mice. Mice were sacrificed 12 h after peptide injection and splenocytes were analyzed for expression of activation markers. The majority of MBP-specific T cells in the MBP+/+ recipients expressed CD25 (69.5%) and CD69 (92.1%) in response to MBP121–140 peptide but not to MBPAc1-11 (2.9% CD25+ and 16.6% CD69+). Analyses of BrdU incorporation over a 72-h period following peptide administration showed that this single dose of MBP121–140 peptide triggered approximately half of the MBP-specific T cells residing in long-term MBP+/+ recipients to proliferate (Fig. 3B). The percentage of BrdU+ transgenic T cells in MBP–/– recipients following a MBP121–140 peptide pulse was somewhat higher, suggesting that long-term exposure to endogenous MBP in the MBP+/+ recipients may impair the proliferative response of some of the MBP-specific T cells. Nevertheless, at least half of the transgenic T cells were not anergic in MBP+/+ hosts.
Regulatory T cells suppress Th1 cytokine production by MBP-specific T cells
The fact that T cell-deficient recipients of MBP121–140-specific T cells succumb to autoimmune disease within 7–10 days indicates that regulatory T cell activity is required shortly after transfer. Because regulatory T cells do not prevent proliferation, expansion, or trafficking of the MBP-specific T cells, we analyzed cytokine production by MBP-specific T cells in response to an in vivo MBP peptide pulse 6 days after transfer into either wild-type or RAG–/– recipients (Fig. 4A). In contrast to assays involving in vitro restimulation, this method allows assessment of the T cell response triggered in the in vivo environment. MBP-specific transgenic T cells transferred into RAG–/– mice exhibited a Th1 phenotype in which most transgenic T cells produced IFN- and some produced IL-2 and TNF-. In contrast, IFN- production was strongly suppressed in transgenic T cells transferred into wild-type recipients. The percentages of transgenic T cells producing TNF- and IL-2 were also reduced in wild-type recipients compared with RAG–/– recipients. We have previously shown that naive B cells in wild-type mice present endogenous MBP (22), raising the possibility that interactions with B cells may contribute to the decrease in the percentage of Th1 cytokine-producing MBP-specific T cells seen in wild-type vs RAG–/– recipients. This does not appear to be the case, however, as no differences in cytokine production were observed when MBP-specific T cells were transferred into unmanipulated RAG–/– recipients compared with RAG–/– recipients that were reconstituted with highly purified B cells 7 days before transgenic T cell transfer (data not shown). No IL-4 production was detected in any recipients, indicating that the decrease in Th1 cytokine production in wild-type recipients did not reflect a shift to a Th2 effector phenotype. Cytokines produced by MBP-specific T cells transferred into MBP–/– mice reflect their naive phenotype in that low levels of TNF- and IL-2 but no IFN- are produced (Fig. 4A). Compared with MBP–/– recipients, a lower percentage of transgenic T cells in MBP+/+ recipients produced TNF- and IL-2 and a higher percentage produced IFN-, demonstrating that regulation of the transgenic T cells in MBP+/+ hosts does not simply maintain a naive phenotype. No cytokine production was observed in mice injected with MBPAc1-11 (data not shown). To determine whether the lymphopenic environment in RAG–/– recipients affected the cytokine profile of MBP-specific T cells, we analyzed cytokines produced by transgenic T cells after transfer into RAG–/–MBP–/– mice. The percentage of TNF-+ and IL-2+ MBP-specific T cells in RAG–/–MBP–/– recipients was comparable to the percentages seen in both RAG–/– and MBP–/– recipients. A higher percentage of transgenic T cells produced IFN- in RAG–/–MBP–/– compared with RAG+/+MBP–/– mice, indicating that homeostatic expansion influences production of IFN-. However, the percentage of IFN--producing transgenic T cells in RAG–/–MBP–/– recipients is dramatically reduced compared with RAG–/–MBP+/+ recipients (12.6 vs 75.2%, respectively), demonstrating that the high percentage of IFN--producing T cells in RAG–/–MBP+/+ recipients reflects an Ag-specific response to endogenous MBP.
|
production (26). RNA isolated from transgenic T cells purified from RAG–/– recipients (6 days posttransfer), wild-type recipients (6 and 30 days posttransfer) and a MBP–/– TCR-transgenic mouse was analyzed by real-time RT-PCR. Surprisingly, T-bet expression by MBP-specific T cells was increased in all recipient mice compared with naive transgenic T cells (Fig. 4B). No difference in T-bet expression was seen in MBP-specific T cells isolated from RAG–/– vs wild-type recipients 6 days posttransfer (p = 0.834), despite the fact that IFN- production is strongly suppressed at this time point in wild-type recipients. T-bet expression was actually elevated in MBP-specific T cells residing in wild-type recipients for 30 compared with 6 days (p = 0.034). These data indicate that the suppression of IFN- in MBP-specific T cells transferred into wild-type mice occurs at a point down-stream of T-bet expression. APC activation in vivo abrogates protection at 5 but not at 30 days after MBP-specific T cell transfer
To determine the importance of preventing APC maturation in this model of autoimmunity, we administered agonistic anti-CD40 Ab and LPS (anti-CD40/LPS) to wild-type recipients at both 5 and 30 days after transfer of MBP-specific T cells, as well as to control mice that had not received transgenic T cells. All mice exhibited an initial weight loss in the first 2 days after injection of anti-CD40/LPS (10–15% of starting weight). However, wild-type mice injected with anti-CD40/LPS 5 days after MBP-specific T cell transfer continued to lose weight and exhibited neurological symptoms (Fig. 5A). Histological analyses revealed inflammatory infiltrates in the CNS and in tissues within and surrounding peripheral nerves (data not shown). Administration of either anti-CD40 Ab or LPS alone at day 5 posttransfer also induced disease (data not shown). In contrast, both control mice (data not shown) and mice that received MBP-specific T cells 30 days earlier recovered their initial weight loss and exhibited no clinical signs up to 30 days after anti-CD40/LPS injection (Fig. 5A).
|
Day 30 transgenic T cells in long-term MBP+/+ recipients are hyporesponsive when triggered by activated APCs
To investigate why anti-CD40/LPS abrogates protection when administered early but not late after transfer of MBP-specific T cells, we analyzed the cytokine responses of transgenic T cells in day 6 and day 30 recipients after anti-CD40/LPS administration. Anti-CD40/LPS was administered 4 days after transfer of MBP-specific T cells and cytokine production in response to an in vivo MBP peptide pulse was determined 2 days later (Fig. 5C). The percentage of transgenic T cells producing IFN- increased dramatically in recipients that received anti-CD40/LPS compared with those that did not (56.0 and 13.6%, respectively). A small but significant increase was also seen in the percentage of IL-2-producing transgenic T cells after anti-CD40/LPS administration. Surprisingly, the opposite result was observed in day 30 recipients. A smaller percentage of transgenic T cells in day 30 recipients produced IFN-, IL-2, and TNF- after receiving anti-CD40/LPS compared with day 30 recipients in which APCs were not activated (Fig. 5D). Thus, while APC activation triggers increased Th1 cytokine production by MBP-specific T cells 5 days after transfer into MBP+/+ mice, there is less Th1 cytokine production by MBP-specific T cells in day 30 recipients when APCs are activated than when they are not activated. The absolute number of transgenic T cells present in day 30 recipients that had received anti-CD40/LPS was equivalent to the number in untreated mice. Furthermore, no increase in apoptosis, as detected by annexin V staining, was observed for transgenic cells in anti-CD40/LPS-treated vs untreated mice. Therefore, the lower percentage of Th1 cytokine-producing transgenic T cells in anti-CD40/LPS-treated day 30 recipients does not reflect increased susceptibility to activation-induced cell death.
A possible explanation for these findings is that APCs presenting endogenous MBP during the initial regulation of MBP-specific T cells are altered such that, upon subsequent activation, they exert a suppressive effect on MBP-specific T cells. In this case, cytokine production by all MBP-specific T cells in day 5 plus day 30 recipients should be decreased when anti-CD40/LPS is administered compared with untreated recipients. Activated APCs in day 5 plus day 30 recipients triggered a smaller percentage of the original (day 30) MBP-specific T cells to produce Th1 cytokines in response to an in vivo MBP peptide pulse (Fig. 5E), as was observed in day 30 recipients that did not receive a second dose of MBP-specific T cells. However, the second dose of MBP-specific T cells (day 5) in day 5 plus day 30 recipients increased IL-2, IFN-, and TNF- production in response to the same activated APCs (Fig. 5F). Thus, the APCs in mice that have regulated MBP-specific T cells have not acquired a dominant suppressive phenotype, rather the day 30 T cells appear to have altered their response to MBP presented by activated APCs, compared with the day 5 T cells. Although day 5 T cells increased Th1 cytokine production in response to activated APCs in these day 5 plus day 30 recipients, they produced less IFN- after anti-CD40/LPS injection in the presence of day 30 MBP-specific T cells than they did in naive recipients (Fig. 5, F vs C), which may account for the lack of disease observed in day 5 plus day 30 mice that received APC-activating stimuli.
MBP-specific T cells in long-term recipient mice express suppressive cytokines in response to APC activation but do not become Foxp3+
To determine whether MBP-specific T cells residing in wild-type recipients for 30 days acquired a regulatory phenotype, the expression levels of Foxp3, IL-10, and TGF-1 were analyzed in MBP-specific T cells before and after transfer into MBP+/+ recipients. A low level of Foxp3 expression was detected in transgenic T cells before transfer (Fig. 6A, naive transgenic), which presumably reflects the presence of a small number of CD25+ T cells in RAG+/+MBP–/– TCR-transgenic mice (data not shown). RAG–/–MBP–/– TCR-transgenic mice could not be used as T cell donors in our experiments because MBP–/–RAG–/– mice have poor viability on the B10.PL background (data not shown). However, the Foxp3+ T cells present in the T cell population isolated from MBP–/–RAG+/+ TCR-transgenic mice were not sufficient to protect T cell-deficient adoptive transfer recipients from autoimmune disease (Fig. 1). No increase in Foxp3 expression was observed in MBP-specific transgenic T cells 6 days after transfer into RAG–/– or wild-type recipients. Foxp3 expression was also not increased in MBP-specific T cells 30 days after transfer into wild-type mice, with or without administration of anti-CD40/LPS (Fig. 6A). These data indicate that MBP-specific T cells in long-term recipients do not convert to a Foxp3+ phenotype, nor is there a selective outgrowth of Foxp3+ T cells that were initially transferred. Interestingly, APC activation in long-term recipients induced MBP-specific T cells to respond to a MBP peptide pulse by expressing increased levels of both IL-10 and TGF-1 mRNA (Fig. 6, B and C), indicating that these T cells have altered their programmed effector response to Ag presented by activated APCs without becoming Foxp3+.
|
The unusual response of day 30 but not day 5 MBP-specific T cells to activated APCs suggested that some intrinsic properties of these T cells may have changed as a result of regulation in wild-type recipients. To investigate this possibility, MBP-specific T cells were isolated from day 30 recipients and retransferred into new wild-type mice. The retransferred T cells proliferated in the new recipients (Fig. 7A), although the rate of proliferation was somewhat slower than that of naive transgenic T cells transferred into wild-type mice (Fig. 5B), suggesting that MBP-specific T cells in long-term recipients have adapted to their environment by decreasing their responsiveness to endogenous MBP over time. Interestingly, transgenic T cells isolated from day 30 recipients and retransferred into different day 30 recipients did not divide after 5 days in the new recipients, even though the amount of endogenous MBP epitopes is presumably the same.
|
Day 30 MBP121–140-specific T cells regain pathogenicity in the absence of regulatory T cells
In light of the intrinsic ability of day 30 MBP-specific T cells to suppress Th1 cytokines in response to activated APCs, we investigated whether the pathogenic potential of day 30 T cells was permanently silenced. As shown in Fig. 8, MBP-specific T cells isolated from day 30 wild-type recipients and retransferred into RAG–/– mice induced autoimmune disease. The number of MBP-specific T cells retransferred in these experiments (3 x 105) represents 14% of the average number of MBP-specific T cells persisting in the spleen of day 30 recipients. The onset of disease induced by retransfer of day 30 T cells into RAG–/– mice was delayed by 2–3 days compared with disease induced by transfer of naive MBP-specific T cells. This delay may reflect the slower proliferation observed for day 30 T cells retransferred into wild-type mice compared with naive T cells (Fig. 5B). Thus, the tolerance exhibited by MBP-specific T cells in day 30 recipients is a dynamic, rather than terminally differentiated, phenotype.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Regulatory T cells did not prevent the expansion of MBP-specific T cells in lymphoid tissues or their trafficking into the CNS. The greater expansion of MBP-specific T cells in the periphery and CNS of T cell-deficient mice may reflect combined effects of lymphopenic conditions and/or the inflammatory milieu associated with the onset of autoimmunity, in addition to the Ag-specific proliferation that occurs in both T cell-deficient and wild-type recipients. Interestingly, the MBP-specific T cell number in the CNS of wild-type mice remained as high at 28 days posttransfer as at 7 days posttransfer, even though these mice remain healthy. MBP-specific T cells in the CNS of wild-type recipients down-regulated the expression of the transgenic TCR (data not shown), consistent with recognition of their cognate Ag in situ. Thus, the population of transgenic T cells persisting in the CNS 28 days after transfer into wild-type mice could be due to retention of Ag-specific T cells in the target organ that has the highest concentration of MBP.
The most striking effect of regulatory T cells on MBP121–140-specific T cells during their initial encounter with endogenous MBP is suppression of their Th1 cytokine production, particularly IFN-. The majority of MBP-specific T cells differentiate into IFN--producing cells in RAG–/– recipients due to interaction with endogenous MBP and not due to lymphopenia-induced expansion (Fig. 3). Regulatory T cells strongly suppressed this Th1 response without inducing a shift toward a Th2 phenotype. Suppression of IFN- production by self-reactive T cells in vivo is consistent with some (29, 30) but not all (31, 32) previous studies of the in vivo effects of regulatory T cells on effector cytokine production. In a diabetes model, Sarween et al. (29) found that regulatory T cells inhibited both IFN- production by effector T cells and target organ infiltration, which was attributed to lack of IFN--mediated up-regulation of CXCR3 expression. In our model, MBP-specific T cells infiltrate the CNS, consistent with the observation that CXCR3 is not required for trafficking across the blood brain barrier (33). DiPaolo et al. (30) also found that regulatory T cells suppress IFN- production by effector T cells without inhibiting Ag-specific proliferation or target organ infiltration in an animal model of autoimmune gastritis. However, IFN- suppression was associated with down-regulation of T-bet expression in this model. Surprisingly, we found that the level of T-bet mRNA is as high in MBP-specific T cells in wild-type recipients as it is in MBP-specific T cells in RAG–/– recipients, which produce high levels of IFN-. These results demonstrate that regulatory T cells can mediate down-regulation of IFN- expression independent of changes in T-bet transcription. A recent study that examined patterns of gene expression in MBP-specific regulatory T cells generated via chronic peptide administration found elevated levels of T-bet expressed in these cells, despite reduced IFN- production (34). This study suggested that T-bet expression was required to suppress IL-2 production in tolerized T cells. Although we have not examined the function of T-bet in MBP-specific T cells undergoing tolerance, our data are consistent with this hypothesis.
This initial suppression of Th1 cytokines by regulatory T cells prevents pathogenicity while the MBP121–140-specific T cells appear to adapt to the levels of endogenous MBP by slowing proliferation without becoming anergic (Fig. 3). Our results suggest that the proliferative rate of MBP-specific T cells reflects the amount of perceived cognate Ag when presented by nonactivated APCs. Day 30 T cells proliferate more rapidly after retransfer into naive recipients than after retransfer into day 30 recipients, perhaps because the number of retransferred T cells is smaller than the number of resident MBP-specific T cells in day 30 recipients, resulting in less competition for endogenous MBP and thus a functionally higher level of available Ag. Alternatively, the context of Ag presentation may differ in day 30 recipients vs naive mice, such that day 30 mice are less conducive to the expansion of Ag-experienced, MBP-specific T cells (a scenario originally suggested by Tanchot et al. (35) in their studies of adaptive T cell tolerance). Although day 30 T cells increased proliferation after retransfer into naive wild-type mice, the rate was slower than the rate of proliferation observed for naive MBP-specific T cells transferred into wild-type recipients (compare Figs. 5B to 7A). This result indicates that the day 30 MBP-specific T cells have become intrinsically less responsive to endogenous MBP than naive MBP121–140-specific T cells. A similar system of adaptive tolerance has been described before in the absence of regulatory T cells, suggesting that adaptation may be a consequence of persistent antigenic stimulation (35). However, regulatory T cells are required in our system to prevent the initial onset of autoimmune disease and allow adaptive tolerance to occur.
Epidemiological studies suggest that an infectious agent is involved in the pathogenesis of MS (36, 37, 38). Therefore, we analyzed the ability of regulatory T cells to suppress MBP121–140-specific T cells when they encounter endogenous MBP on activated APCs, as might occur during an infection. The inability of regulatory T cells to prevent disease when APCs were activated 5 days after MBP-specific T cell transfer is consistent with the notion that suppression by regulatory T cells may depend on preventing DC maturation. Unexpectedly, APC activation 30 days after MBP-specific T cell transfer did not trigger disease, even though the T cells are not anergic at this later time point. Investigation of the basis for this lack of pathogenicity led to the discovery that day 30 MBP-specific T cells have differentiated into a novel phenotype in which they express elevated levels of IL-10 and TGF-1 RNA and produce lower levels of Th1 cytokines in response to activated, but not nonactivated, APCs presenting MBP. This pattern of cytokine expression represents a signature for MBP-specific T cells that have adapted to an environment containing endogenous MBP. Day 30 MBP-specific T cells did not increase expression of Foxp3, in contrast to another study in which Ag-specific T cells persisting in mice recovering from a systemic autoimmune disease became Foxp3+ (6). MBP-specific T cells in day 30 recipients instead resemble Tr1 regulatory T cells that are induced to express TGF-1 and IL-10 in the absence of Foxp3 (39, 40). Because experiments that neutralize the activity of IL-10 cannot be conducted in mice that receive LPS in vivo (41), we could not investigate a functional role for TGF- and IL-10 production by day 30 T cells in suppressing their own pathogenicity in response to APC activation. Interestingly, unlike Tr1 cells, day 30 MBP-specific T cells do not represent a terminally differentiated phenotype. Instead, their tolerant state depends on the continuous presence of regulatory T cells, as illustrated by the ability of day 30 T cells to mediate autoimmune disease when transferred into regulatory T cell-deficient hosts.
Further insights into the long-term tolerant state of MBP-specific T cells were obtained using day 5 plus day 30 MBP-specific T cell recipients. These experiments showed that the unusual suppressive response of day 30 MBP-specific T cells to activated APCs was an intrinsic property that MBP-specific T cells acquired over time, because day 5 MBP-specific T cells increased production of IFN- and IL-2 after anti-CD40/LPS administration, while day 30 T cells decreased production of these cytokines in response to the same activated APCs (Fig. 5, E and F). Retransfer of day 30 MBP-specific T cells into naive wild-type mice followed by anti-CD40/LPS treatment confirmed that the suppressive response to APC activation by day 30 MBP-specific T cells was T cell intrinsic. This result suggested that the suppressive response of the day 30 T cells may represent a form of infectious tolerance that could explain the lack of disease observed in day 5 plus day 30 recipients following APC activation. However, our results do not support this hypothesis as day 30 T cells retransferred into new wild-type recipients did not prevent disease when a second set of naive MBP-specific T cells was injected and the APCs were activated. Thus, the protection from autoimmunity following APC activation in the original day 5 plus day 30 recipients must depend in part on as yet unidentified changes in the host environment that occur during regulation of the initial set of MBP-specific T cells.
Our results provide a mechanistic explanation for how regulatory T cells can suppress autoreactive T cells under both noninflammatory and inflammatory conditions. Typically, self-reactive T cells that escape central tolerance will encounter their Ag in the periphery on nonactivated APCs. If the avidity of this interaction is sufficiently high to trigger proliferation, regulatory T cells prevent the self-reactive T cells from differentiating into Th1 effectors and allow them to become tolerant to the level of endogenous Ag presented. At the same time, regulatory T cells "mark" these T cells as self-reactive by inducing a phenotype that is hyporesponsive with respect to Th1 cytokines and associated with transcription of suppressive cytokines when Ag is encountered on activated APCs. This tolerant state allows regulatory T cells to retain control specifically over T cells that have previously demonstrated self-reactivity even when they subsequently encounter their Ag under inflammatory conditions. Although this system has many advantages, it leaves the organism vulnerable to autoimmunity mediated by T cells that had been previously tolerized if regulatory T cell numbers decrease or their activity is impaired.
|
Acknowledgments |
|---|
Abs, Molly Ottele, Hannah Simkins, and Neal Mausolf for expert technical assistance, and Ingunn Stromnes and Qingyong Ji for critical reading of this manuscript. |
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported by grants to J.G. from the National Institutes of Health (NS043417) and from the National Multiple Sclerosis Society (3328-A-4). S.E.C. was supported in part by Public Health Service National Research Service Award T32 GM07270 from the National Institute of General Medical Sciences. E.S.H. was supported by a Samuel and Althea Stroum Endowed Diabetes Fellowship. 
2 Current address: Department of Immunology, Howard Hughes Medical Institute, National Jewish Medical Center, Denver, CO.
3 Address correspondence and reprint requests to Dr. Joan Goverman, Department of Immunology, University of Washington, Box 357650, Seattle, WA 98195-7650. E-mail address: goverman{at}u.washington.edu
4 Abbreviations used in this paper: Tr1, type I regulatory T cell; DC, dendritic cell; MBP, myelin basic protein; MS, multiple sclerosis.
Received for publication August 4, 2006. Accepted for publication October 30, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
-secreting Th3 regulatory cells. Immunol. Rev. 182: 207-214. [Medline]
) and TCR
) mice were injected with 3 x 106 transgenic T cells. At the indicated times posttransfer, recipients were perfused and cells collected from the spleen (A) and CNS (B). Percentages of CD4+V
). For splenocytes, the average number was 5.45 x 105. *, Statistically significant differences (p 
0.01) compared with uninjected mice.
), RAG–/–MBP–/– (
), or RAG+/+MBP–/– (
) recipients. Cytokine production by genetically marked transgenic splenocytes was measured directly ex vivo 6 days after transfer. ND indicates IL-4 was not determined for RAG–/–MBP–/– recipients. B, T-bet expression by transgenic T cells purified from the indicated recipients was measured by real-time RT-PCR. Naive transgenic T cells were obtained from MBP–/– mice. T-bet expression is normalized to 
) after transfer of 3 x 106 transgenic T cells and weight loss was monitored after anti-CD40/LPS injection. Anti-CD40/LPS was also injected into day 5 plus day 30 recipients 5 days after the second injection of transgenic T cells; •, day 5 plus day 30 recipients that did not develop disease (n = 10); 
