In VivoMechanisms of Acquired Thymic Tolerance

CELLULAR IMMUNOLOGY ARTICLE NO.

179, 165–173 (1997)

CI971165

In Vivo Mechanisms of Acquired Thymic Tolerance1 Wanjun Chen,* Shohreh Issazadeh,* Mohamed H. Sayegh,† and Samia J. Khoury* *M.S. Unit, Center for Neurologic Diseases, and †Laboratory of Immunogenetics and Transplantation, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115 Received February 11, 1997; accepted June 13, 1997

Injection of antigen into the thymus of adult animals induces specific systemic tolerance, but the mechanisms of acquired thymic tolerance are not well understood. To investigate these mechanisms we used a model of intrathymic injection of ovalbumin (OVA) in BALB/c mice. We show an antigen-specific decrease in proliferative responses to OVA, as well as a significant decrease in antigen-specific IL-2 secretion and IFNg production by splenocytes and lymph node cells of tolerant mice. Addition of recombinant IL-2 in vitro reversed the defect in IFN-g production by cells from OVA-tolerized animals, but did not reverse the proliferation or IL-2 production defects. By using an adoptive transfer system, where a small population of OVA peptide-specific CD4/ TCR transgenic T cells are transferred into syngeneic normal recipients, we show an absence of peripheral antigen-dependent clonal expansion of transferred CD4/ TCR transgenic cells in tolerant mice in vivo. There was an increase in clonotype-positive T cells in the thymus after immunization, confirming that activated T cells circulate through the thymus. Furthermore, thymectomy after intrathymic injection abrogates the effect of acquired thymic tolerance and restores antigen-dependent clonal expansion in vivo. We conclude that intrathymic injection of antigen induces Th1 cell unresponsiveness and prevents the peripheral expansion of antigen-specific CD4/ T cells in vivo. This is the first demonstration that in acquired thymic tolerance antigen-specific T cells circulate to the thymus where they may be anergized or ultimately deleted. q 1997 Academic Press

transplantation (2–5) as well as autoimmune (6–10) models, respectively. We have previously shown that injection of guinea pig myelin basic protein (MBP) or its encephalitogenic peptide (p71–90) into the thymus induces systemic tolerance and prevents the development of experimental autoimmune encephalomyelitis (EAE) in the Lewis rat model (9). We have also shown that neither perithymic, intrasplenic, nor intravenous injection of the same amount of antigen can reproduce the effects of intrathymic injection (9, 11, 12) and that acquired thymic tolerance is antigen specific (9, 11) and restricted to the immunodominant peptide of the antigen (9, 13). Immunohistologically, there was marked reduction of mononuclear cell infiltrates and absence of activation and inflammatory cytokines in the brains of intrathymically tolerized animals (9). More recently, we have demonstrated that thymic dendritic-enriched cells mediate the induction of acquired thymic tolerance in this model (12). A potential site for dendritic cell interaction with antigen-specific T cells could be the thymus where activated T cells can circulate (14, 15). However, it is still unclear how IT injection of antigen affects T cell function in vivo and where this interaction takes place. To investigate these mechanisms we used a model of IT injection of OVA in BALB/ c mice and an adoptive transfer system where small numbers of TCR transgenic T cells are transferred into syngeneic normal recipients (16, 17). Our data show that IT injection of antigen results in specific Th1 cell tolerance and inhibits peripheral expansion of antigenspecific CD4/ T cells in vivo. Our data also suggest that the antigen-specific T cells circulate to the thymus where they may be anergized or ultimately deleted.

INTRODUCTION The thymus plays a major role in development of self-tolerance (1), and recent evidence indicates that the thymus plays an important role in acquired tolerance. Intrathymic (IT)2 injection of alloantigen or autoantigens induces specific tolerance in experimental 1

Supported by research grants from the National Multiple Sclerosis Society (RG-2589-A-2) and NIH (R29 AI-34965-03). 2 Abbreviations used: IT, intrathymic; OVA, ovalbumin; HEL, hen egg lyzozyme.

MATERIALS AND METHODS Animals. BALB/c mice (The Jackson Laboratory, Bar Harbor, ME), aged 6–8 weeks, were maintained in our animal facility and used for all experiments. The DO11.10 TCR transgenic mice extensively backcrossed onto BALB/c background were a gift from Dr. D. Y. Loh (18) and were bred and screened for expression of the TCR transgenes in our pathogen-free animal facility, as previously described (16).

165

AID

CI 1165

/

6c26$$$161

08-11-97 08:25:21

cial

0008-8749/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.

166

CHEN ET AL.

Intrathymic injection. BALB/c mice were injected intrathymically with 100 mg of chicken ovalbumin (OVA) or 100 mg hen egg lyzozyme (HEL) (Sigma, St. Louis, MO) dissolved in 50 ml of sterile PBS (Biowhittaker, Walkersiville, MD) as previously described (9). Intrathymic injection was done 48 hr before immunization. Each experimental group consisted of three to five mice. Thymectomy. BALB/c mice were anesthetized and thymectomized through a strenal incision. Care was taken to remove all thymic tissue. The day after thymectomy, each BALB/c mouse received 5 1 106 CD4/ KJ 1-26/ cells, as described below. The mice were then immunized with OVA/CFA. Immunization. BALB/c mice were immunized in the footpads and axillae with 100 ml of OVA/CFA containing 2 mg of OVA in PBS emulsified with 2 mg/ml of CFA (Difco, Detroit, MI). On Days 12–14 postimmunization the draining lymph nodes and spleen were removed aseptically and meshed carefully to prepare single-cell suspensions. The cells from each group of mice were pooled for the proliferation and cytokine experiments but cells from individual mice were used for FACS staining. Cell cultures. The cell suspensions were washed twice before resuspending in DMEM (Biowhittaker) supplemented with 10% (v/v) heat-inactivated fetal calf serum, 2 mM glutamine, 15 mM Hepes, 1% nonessential amino acids, 1 mM sodium pyruvate (all from Biowhittaker), penicillin (100 u/ml), streptomycin (100 mg/ml), and 20 mM 2-ME (Sigma). Red blood cells were lysed with ACK lysing buffer (0.15 M NH4Cl, 10 mM KHCO3 , and 0.1 mM Na2 EDTA), and 4 1 105 cells in 200 ml per well were cultured in round-bottom microtiter plates (Costar, Cambridge, MA) and stimulated with OVA, mycobacterium tuberculosis (MT) (Difco), or anti-mouse CD3 monoclonal antibody (2.5 mg/ml) (Pharmingen, San Diego, CA). Proliferation was measured by the standard 72-hr lymphocyte proliferation assay. For cytokine production, the cells were cultured in X-vivo serum-free medium (Biowhittaker). Cell-free supernatants were collected after 48 hr for measurement of IL-2, IL-4, IFN-g, and IL-10 production. Results of proliferative and cytokine studies were similar for both splenocytes and lymphocytes from draining lymph node cells, and, therefore, only data for splenocytes will be shown. In some experiments recombinant IL-2 (Boeringer-Mannheim, Indianapolis, IN) was added at a concentration of 5 u/ml.

Adoptive transfer and flow cytometry studies. BALB/ c mice were injected intrathymically with OVA (100 mg) or HEL (100 mg) on Day 02. OVA TCR transgenic cells were obtained from the spleen of DO11.10 mice aged 6– 8 weeks and injected intraperitoneally in the BALB/c recipients on Day 01, as previously described (16, 17). Each BALB/c mouse received 5 1 106 CD4/ KJ 1-26/ cells. The mice were immunized with OVA/CFA on Day 0; some mice were used without immunization. Five days later, thymus, popliteal, and axillary lymph node cells were harvested and counted. In some experiments the lymph nodes and thymus were harvested on Day 2 or Day 10 postimmunization. To determine the percentage of KJ 1-26/ cells, 5 1 105 thymocytes or lymph node cells were resuspended in cold PBS / 1% BSA (Irvine Scientific, Santa Ana, CA) and 0.02% sodium azide and then were incubated with KJ 1-26 monoclonal antibody (19) (mouse IgG2a, a kind gift of Dr. D. Y. Loh) for 30 min on ice. The cells were washed twice and then incubated with FITC-labeled goat anti-mouse IgG (H / L) (Caltag, San Francisco, CA) for 30 min on ice; negative control tubes were stained with goat anti-mouse IgG only. Cells were washed and then incubated with CD4-PE monoclonal antibody (Caltag) for an additional 30 min. After several washes, 10,000 to 50,000 events were collected on a Becton–Dickinson FACS sorter and analyzed using Lysis II software. The CD4/ cells were gated and the percentage of KJ 1-26/ cells was calculated. Significance was calculated using a two-tailed Student’s t test. In some experiments thymocytes were treated with antiCD8 mAb and rabbit complement (Accurate Chemical & Scientific Corp., Westbury, NY) before staining. Significance was calculated using a two-tailed Student’s t test. mAbs. The following mAbs were used: anti-CD3e (hamster IgG, Pharmingen); anti-CD4-PE mAb (Caltag); mAb KJ 1-26 (19) (mouse IgG2a, ascites); and purified anti-mouse CD8a (rat IgG2a, clone 53-6.7; Pharmingen). Statistical analysis. For comparison of proliferative responses and cytokine production between experimental groups across three to four experiments, we analyzed the percentage difference between experimental groups at each dose using the MANOVA test for repeated measures, where the repeated measures were the different experiments. RESULTS IT Injection of OVA Induces Antigen-Specific Th1 Cell Unresponsiveness in Vivo

ELISA of cytokines. Quantitative ELISA for IL-2, IL-4, IL-10, and IFN-g was performed using paired monoclonal antibodies specific for the corresponding cytokine according to manufacturer’s recommendations (Pharmingen). Standard curves were generated using known amounts of purified recombinant murine IL-2, IL-4, IL-10, or IFN-g (Pharmingen).

Spleens from BALB/c mice immunized with OVA/ CFA were collected on Day 14 postimmunization and tested for response to stimulation with OVA or MT in vitro. As seen in Fig. 1A, cells recovered from spleens of mice injected intrathymically with OVA had significantly lower proliferative responses and decreased IL-

AID

cial

CI 1165

/

6c26$$$162

08-11-97 08:25:21

MECHANISMS OF ACQUIRED THYMIC TOLERANCE

167

FIG. 1. Effects of IT injection of OVA on T cell responses in vitro. BALB/c mice were intrathymically injected with OVA (open symbols) or HEL (closed symbols) and then immunized with OVA/CFA 48 hr later. Splenocytes were obtained on Day 14 after immunization. Cells were pooled from three mice per group, proliferation was measured by [3H]thymidine incorporation, and cytokines were measured by ELISA in the cell-free supernatants. The x-axis represents the dose of OVA (A, C) or MT (B) used in vitro. Proliferation and cytokine production after addition of rIL-2 (5 U/ml) to OVA-stimulated cultures in vitro are shown in C. The results shown are representative of three experiments.

2 and IFN-g production to OVA, compared to cells from control mice over a range of OVA concentrations in vitro (P õ 0.03 by MANOVA across four experiments as outlined under Materials and Methods). The suppression of immune response was antigen-specific, since proliferation and IL-2 and IFN-g production to stimulation with MT were similar in cells obtained from OVA- and HEL-injected mice (Fig. 1B). To determine whether IT injection of antigen is associated with a switch to Th2-type cytokines, we measured the production of IL-4 and IL-10 by splenocytes of tolerant and control mice. As shown in Fig. 1A, there

was no significant increase of antigen-specific IL-10 in the splenocyte cultures from tolerant mice. OVA-specific IL-4 was not detected in any antigen-stimulated cell culture, although CD3-induced IL-4 production was equivalent in experimental and control groups. Similar results were obtained by stimulation of draining lymph node cells.

AID

cial

CI 1165

/

6c26$$$162

08-11-97 08:25:21

Recombinant IL-2 in Vitro Reverses Suppression of IFN-g Production To investigate the mechanisms of this Th1 unresponsiveness, spleen cells were cultured with antigen in the

168

CHEN ET AL.

presence of rIL-2 (5 u/ml) for 3 days. As shown in Fig. 1C, rIL-2 caused a reversal of the suppressed OVAspecific IFN-g production by spleen cells of tolerant mice especially at high antigen dose where the IFNg production was equal to that from control cultures. However, OVA-specific proliferation and IL-2 production remained lower than in the cultures from HELinjected controls in spite of the presence of rIL-2. IT Injection of OVA Inhibits the Expansion of Antigen-Specific CD4/ T Cells in Vivo To investigate the effect of IT injection on the expansion of the antigen-specific cell population in vivo we used a sensitive system of adoptive transfer of DO11.10 TCR transgenic mouse cells specific for chicken OVA peptide 323–339 bound to I-Ad (16, 17) into normal BALB/c mice. These cells can then be tracked with the anti-clonotypic monoclonal antibody KJ 1-26 (19), which does not detect any cells in normal BALB/c mice either before or after immunization with OVA (Fig. 2), as previously reported (16). Adoptive transfer of transgenic CD4/ KJ 1-26/ T cells (5 1 106 per mouse) to naive BALB/c mice results in a very small population (1.73 { 0.13%) of KJ 1-26/ T cells in the CD4 population of draining lymph nodes (0.63 { 0.15 1 105 KJ 1-26/ cells) that can be detected by flow cytometry after 5 days (Fig. 2). After immunization with OVA/CFA, the percentage of KJ 1-26/ T cells increases significantly (P õ 0.005) to 4.3 { 0.25% of CD4/ cells (2.99 { 0.53 1 105 KJ 1-26/ cells) as was previously described (16). Interestingly, preinjection of OVA intrathymically 48 hr before immunization (24 hr before adoptive transfer of transgenic cells) almost completely prevented the increase of KJ 1-26/ cells which constituted 2.1 { 0.26% of CD4/ cells (P Å 0.004 versus immunized controls; P Å NS compared to nonimmunized controls) after immunization (0.58 { 0.47 1 105 KJ 1-26/ cells). The specificity of this observation was confirmed when we found that the KJ 1-26/ cell population in HELinjected control mice increased similar to immunized controls to 3.67 { 0.29% of CD4/ cells (P Å 0.016 versus OVA-tolerized mice) (1.97 { 0.31 1 105 KJ 1-26/ cells). Injection of 100 mg of OVA intraperitoneally instead of intrathymically did not prevent clonal expansion; the percentage of KJ 1-26/ cells was 3.7 { 0.9% of CD4/ cells (2.5 { 1.2 1 105 KJ 1-26/ cells). The data presented are from three animals per group; the experiment was repeated four times with the same results. Figure 2 shows the staining from a representative animal per group. These data clearly demonstrate that IT injection of antigen inhibits the expansion of antigenspecific CD4/ T cells in vivo.

restimulation in vitro, we cultured draining lymph node cells from recipients of the transgenic cells with OVA (100 mg/ml) for 3 days. On Day 5 postimmunization the predominant population of cells responding to OVA are the transgenic KJ 1-26/ cells (16). Cells from tolerant mice had a marked decrease in OVA-specific T cell proliferation and IFN-g and IL-2 production, compared to cells of HEL-injected control mice (P õ 0.03 by MANOVA across four experiments as outlined under Materials and Methods) (Fig. 3A). Similar results were obtained when the cells were stimulated with p323–339 of OVA (sequence was obtained from published data (18) and synthesized at the Biopolymer Center, Center for Neurologic Diseases, Brigham and Women’s Hospital, Boston, MA) (not shown). The response of lymph node cells from tolerized mice to the control antigen MT was not significantly affected and was equivalent to the response of lymph node cells from HEL-injected controls (data not shown). Since the number of KJ 1-26/ cells in the lymph nodes from OVA IT-tolerized mice is lower than that from the control mice, the proliferation and cytokine production were adjusted to cpm per 10,000 KJ 1-26/ cells or pg/ml per 10,000 KJ 1-26/ cells. The adjusted results are shown in Fig. 3B and are similar to those in Fig. 3A. Similar to our observations with lymphocytes of intrathymically tolerized BALB/c mice, addition of recombinant IL-2 in vitro caused a reversal the IFN-g production but not IL-2 production in the cultures from tolerant mice. T Cells Activated by in Vivo Immunization Circulate to the Thymus

To study whether CD4/ KJ 1-26/ T cells inactivated by IT injection of OVA could be activated by antigen

After adoptive transfer of transgenic CD4/ KJ 1-26/ T cells (5 1 106 per mouse) to naive BALB/c mice there is no detectable staining with KJ 1-26 antibody in the thymus after 6 days (Table 1). Since the detection of these KJ 1-26/ cells may be obscured by the large number of double-positive cells in the thymus of naive animals, we enriched for single-positive CD4 cells by incubating thymocytes with anti-CD8 mAb and rabbit complement. The percentage of KJ 1-26/ cells remained undetectable even after enriching for CD4 single-positive cells. However, 5 days after immunization with OVA/CFA (6 days after transgenic cell transfer) there is an increase in KJ 1-26/ cells in the thymus to 0.33% of thymocytes (0.21 1 106 cells). Injection of HEL in the thymus followed by immunization with OVA/CFA does not significantly change the number of KJ 1-26/ cells compared to immunized animals (0.23 { 0.08 1 106 cells) even though the percentage of KJ 1-26/ cells appeared to increase (0.79 { 0.2% of thymocytes) due to a decrease in the total number of thymocytes. Similarly, in animals injected intrathymically with OVA followed by immunization with OVA/CFA, the number of KJ 126/ cells in the thymus was not significantly changed (0.16 { 0.06 1 106) even though the percentage of KJ

AID

cial

TCR Transgenic T Cells Have Decreased Proliferation and Cytokine Production upon Restimulation with OVA in Vitro

CI 1165

/

6c26$$$162

08-11-97 08:25:21

MECHANISMS OF ACQUIRED THYMIC TOLERANCE

169

FIG. 2. Failure of expansion of antigen-specific T cells after IT injection of OVA. Flow cytometric analysis was performed on pooled popliteal and axillary lymph node cells from each mouse stained with KJ 1-26 monoclonal antibody followed by FITC-labeled goat antimouse IgG and PE-labeled anti-CD4 monoclonal antibody. CD4/ cells were gated and the percentage of KJ 1-26/ cells is displayed on the contour plot. The top panel shows from left to right cells from an immunized BALB/c mouse that did not receive transgenic cells, cells from a naive BALB/c mouse injected with 5 1 106 TCR transgenic T cells 6 days before analysis, and cells from a BALB/c mouse that received TCR transgenic T cells and was immunized with OVA/CFA 5 days prior to analysis. In the lower panel we show cells from a BALB/c mouse preinjected intrathymically with HEL on the day prior to transgenic T cell transfer and immunized 5 days prior to analysis, cells from a BALB/c mouse preinjected intrathymically with OVA on the day prior to transgenic T cell transfer and immunized 5 days prior to analysis, and cells from a BALB/c mouse preinjected intraperitoneally with OVA on the day prior to transgenic T cell transfer and immunized 5 days prior to analysis. The results shown are for a representative animal from each group, with 3 mice/group.

1-26/ cells appeared to be increased (1.14 { 0.18% of thymocytes) again due to a decreased number of total thymocytes. Injection of OVA intrathymically without immunization did not cause an increase in the number or percentage of KJ 1-26/ cells in the thymus (õ0.01% of thymocytes) even though the total number of thymocytes decreased significantly. Table 1 shows the number of thymocytes as well as the number and percentage of KJ 1-26/ cells in each experimental group. These data show that on Day 5 postimmunization, the number of KJ 1-26/ cells in the thymus is equivalent in all the groups that received transgenic cells and were immunized regardless of the antigen injected in the thymus. Immunization and the surgical procedure associated with intrathymic injection caused a significant loss of thymocytes presumably due to stress. Proliferation of thymocytes taken on Day 5 postim-

munization to OVA in vitro was measured and showed that thymocytes from OVA-tolerized mice had a Dcpm of 2000 per 104 KJ 1-26/ cells compared to a proliferation of 18,500 per 104 KJ 1-26/ cells in the HEL-injected mice. Thymocytes of naive mice and of mice injected IT with OVA but not immunized did not proliferate to OVA in vitro. These data suggest that the KJ 1-26/ cells in the thymus of OVA-injected animals have been inactivated.

AID

cial

CI 1165

/

6c26$$$162

08-11-97 08:25:21

Thymectomy prior to Immunization Abrogates the Effect of Intrathymic Antigen Injection Our hypothesis is that activated antigen-specific cells circulate to the thymus and are inactivated. Another possibility is that thymic cells (possibly dendritic cells) migrate from the thymus to interact with anti-

170

CHEN ET AL.

FIG. 3. Unresponsiveness of antigen-specific T cells from intrathymically tolerized mice after restimulation with OVA in vitro. Mice were intrathymically injected with HEL or OVA, and 24 hr later each mouse received 5 1 106 KJ 1-26/ T cells and was immunized with OVA/CFA 24 hr after transfer. (A) Lymph node cells from each group were collected on Day 5 after immunization and pooled. Naive BALB/c mice that received transgenic cells without immunization are the closed circles, mice injected intrathymically with OVA and then immunized with OVA/ CFA are the open squares, and mice injected intrathymically with HEL and then immunized with OVA/CFA are the closed squares. Proliferation was measured by [3H]thymidine incorporation and cytokines were measured by ELISA in the cell-free supernatants. (B) Proliferation and cytokine production by lymph node cells reported as Dcpm/per 10,000 KJ 1-26/ cells or pg/ml of cytokine per 10,000 KJ 1-26/ cells.

gen-specific cells and anergize them in the periphery. To distinguish these possibilities, we injected BALB/c mice intrathymically with OVA and then thymecto-

mized the animals 48 hr later. The next day the mice received KJ 1-26/ cells and were immunized with OVA. Control groups included mice injected with OVA IT

TABLE 1 T Cells Activated by in Vivo Immunization Circulate to the Thymus

IT

SQ

No. of animals

— OVA — HEL OVA

— — OVA/CFA OVA/CFA OVA/CFA

2a 1 2a 3 5

No. of thymocytes (1106 { SE) 185

{ 23 64.5 { 28.8 { 11.6 {

25 2.5 7 3.7

No. of KJ 1-26/ cells (1106 { SE)

KJ 1-26/ cells (%) { SE

õ0.02 õ0.02 0.21 0.23 { 0.08 0.16 { 0.06

õ0.01 õ0.01 0.33 0.79 { 0.2 1.14 { 0.18

Note. The antigen injected intrathymically is listed under the column IT; the immunizing antigen is listed under the column SQ. The mice were all of the same age; the total number of thymocytes is significantly different between groups (P õ 0.001 by ANOVA) but no significant differences were found between the number of KJ 1-26/ cells. This experiment was repeated four times with similar results. a Thymocytes from this group were pooled before staining.

AID

CI 1165

/

6c26$$$163

08-11-97 08:25:21

cial

171

MECHANISMS OF ACQUIRED THYMIC TOLERANCE

TABLE 2 Thymectomy prior to Immunization Abrogates the Effect of Intrathymic Antigen Injection

Thymectomy No IT OVA IT OVA IT No IT

No No Yes Yes

No. of KJ 1-26/ cells (1105 { SE)

No. of LNCs (1106 { SE) 27.6 8.65 20.7 9.3

{ { { {

6.4 0.8 4.2 3.3

5.1 0.03 2.74 1.75

{ { { {

0.2 0.01* 0.8 0.8

KJ 1-26/ cells (%) { SE 2.74 0.7 2.84 2.62

{ { { {

1.3 0.2* 0.6 0.3

Note. There were two to three animals per group. * P õ 0.001 compared to the other groups by Mann–Whitney test. No significant differences existed between the OVA IT thymectomized group and the immunized only or thymectomized and immunized controls.

without thymectomy, mice immunized only, and mice thymectomized and immunized without IT injection. As seen in Table 2, the failure of expansion usually seen in the intrathymically injected group (KJ 1-26/ cells 0.7 { 0.21%) does not occur when the animals are thymectomized prior to immunization (KJ 1-26/ cells 4.11 { 1.32%). Proliferation and IL-2 and IFN-g production in this thymectomized group are indistinguishable from those in control nontolerized animals. These findings indicate that the transferred KJ 1-26/ cells do not become tolerized in the absence of the thymus and argue against thymic cells migrating to the periphery to anergize antigen-specific T cells. Intrathymic Injection of OVA Alters the Kinetics of Antigen-Specific Cells in the Thymus Thymocytes from animals injected intrathymically with OVA or HEL followed by administration of KJ 126/ cells and immunization with OVA/CFA were collected on Day 2, Day 5, and Day 10 postimmunization to study the kinetics of circulation to the thymus. As seen in Fig. 4, KJ 1-26/ cells are detectable in the thymus on Day 2 postimmunization (0.05 { 0.04 1 106 for the HEL group and 0.07 { 0.06 1 106 for the OVA group); their number increases on Day 5, and in the case of HEL-injected mice their number remains relatively stable on Day 10 (0.14 { 0.02 1 106 KJ 1-26/ cells). However, in the case of OVA-tolerized animals the number of KJ 1-26/ cells was significantly decreased on Day 10 postimmunization compared to HEL-injected controls (0.03 { 0.02 1 106 KJ 1-26/ cells, P Å 0.03). These data suggest that in OVA-injected animals, antigen-specific T cells may be deleted in the thymus, although recirculation to the periphery cannot be ruled out.

has been confirmed by parallel studies showing failure of injection of antigen systemically or into other lymphoid organs to induce tolerance in these models. In the autoimmune disease model EAE, intrathymic injection of MBP or its major encephalotigenic peptide (71–90), but not a nonencephalotigenic peptide (21– 40) of MBP prevents disease in the Lewis rat (9) and this is associated with antigen-specific suppression of proliferation of the lymph node T cells of these animals (9). We have previously shown that, in this model, there is downregulation of Th1 cytokines in the brain without evidence of Th2 upregulation (9). Thymectomy up to Day 7 after immunization abrogates the effect of acquired thymic tolerance (9). In a more recent study we have also shown that thymic dendritic-enriched cells mediate the induction of acquired thymic tolerance (12). Thymic dendritic cells harvested from animals injected intrathymically with p71–90 protect naive recip-

DISCUSSION Induction of specific tolerance with intrathymic injection of antigen was initially described by Waksman and colleagues in the 1960s–1970s (20–22) and recently applied to several transplantation (4, 5) and autoimmune models (6–10). The uniqueness of the thymus

AID

CI 1165

/

6c26$$$163

08-11-97 08:25:21

FIG. 4. Kinetics of antigen-specific cells circulating through the thymus. Mice were intrathymically injected with OVA (open symbols) or HEL (closed symbols), and 24 hr later they received 5 1 106 TCR transgenic T cells (KJ 1-26/) per mouse; they were immunized with OVA/CFA 48 hr after the initial intrathymic injection. Each point represents the mean number of KJ 1-26/ cells in the thymus for the group (2–3 per group) with the standard error.

cial

172

CHEN ET AL.

ients from EAE when injected systemically. Interestingly, we also showed that the protective effect of thymic dendritic cells is abrogated in thymectomized animals, suggesting that the interaction of dendritic cells and T cells occurs in the thymus (12). Our hypothesis is that primed Th1 cells circulate to the thymus and become inactivated. To demonstrate specific Th1 tolerance we now used a mouse model with injection of ovalbumin intrathymically. We show that spleen and lymph node cells from tolerized animals have an antigen-specific suppression of proliferation and suppression of IFN-g and IL-2 cytokine production. We did not detect an increase in production of IL-4 or IL-10 in vitro. These results indicate that injection of antigen intrathymically leads to an antigen-specific Th1 cell unresponsiveness (9, 11, 13, 23). Exogenous IL-2 did not reverse the antigen-specific unresponsiveness. Stimulation of anergic cells with IL2 causes them to proliferate and to undergo a partial or complete reversal of the proliferative defect, when restimulated with antigen and APCs (24–26). However, anergic CD4/ T cells tolerized in vivo are resistant to exogenous IL-2 (27). In contrast to IL-2 production and proliferation, we found that exogenous IL-2 can reverse the IFN-g production completely when cells are restimulated with antigen in vitro. This is consistent with the described hierarchy of suppression for different lymphokines when T cells are anergized, IL-2 production being most affected, IL-3 less so, and IFN-g the least (26, 28). Since IL-2 is required for IFN-g production, tolerant T cells that cannot produce endogenous IL-2 will also have inhibited IFN-g production. Hence, when exogenous IL2 is supplied, the production of IFN-g by these tolerant T cells is restored. The reversal of IFN-g production by exogenous IL-2 to the level seen in control cells argues for the possibility of anergy of antigen-specific T cells. The novelty of our studies is in the data obtained using adoptive transfer of TCR transgenic cells as an approach to track antigen-specific cells in vivo (16, 17, 29). We show failure of expansion of clonotypepositive T cells in the lymph nodes of tolerized animals. In vitro stimulation of lymph node cells after adoptive transfer of TCR transgenic cells shows that the KJ 126/ cells taken from OVA-tolerized mice have decreased proliferation and cytokine production compared to control mice, even after adjusting for the initial number of KJ 1-26/ cells in the culture, confirming that neither the native primed lymph node cells nor the CD4/ KJ 1-26/ T cells from tolerized animals are responsive to OVA. These data confirm that intrathymic injection of OVA results in anergy of antigenspecific T cells in vivo. How do peripheral T cells become anergized? Here we show data demonstrating that peripheral T cells circulate to the thymus. The number of activated clonotypespecific cells increases in the thymus after immunization and thymectomy abrogates the effect of intrathymic

injection, suggesting that activated cells circulate to the thymus rather than thymic cells migrating out to the periphery. One possibility is that activated T cells are anergized in the thymus before recirculating to the periphery in a dynamic process such that new activated cells migrate to the thymus as anergized cells migrate out of the thymus. Another possibility is that a subset of T cells anergized in the thymus recirculate to the periphery and anergize other T cells (30–32). It has been suggested that this process is dependent on IL-4 (32) and biases T cells toward a Th2 phenotype (33). These possibilities are currently under investigation. We also show an increase in the clonotypic population in the thymus in immunized animals that received KJ 1-26/ cells systemically. This is the first demonstration that antigen-specific T cells primed by immunization in vivo circulate through the thymus and confirms previous observations that systemically administered in vitro activated cells circulate through the thymus (14, 15). The kinetic studies are particularly interesting because they show similar kinetics of KJ 1-26/ cells in OVA- or HEL-injected animals on Day 2 and Day 5 after immunization, but on Day 10 the number of antigen-specific cells in the OVA-tolerized animals decreased significantly suggesting that they may be deleted in the thymus. To confirm the occurrence of deletion we will need to amplify this response by injecting antigen intrathymically in TCR transgenic mice (18, 34). It is likely that this is a dynamic process where anergic T cells undergo programmed cell death and are deleted (35). In summary, our results taken together with our recent observations that thymic dendritic cells are responsible for induction of acquired thymic tolerance (12) suggest that primed T cells circulating through the thymus interact with thymic dendritic cells which have been preexposed to the antigen leading to systemic tolerance most likely by mechanisms of anergy and/or ultimate deletion of these T cells (36).

AID

cial

CI 1165

/

6c26$$$163

08-11-97 08:25:21

ACKNOWLEDGMENTS We thank Dr. D. Loh for providing the TCR transgenic mice and the clonotypic antibody. We also thank Chang A. Kwok for her technical assistance.

REFERENCES 1. Kappler, J. W., Roehm, N., and Marrack, P. C., Cell 49, 273, 1987. 2. Posselt, A. M., Barker, C. F., Tomaszewski, J. E., Markmann, J. F., Choti, M. A., and Naji, A., Science 249, 1293, 1990. 3. Remuzzi, G., Rossini, M., Imberti, O., and Perico, N., Lancet 337, 750, 1991. 4. Remuzzi, G., Perico, N., Carpenter, C. B., and Sayegh, M. H., J. Am. Soc. Nephrol. 5, 1639, 1995. 5. Naji, A., Curr. Opin. Immunol. 8, 704, 1996. 6. Posselt, A. M., Barker, C. F., Friedman, A. L., and Naji, A., Science 256, 1321, 1992.

MECHANISMS OF ACQUIRED THYMIC TOLERANCE 7. Koevary, S. B., and Blomberg, M., J. Clin. Invest. 89(2), 512, 1992. 8. Gerling, I. C., Serreze, D. V., Christianson, S. W., and Leiter, E. H., Diabetes 41, 1672, 1992.

20. 21. 22. 23.

9. Khoury, S. J., Sayegh, M. H., Gallon, L., Hancock, W. W., Carpenter, C. B., and Weiner, H. L., J. Exp. Med. 178, 559, 1993.

24.

10. Goss, J. A., Nakafusa, Y., Roland, C. R., Hickey, W. F., and Flye, M. W., J. Immunol. 153, 3890, 1994. 11. Sayegh, M. H., Perico, N., Imberti, O., Hancock, W. W., Carpenter, C. B., and Remuzzi, G., Transplantation 56, 461, 1993.

25. 26.

12. Khoury, S. J., Gallon, L., Chen, W., et al., J. Exp. Med. 182, 357, 1995.

27. 28.

13. Sayegh, M. H., Perico, N., Gallon, L., et al., Transplantation 58, 125, 1994.

29.

14. Agus, D. B., Surh, C. D., and Sprent, J., J. Exp. Med. 173, 1039, 1991.

30.

15. Naparstek, Y., Holoshitz, J., Eisenstein, S., et al., Nature 300, 262, 1982. 16. Kearney, E., Pape, K., Loh, D., and Jenkins, M., Immunity 1, 327, 1994. 17. Kearney, E. R., Walunas, T. L., Karr, R. W., et al., J. Immunol. 155, 1032, 1995. 18. Murphy, K. M., Heimberger, A. B., and Loh, D. Y., Science 250, 1720, 1990. 19. Haskins, K., Kubo, R., White, J., Pigeon, M., Kappler, J. W., and Marrack, P., J. Exp. Med. 157, 1149, 1983.

AID

CI 1165

/

6c26$$$164

08-11-97 08:25:21

31. 32. 33. 34. 35. 36.

cial

173

Toullet, F., and Waksman, B., J. Immunol. 97(5), 686, 1966. Horiuchi, A., and Waksman, B. H., J. Immunol. 100, 974, 1968. Ellison, G., and Waksman, B., Immunology 105(2), 322, 1970. Hancock, W. W., Khoury, S. J., Carpenter, C. B., and Sayegh, M. H., Am. J. Pathol. 144, 1149, 1994. Essery, G., Feldmann, M., and Lamb, J. R., Immunology 64, 413, 1988. Schwartz, R. H., Science 248, 1349, 1990. Beverly, B., Kang, S. M., Lenardo, M. J., and Schwartz, R. H., Int. Immunol. 4, 661, 1992. Quill, H., J. Immunol. 156, 1325, 1996. Jenkins, M. K., Chen, C., Jung, G., Mueller, D. L., and Schwartz, R. H., J. Immunol. 144, 16, 1990. Judge, T. A., Tang, A., Spain, L. M., Gratiot-Deans, J., Sayegh, M. H., and Turka, L. A., J. Immunol. 156, 2294, 1996. Lombardi, G., Sidhu, S., Batchelor, R., and Lechler, R., Science 264(5165), 1587, 1994. Qin, S., Cobbold, S., Pope, H., et al., Science 259, 974, 1993. Davies, J. D., Martin, G., Phillips, J., Marshall, S. E., Cobbold, S. P., and Waldmann, H., J. Immunol. 157, 529, 1996. Lombardi, G., Hargreaves, R., Sidhu, S., et al., J. Immunol. 156(8), 2769, 1996. Critchfield, J. M., Racke, M. K., Zuniga-Pflucker, J. C., et al., Science 263, 1139, 1994. Gao, E., Lo, D., and Sprent, J., J. Exp. Med. 171, 1101, 1990. Remuzzi, G., Noris, M., Benigni, A., Imberti, O., Sayegh, M. H., and Perico, N., J. Exp. Med. 180, 1967, 1994.