Loss of the Surface Antigen 3G11 Characterizes a Distinct Population of Anergic/Regulatory T Cells in Experimental Autoimmune Encephalomyelitis¹

Immunologic tolerance is a basic property of the immune system that provides for self/nonself discrimination to protect the host from external pathogens without reacting against self (1, 2). In the periphery, autoreactive T cells specific for tissue Ags may achieve tolerance in different ways: clonal deletion, active suppression, and anergy (3, 4, 5, 6). In T cell anergy, lymphocytes are functionally inactivated after an Ag encounter, but remain alive for an extended period of time in a hyporesponsive state (7, 8). Anergy can be induced by various manipulations in vitro (9, 10, 11) and in vivo (11, 12) and has often been characterized by the loss of IL-2 production in addition to nonproliferation (13, 14). However, T cell anergy was, for a long time, a functionally defined state of hyporesponsiveness. Although down-regulation of the cell surface molecule 6C10 on CD4⁺ T cells has been suggested to represent an anergic status induced by superantigen (15), a distinct cell surface marker for anergy in autoimmune diseases has not been identified, making it impossible to purify and to distinguish populations of anergic and nonanergic T cells.

Regulatory T cells (Tregs)³ can be defined as CD4⁺ T cells that inhibit immunopathology or autoimmune disease in vivo (16, 17). Naturally occurring CD4⁺CD25⁺ Tregs are generated in the thymus and represent 5–10% of the CD4⁺ T lymphocytes in healthy adult mice and humans. Peripherally induced Tregs can be induced by administration of Ag in vivo during anergy/tolerance (18, 19). Because the major effect of IL-10 is to inhibit APC function, including the production of proinflammatory cytokines, it is likely that the function of Tregs is dependent on IL-10 only when cells of the innate immune system are involved (20). In addition to CD4⁺CD25⁺ Tregs, IL-10 Tregs, Th2 cells, and APCs also produce IL-10 and thus can control the magnitude of an immune response and limit immunopathology (21). At present, peripherally induced Treg activity is a functionally defined state, with no specific marker.

3G11, an mAb, has been found to recognize determinants that are composed of lymphoid cell gangliosides, independent of Thy-1 expression (22, 23). The 3G11 molecule is expressed on both thymocytes and peripheral T cells of all mouse strains, predominantly on the membranes of CD4⁺ T cells. 3G11⁺ T cells stimulated by mitogen produced a large amount of IL-2, but the 3G11⁻ population did not (22). We thus hypothesized that the level of 3G11 expression might reflect the functional state of CD4⁺ T cells in autoimmune diseases and could be a potential cell surface marker for the anergic state. In the present study we tested this hypothesis in an i.v. tolerance model in myelin basic protein (MBP) TCR transgenic mice with experimental autoimmune encephalomyelitis (EAE), a Th1 cell-mediated autoimmune disease of the CNS (24, 25). Our data demonstrate that loss of the 3G11 molecule on autoantigen-reactive CD4⁺ T cells represents a distinct anergic/Treg population.

Mice transgenic for a TCR specific for the Ac1–11 peptide of MBP (MBP_1–11) were generated by introducing the rearranged TCR α- and β-chain genes into the germline of C57BL/6 mice and extensively backcrossed to PL/J mice (26) (gift from Dr. R. Caspi (National Institutes of Health, Bethesda, MD) and the late Dr. C. Janeway (Yale University, New Haven, CT)). Mice were screened for the expression of MBP-specific TCR by FACS analysis with an anti-clonotypic mAb 19G (26). All mice were housed at University of Pennsylvania and Thomas Jefferson University animal care facilities.

MBP_1–11 (Ac-ASQKRPSQRHG) was made using a peptide synthesizer and was purified through HPLC. Con A was purchased from Sigma-Aldrich. Mouse rIL-2, rIL-4, and rIL-10 and mAbs to these cytokines were purchased from BD Pharmingen. Anti-mouse 3G11 IgM mAb (gift from Dr. M. Greene, University of Pennsylvania, Philadelphia, PA) was prepared as previously described (15). The Ab is IgMκ, with some association of γ2b H chain contributed by MPC11. 19G mAb was a gift from the late Dr. C. Janeway.

Female MBP TCR transgenic mice, 8–10 wk of age, were immunized with 200 μg of MBP_1–11 in 0.1 ml of PBS emulsified with the same volume of CFA containing 4 mg/ml Mycobacterium tuberculosis H37 RA (Difco). Pertussis toxin (200 ng/mouse/injection; List Biological) was given i.p. at the time of immunization and 48 h later. MBP_1–11 (200 μg/mouse) was i.v. injected into 10 mice from day 0 and then every 3 days. The same volumes of PBS were injected in 10 mice in parallel as a control. EAE was scored daily in a blind fashion as follows (27): 1, limp tail or waddling gait with tail tonicity; 2, waddling gait with limp tail (ataxia); 2.5, ataxia with partial limb paralysis; 3, full paralysis of one limb; 3.5, full paralysis of one limb with partial paralysis of second limb; 4, full paralysis of two limbs; 4.5, moribund; and 5, death. Mice were examined daily in a blind fashion for signs of EAE. All work was performed in accordance with Thomas Jefferson University guidelines for animal use and care.

Mice were killed when the disease reached its peak (16 days postinfection (p.i.)). Mononuclear cells (MNCs) were harvested from inguinal, popliteal, brachial, axillary, and cervical LN, spleen and thymus of MBP_1–11-injected and PBS-injected EAE mice. Erythrocytes in the cell pellets from spleen were hemolyzed by added NH₄Cl-Tris buffer for 5 min at room temperature, followed by washing. CD4⁺ T cells from spleens of i.v. MBP_1–11-treated (MBP_1–11-i.v.) animals were purified as previously described (15) by depleting non-CD4 cells with complement fixation and Abs directed against heat-stable Ag (J11d), I-Ek (14-4-4), and CD8 (3.155) and centrifuging over a cushion of lympholyte M (Cedarlane Laboratories). The purity of CD4⁺ cells was confirmed by FACScan. On the average, this preparation yielded 92.6% CD4⁺ purification of viable cells. To separate 3G11⁻ and 3G11⁺ populations, purified CD4⁺ T cells were cultured with anti-3G11 Abs (IgM), washed, and then cultured with rat anti-mouse IgM microbeads. 3G11⁻ and 3G11⁺ cells were collected by MACS magnetic cell sorting, using a type AS column following the manufacturer’s protocol (Miltenyi Biotec). On the average, magnetic sorting for 3G11 expression yielded CD4⁺ populations that were 82% 3G11⁻ and 95.0% 3G11⁺.

To study 3G11 expression on T cells in different immune organs and the CNS from MBP_1–11-i.v. and PBS-i.v. mice, 1 × 10⁶ MNCs from LN, spleen, thymus, and spinal cord were stained with anti-mouse CD4, CD8, 19G, and 3G11 mAb. Anti-3G11 and anti-19G Ab staining was additionally labeled using indirect fluorescence-conjugated Abs. Lymphocytes were gated, and fluorescence was analyzed using CellQuest (BD Biosciences) software. CD4⁺19G⁺ cells that were MBP TCR transgenic T cells were gated for analysis of expression of 3G11. Data representing 10,000 events were presented as a histograph. To determine the naive/activation status of CD4⁺ T cells, cells were stained with anti-CD4, anti-19G, anti-3G11, and anti-CD62L mAb simultaneously. Data were analyzed using CellQuest.

First, purified 3G11⁻ and 3G11⁺ CD4⁺ T cells were cultured separately (5 × 10⁴/200 μl) in triplicate with MBP_1–11 (10 μg/ml) and Con A (5 μg/ml) without Ag/mitogen. Cultures were also complemented with or without 50 IU/ml mouse rIL-2. Irradiated (2200 rad) syngeneic splenocytes (5 × 10⁵/well) of naive mice were plated as APCs. After 60 h of incubation, the cells were pulsed for 12 h with 1 μCi of [³H]methylthymidine (sp. act., 42 Ci/mmol). Thymidine incorporation was measured by a scintillation counter.

For the coculture experiments, 3G11⁻ T cells were cocultured with splenocytes of PBS-i.v. mice in individual wells (5 × 10⁵/200 μl) in triplicate with MBP_1–11 (10 μg/ml), with Con A (5 μg/ml) without Ag/mitogen. The ratios of splenocytes and 3G11⁻ cells in the wells were 1:0, 1:0.1, and 1:0.3, respectively, and the final volume remained 200 μl/well. 3G11⁺ cells were added at the same ratios as in the control culture. Proliferative responses were determined.

Purified 3G11⁻CD4⁺ and 3G11⁺CD4⁺ T cells were cultured with irradiated naive splenocytes (APCs) in 24-well plates, respectively. Nonisolated splenocytes from MBP_1–11-i.v. and PBS-i.v. mice in the same plate served as the control. Cells were cultured in the presence and the absence of MBP_1–11 (10 μg/ml) with or without rIL-2 (50 U/ml). Forty-eight hours later, supernatants were harvested, and production of IL-2, IFN-γ, and IL-10 was determined by ELISAs using paired mAbs according to the manufacturer’s instructions (Harlingen). The concentrations of the cytokines detected were automatically calculated based on the standard curves obtained from known concentrations.

Splenocytes from EAE mice were cultured at 5 × 10⁵/200 μl in triplicate with MBP_1–11 (10 μg/ml). Supernatants from cultures of 3G11⁻ T cells were pooled and added at 50, 25, 12.5, 6.25, 3.13, 1.56, and 0 μl, respectively, with the final volume kept at 200 μl. In Ab-blocking experiments, monoclonal anti-mouse IL-4 (10 μg/ml), anti-IL-10 (10 μg/ml), anti-IL-4 with anti-IL-10, and purified polyclonal mouse IgG (10 μg/ml) were added to the wells cultured with 50 μl of supernatants of 3G11⁻ T cells. In parallel, supernatants from cultures of 3G11⁺ T cells were added to separate wells in the same amounts as supernatants of 3G11⁻ T cells. rIL-4 at 200 U/ml, rIL-10 at 5 ng/ml, and rIL-4 with rIL-10 were added to the wells cultured with 50 μl of supernatants from 3G11⁺ T cells. Proliferative responses were determined.

To determine the potential in vivo regulatory function of 3G11⁻ T cells, we adoptively transferred these cells to EAE mice. Due to the shortage of MBP TCR transgenic mice, in the current experiments we used myelin oligodendrocyte glycoprotein (MOG) TCR transgenic mice. We have shown that these mice are an excellent alternative to MBP TCR transgenic mice for EAE and tolerance studies. Both strains are TCR transgenic for a myelin Ag and develop chronic progressive EAE. More importantly, we have shown that MOG and MBP peptides are comparably effective in i.v. tolerance induction in EAE (28, 29). EAE was induced as previously described (29). 3G11⁻CD4⁺ T cells and 3G11⁺CD4⁺ T cells were purified from spleens of MOG TCR transgenic mice, stimulated with MOG_35–55, and i.v. injected into MOG-induced EAE mice. A single dose of 3G11⁻CD4⁺ T cells at 10 or 40 million was i.v. injected into EAE mice at the time of immunization. Control mice received the same number of purified 3G11⁺CD4⁺ T cells or PBS only. Mice were examined daily in a blind fashion for signs of EAE.

Mann-Whitney U test was used for the comparison of average clinical scores, and Student’s t test was used for comparing other parameters among different groups. All tests were two-sided.

To induce autoantigen-specific tolerance, MBP_1–11 was i.v. injected at the beginning of EAE induction in MBP TCR transgenic mice. Clinical EAE was completely suppressed in MBP_1–11-i.v. mice (Fig. 1,A). Histological study showed typical infiltrates in PBS-i.v. mice, but few infiltrates were found in MBP_1–11-i.v. mice (data not shown). Similar absolute numbers of MNCs were harvested from MBP_1–11-i.v. mice and PBS-i.v. mice in lymph nodes (33.0 ± 6.1 vs 36.2 ± 8.8 × 10⁶ for each mouse; n = 5), spleen (74.4 ± 17.3 vs 76.7 ± 12.5 × 10⁶), or thymus (9.0 ± 1.2 vs 8.2 ± 1.3). More importantly, the percentage of CD4⁺19G⁺ (a marker for MBP TCR transgenic T cells) in the periphery (spleen as a representative) and the thymus was similar in both groups (Fig. 1,B). These results indicate that the tolerance is not due to clonal deletion. Splenocytes of MBP_1–11-i.v. mice produced no IL-2 or IL-4, significantly lower levels of IFN-γ (p < 0.001), and higher IL-10 (p < 0.01; Fig. 2,A). Proliferative responses to MBP_1–11 were nearly completely suppressed in MNCs of MBP_1–11-i.v. animals (p < 0.001), and such hyporesponsiveness was not reversed by IL-2 (Fig. 2 B). Thus, the suppression of clinical EAE was closely related to i.v. MBP-induced in vivo anergy of MBP-specific T cells according to criteria described recently (7).

To define the correlation between 3G11 expression on MBP-specific T cells and tolerance induction, MNCs from lymph nodes, spleen, thymus, and spinal cord from MBP_1–11-i.v. and PBS-i.v. mice were isolated and labeled with Abs against CD4, CD8, 3G11, and 19G, an mAb specific to MBP-reactive T cells. We found that the expression of 3G11 on CD4⁺19G⁺ cells was dramatically decreased in central and peripheral immune organs and the CNS. In lymph nodes, 72.2% of CD4⁺19G⁺ cells of PBS-i.v. mice expressed 3G11, whereas this percentage was 31.5% in MBP_1–11-i.v. mice. Similar results were shown in spleen (52.3 vs 35.6%), thymus (56.4 vs 21.5%), and spinal cord (59.0 vs 28%; Fig. 3). In the thymus, a similar level of 3G11 expression was found at the CD4⁻8⁻ double-negative stage (26.3% in PBS-i.v. mice vs 31.3% in MBP_1–11-i.v. mice), and 3G11 expression was dramatically down-regulated on CD4⁺8⁺ double-positive cells (70.6% in PBS-i.v. mice vs 12.7% in MBP_1–11-i.v. mice). These data indicate that the loss of 3G11 is positively correlated with in vivo anergy status, and that in the thymus, the loss of 3G11 expression after i.v. tolerance mainly occurred at the CD4⁺8⁺ double-positive stage.

To study the relation between the loss of 3G11 expression and T cell activation/memory status, we examined the expression of 3G11 and CD62L on CD4⁺19G⁺ T cells from draining lymph nodes of MBP_1–11-i.v. mice and PBS-i.v. mice. In MBP_1–11-tolerized mice, a portion of 3G11⁻ cells was CD62L^low (26.0%), consistent with a memory/activation phenotype; 9.8% of 3G11⁺ cells were CD62L^low in these mice. Similar results were observed in PBS-i.v. mice (16.6% in 3G11⁻ T cells vs 11.1% in 3G11⁺ cells). These results suggest that anergic T cells undergo an activation/memory process similar to that of nonanergic T cells.

To investigate the functional state of 3G11⁻CD4⁺ cells in MBP_1–11-i.v. mice, we separated 3G11⁻CD4⁺ cells from 3G11⁺CD4⁺ T cells. 3G11⁻CD4⁺ cells did not proliferate when stimulated with autoantigen MBP_1–11 (Fig. 4,A) and failed to produce IL-2 (Fig. 4,B). This state of hyporesponsiveness could not be reversed by rIL-2 (Fig. 4). On the contrary, 3G11⁺ T cells retained a potent capacity to proliferate with autoantigen and produced comparable amounts of IL-2 as MNCs from PBS-i.v. mice. No response to control Ag OVA was found in either 3G11⁻ or 3G11⁺ T cells (data not shown), and both 3G11⁻ and 3G11⁺ T cells showed a similar high proliferation response to mitogen Con A (Fig. 4 A).

To identify the phenotype of 3G11⁻ T cells, we analyzed cytokine production from the culture supernatant of 3G11⁻ and 3G11⁺ T cells. As shown in Fig. 4B, 3G11⁻ T cells produced no detectable IL-2 and a low level of IFN-γ, but high levels of IL-10 (p < 0.001) when stimulated with MBP_1–11 at 10 μg/ml. In contrast, 3G11⁺ T cells produced high levels of IFN-γ and IL-2, but not of IL-10 when stimulated with autoantigen, similar to the cytokine patterns of MNCs from PBS-i.v. mice. IL-4 was undetectable in both 3G11⁺ and 3G11⁻ T cells (data not shown). Both 3G11⁺ and 3G11⁻ T cells produced high levels of these cytokines, and no significant difference was found when they were stimulated with Con A (data not shown).

Although a large number of CD4⁺ T cells in animals that were tolerized with MBP_1–11-i.v. lost the expression of 3G11 Ag on their surfaces, a certain number of 3G11⁺ T cells retained 3G11 expression (as shown in Fig. 3) and responded well to autoantigen in vitro (Fig. 4). It was unclear why the presence of significant numbers of MBP-reactive T cells in these tolerized mice failed to cause EAE in vivo. We postulated that increased numbers of 3G11⁻ T cells induced by i.v. tolerance play a role in the active suppression of MBP-reactive T cells. Fig. 5 shows that splenocytes from PBS-i.v. mice had high proliferative responses to MBP_1–11, and when these cells were cocultured with 3G11⁻ T cells at a ratio of 1:0.3, the proliferative response was dramatically suppressed (p < 0.001). In contrast, there were no changes in the proliferative responses to MBP_1–11 of splenocytes that were cultured with and without adding 3G11⁺ T cells. Furthermore, the suppression by 3G11⁻ T cells was Ag specific, because 3G11⁻ T cells did not suppress proliferative responses to Con A (Fig. 5).

The crucial role of soluble factors from 3G11⁻ T cells in the suppression of Ag-reactive cells was addressed by culturing the latter with supernatant of these T cells. As shown in Fig. 6,A, supernatants from 3G11⁺ T cells clearly stimulated the proliferative responses of MBP-reactive T cells in a concentration-dependent manner. In contrast, supernatant from 3G11⁻ T cells significantly suppressed the proliferative responses of MNCs to MBP_1–11 in a dose-dependent manner (Fig. 6,B). The stimulating effect of supernatant from 3G11⁺ T cells could be suppressed by exogenous rIL-4 or IL-10 (Fig. 6,C). In contrast, the suppression by supernatants of 3G11⁻ T cells could be partly reversed by anti-IL-10 (p < 0.05; Fig. 6 D), confirming that the function of 3G11⁻ T cells is at least partially mediated by immunoregulatory cytokines.

To determine the in vivo regulatory function of 3G11⁻ T cells, we transferred these cells into B6 mice immunized with MOG_35–55 peptide. As shown in Fig. 7, a single injection of 3G11⁻CD4⁺ T cells significantly suppressed clinical EAE at both doses of cells. In contrast, transfer of 3G11⁺ T cells did not suppress EAE and, in fact, enhanced the clinical disease at a high cell number (Fig. 7). These data together with our in vitro results provide direct evidence that 3G11⁻ T cells are a distinct subpopulation that possesses Treg properties in vitro and in vivo.

This study demonstrates that loss of the surface Ag 3G11 on Ag-specific CD4⁺ T cells represents a unique characteristic for anergic T cells, which also exhibit a Treg phenotype in vitro and effectively suppress clinical EAE in vivo.

Tolerance induction to EAE has been accomplished by administration of specific encephalitogenic Ags in a variety of tolerogenic forms and by various routes, including mucosal, i.v., i.p., and s.c (4, 30). Ag-specific immunological tolerance could be induced via different mechanisms, including clonal deletion, anergy, and active suppression, depending on the route, form of Ag, dose, and time kinetics (3, 27, 28, 31, 32). In our study clonal deletion could be excluded, because MBP_1–11-i.v. and PBS-i.v. mice contained similar total numbers of MNCs in the central and peripheral immune organs and, most importantly, a similar percentage and absolute numbers of MBP-reactive T cells recognized by a clonotypic Ab 19G (26). Instead, our protocol, administration of MBP_1–11 peptide at the beginning of EAE induction, induced Ag-specific T cell anergy in these mice, as shown by suppressed proliferation and IL-2 production. It has been suggested that anergy induced in adaptive tolerance is most often initiated in naive T cells in vivo by stimulation in an environment deficient in costimulation or high in coinhibition. Anergic CD4⁺ T cells fail to proliferate and produce IL-2 and IFN-γ upon autoantigen stimulation, but produce a high level of IL-10, and the hyporesponsiveness cannot be reversed by exogenous IL-2 (7). Likewise, this type of anergic CD4⁺ T cell has been induced in our model. To search for a cell surface marker to identify anergic T cells, we tested the expression of the 3G11 molecule on transgenic T cells from tolerized and nontolerized mice, because 3G11⁺ T cells stimulated by mitogen produced a large amount of IL-2, but the 3G11⁻ population did not (22). Our findings were directly confirmed by the culture system using purified CD4⁺3G11⁻ T cells, which exhibit most properties of anergic T cells. Furthermore, the hyporesponsiveness in purified 3G11⁻ T cells occurred after autoantigen stimulation, whereas the T cell response to the mitogen Con A remained intact, providing evidence that 3G11⁻ T cells were specifically anergic to the autoantigen.

T cells rendered anergic have classically been characterized as cells that neither proliferate nor produce IL-2 after re-exposure to the cognate Ag or other TCR cross-linking reagents (8). The hyporesponsiveness can be reversed by exogenous IL-2, demonstrating that the anergic T cell can be activated to the extent that it bears a high affinity IL-2R (8). However, recent studies have shown that in most in vivo anergy models, the proliferative block cannot be reversed by IL-2. This block of IL-2R signaling can be achieved in vitro (33, 34) and in vivo (35) and may involve CTLA-4 signaling (36). The expression of low level CD62L on 3G11⁻ T cells was similar to that on 3G11⁺ T cells, suggesting that a fraction of anergic T cells exhibited an activation/memory phenotype, and this may explain the reversal of anergy after discontinuation of tolerogen (7).

Of note is the finding that in MBP_1–11-i.v. animals, a definite number of 3G11⁺ T cells persisted after induction of tolerance, and these cells were fully responsive to the autoantigen MBP_1–11 when stimulated in vitro. However, it was unclear why the presence of a significant number of MBP-reactive T cells in tolerized mice failed to cause EAE in vivo. We postulated that the function of these autoantigen-reactive T cells may be inhibited in vivo by an active immunosuppressive mechanism. Indeed, we found that proliferative responses to autoantigen were dramatically suppressed in splenocytes of PBS-i.v. mice when these cells were cocultured with 3G11⁻ T cells, and this suppression was cell number dependent. 3G11⁻CD4⁺ cells produced a dramatically increased level of IL-10, but no IL-2 and low IFN-γ, indicating a Treg phenotype. More importantly, transfer of purified 3G11⁻ T cells in vivo effectively suppressed clinical EAE, directly confirming the T regulatory property of these cells. The phenomenon of anergic T cells producing IL-10 has been reported in several recent studies (19, 37, 38). More recently, Mana et al. (39) showed that repeatedly injected superantigen induced a tolerant state, resulting from a combination of both clonal anergy and cytokine-mediated immunosuppression involved in the up-regulation of IL-10 and TGF-β. Our present study shows that IL-10 plays an important role in the immunoregulation by 3G11⁻ T cells. 3G11⁻ T cells have also been found to produce IL-4 upon stimulation with anti-CD3 (40). Although IL-4 was not detectable in the present study, IL-4 may also play a role in the immunoregulatory process, because anti-IL-4 could also partially reverse the suppression. Furthermore, there must be other immunoregulatory factors involved, because blocking both IL-4 and IL-10 could not completely reverse the 3G11⁻ supernatant-induced suppression of proliferative responses to autoantigen.

How T cells, after exposure to certain stimuli, become anergic and exhibit immunoregulatory characteristics remains unclear. Four mechanisms might potentially contribute to this phenomenon. First, inhibition of IL-2 production and loss of proliferation could be a consequence of an altered TCR-transduced signal when T cells are used as APCs (41). In that case, T cell Ag promotes a cohort of T cells in which IL-2 secretion has been silenced and the capacity to secrete Th2 cytokines maintained. This cohort of T cells has been produced in vitro, and in this study we have produced them in vivo by the induction of i.v. tolerance. Second, anergizing stimuli could lead to inhibition of the expression of one cytokine gene, but not another in the same cell (42), suggesting the likelihood that some functions were inhibited, but others remained with the same Ag specificity. When compared with Th1 cells, Th2 cells were more difficult to anergize, and this difference was probably due to different signals being generated after engagement of the TCR complex in these two cell types (42). As a result, these T cells induced with anergizing stimuli might result in the selective loss of Th1 characteristics and retention of a Th2 phenotype (43). Third, the function of 3G11⁻ T cells may at least partially overlap with that of CD4⁺CD25⁺ T cells, a well-defined immunoregulatory population (44, 45, 46, 47). Indeed, we found that although CD25⁺ T cells consist of a small proportion (∼5%) of CD4⁺ T cells, almost all CD4⁺CD25⁺ T cells (∼93%) were 3G11⁻, whereas 3G11⁻ T cells were only 30% CD4⁺CD25⁻ T cells (G.-X. Zhang, S. Yu, and A. M. Rostami, unpublished observations), suggesting a close relation between CD25⁺ and 3G11⁻ populations. Fourth, the increased expression of CD25 on 3G11⁻ T cells suggests that consumption of IL-2 may also account for the in vitro suppressive activity of these cells. It has been found that competition for locally produced IL-2 is one of mechanisms underlying the immunosuppression by anergic T cells, because these cells express an increased level of IL-2R (CD25α) (48). CD4⁺CD25⁺ Treg cells compete with naive CD4⁺ T cells for IL-2 and exploit it for the induction of IL-10 production (47). Taken together, our present study shows that Ag-specific anergy to autoimmune diseases can be induced in vivo accompanied by an immunoregulatory response. Furthermore, the findings that 3G11⁻ T cells exhibit Treg properties in vitro and effectively suppress clinical EAE in vivo suggest that 3G11⁻ may be a new Treg marker in addition to CD4⁺CD25⁺.

In summary, the present study provides evidence that loss of the surface molecule 3G11 characterizes a distinct population of anergic/Treg T cells. This is the first demonstration of the ability to identify and purify anergic T cells by a distinct cell surface marker in an autoimmune disease and paves the way for a better understanding of the mechanism of tolerance in autoimmune diseases.

We thank Dr. R. Caspi for the MBP TCR transgenic mice, the late Dr. C. Janeway for permission to use the mice and for the gift of 19G Abs, Dr. V. Kuchroo for MOG TCR transgenic mice, and Drs. M. Greene and H. Maeda for the gift of the antimouse 3G11 Ab.

The authors have no financial conflict of interest.