Activation of islet-specific T cells plays a significant role in the development of type 1 diabetes. In an effort to control T cell activation, we expressed the inhibitory receptor, Ly-49A, on islet-specific mouse CD4 cells. Ag-mediated activation of Ly-49A T cells was inhibited in vitro when the Ly-49A ligand, H-2Dd, was present on APCs. Ag-driven T cell proliferation, cytokine production, and changes in surface receptor expression were significantly reduced. Inhibition was also evident during secondary antigenic challenge. Addition of exogenous IL-2 did not rescue cells from inhibition, suggesting that Ly-49A engagement does not lead to T cell anergy. Importantly, in an adoptive transfer model, Ly-49A significantly delays the onset of diabetes. Together these results demonstrate that the inhibitory receptor Ly-49A effectively limits Ag-specific CD4 cell responses even in the presence of sustained autoantigen expression in vivo.
Accurate identification of individuals with type 1 diabetes (T1D)3 is not possible before significant islet-specific immune destruction occurs. This major hurdle warrants consideration of therapies directed at ongoing disease. Given the importance of T cells in the pathogenesis and progression of T1D, it is not surprising that their removal can prevent disease (1, 2); however, profound immunosuppression makes this approach impractical. More specific immunotherapies are currently being explored at the preclinical and clinical levels for prevention of disease progression. For example, enhancement of immune regulation through brief treatment with anti-CD3 (hOKT3γ1(Ala-Ala)) can improve metabolic control for individuals with recent onset T1D (2, 3). However, patients still require insulin therapy, and it is not clear whether tolerance is restricted to β cell Ags. Oral administration of insulin can induce Ag-specific tolerance effectively in rodents, presumably through generation of regulatory cells (4). Unfortunately, human studies have been disappointing (5). Ag-specific immune deviation can also provide protection from disease; however, genetic background can influence the efficacy of this approach (6, 7). In addition, Ag-specific Th2 cells contribute to disease in some circumstances (8, 9, 10). Given the limitations of current potential therapies, it is likely that multiple approaches may be required to tame the autoreactive immune response present at diagnosis of T1D.
Another strategy aimed at controlling ongoing disease involves design of cells capable of disrupting inflammation through competition with autoreactive T cells for space and nutrients. The mobility of T cells together with their extensive interactions with APCs make them ideal tools to enforce immune regulation in this way. This requires that the therapeutic cell population remain unresponsive to Ag, in the classical sense, after homing to inflamed tissues or draining lymph nodes. Similar mechanisms have been proposed to explain the physiologic purpose of some anergic T cells (11, 12, 13).
A strong linkage exists between T1D susceptibility and the MHC class II locus in humans (HLA-DQA1*0301 and DQB1*0302) and in NOD mice (I-Ag7) (14, 15, 16, 17). As might be anticipated from this linkage, CD4 cells are important mediators of insulitis and hyperglycemia (18, 19). In this study, we examine the ability of an inhibitory receptor, Ly-49A, to limit CD4 cell activation as a first step in artificially controlling Ag-specific CD4 T cell responses.
Ly-49A is a type II transmembrane glycoprotein belonging to the C-type lectin family of inhibitory receptors (20, 21). Ly-49A is normally expressed on NK cells, where it functions as a potent inhibitor of cytotoxicity (20, 22), and on a small fraction of memory CD8 cells, but not CD4 T cells (23, 24). TCR-mediated up-regulation of CD69, cytokine production, and lysis of virus-infected target cells are inhibited by Ly-49A engagement on CD8 cells (23, 25, 26). Ly-49A is also found on a unique subpopulation of DX5+ cells that bear an MHC class I-restricted, lymphocytic choriomeningitis virus gp33-specific, αβ TCR (318 transgenic), but only in mice that also coexpress lymphocytic choriomeningitis virus gp33 (H8 transgenic) (27). Because these cells lack CD8 coreceptors, their physiologic relevance is uncertain. However, Ly-49A-mediated inhibition of cellular function is still evident as CD69 is not up-regulated when Ly-49A is engaged during Ag stimulation. In a different model, transgenic mice expressing Ly-49A on all NK, NKT, CD4, and CD8 T cells have impaired antitumor responses (28). Although CD8 cells mediate tumor rejection, in this system the inability to reject tumors was traced to a failure of CD4 cells to provide adequate help. Due to limitation in the model system, the exact nature of the CD4 cell impairment remains unclear. Collectively, current knowledge suggests that Ly-49A may be an effective way to render T cells unresponsive, thereby limiting the immune response; however, its capacity to inhibit CD4 cells is presently unclear.
Inhibition mediated by Ly-49A requires recognition of allele-specific major histocompatibility molecules (H-2Dd, Dk, and Dp) on potential target cells (29, 30, 31, 32). Upon binding to its ligand H-2Dd, the ITIM in the cytoplasmic domains of Ly-49A become phosphorylated. This leads to recruitment and activation of the Src homology 2 domain-containing tyrosine phosphatase-1 (SHP-1) (25, 33, 34). SHP-1 dephosphorylates src and syk family kinases such as Lck and ZAP-70 (35, 36).
We investigated the ability of Ly-49A to block TCR-mediated activation of CD4 cells using Ly-49A transgenic mice. Ly-49A engagement significantly reduced Ag-stimulated proliferation and cytokine production. Cell surface receptor expression of CD25, CD40L, CD62L, and CD69 also reflected the failure of Ag to activate T cells fully. More importantly, Ly-49A expression disrupted the ability of autoreactive T cells to mediate disease in vivo. Using an adoptive transfer model, the onset of T1D was delayed significantly when APCs in recipient mice expressed Dd. It appears that Ly-49A is capable of preventing Ag-specific activation of autoreactive T cells both in vitro and when Ag is present continuously in vivo.
Materials and Methods
B10.D2 and B10.HTG mice are MHC congenic strains that were purchased from The Jackson Laboratory. Both strains express identical MHC class II genes (I-Ad and I-Ed), but differ in expression of the MHC class I D allele: B10.D2 expresses H-2Dd, whereas B10.HTG expresses H-2Db. TCR-SFE single transgenic mice were kindly provided by H. von Boehmer (Harvard Medical School, Boston, MA) (37). The TCR-SFE TCR recognizes a peptide (110–119; SFERFEIFPK) derived from influenza (A/PR/8/34 Mt. Sinai) hemagglutinin and is restricted to I-Ed. We generated Ly-49A/TCR-SFE double-transgenic mice, as described previously (38, 39). Transgenic Ly-49A expression is controlled by a modified CD4 promoter that lacks the silencer element responsible for repression in CD8 cells (38); therefore, Ly-49A is expressed constitutively by thymocytes as well as peripheral CD4 and CD8 T cells (39). In the case of Ly-49A/TCR-SFE double-transgenic mice, the MHC class II-restricted TCR transgene limits CD8 cell development; thus, nearly all peripheral T cells are CD4 positive and express Ly-49A. Ins-HA/B10.HTG transgenic mice expressing hemagglutinin on islet β cells were kindly provided by D. Lo (La Jolla Institute for Allergy and Immunology, San Diego, CA) and have been described previously (7, 40). TCR-SFE, Ly-49A/TCR-SFE, and Ins-HA mice were maintained on the B10.HTG background. Rag-2knockout(ko)/B10.D2 mice were purchased from Taconic Farms. All mice were housed in specific pathogen-free conditions at the Southern Illinois University School of Medicine Laboratory Animal Facility in accordance with National Institutes of Health and institutional guidelines.
In vitro T cell stimulation assays
CD4 cells were purified from lymph nodes obtained from Ly-49A/TCR-SFE or TCR-SFE mice. Briefly, lymph nodes were harvested, dissociated, washed, and incubated with anti-B220 (RA3-6B2; BD Pharmingen), anti-CD8 (53-6.7; BD Pharmingen), F4/80, and M5/114 (American Type Culture Collection) for 30–60 min at 4°C to deplete B, CD8, and myeloid cells, respectively. Cells were purified using negative selection by incubation with goat anti-rat IgG-conjugated magnetic beads (Qiagen) for 30–60 min at 4°C. CD4 T cells were recovered from cell supernatant following exposure to a magnet, with greater than 90% purity, as assessed by flow cytometry.
To prepare APCs, spleens from B10.D2 (Dd APC) and B10.HTG (Db APC) mice were harvested, mechanically dissociated, and treated with 1.44% NH4Cl to lyse RBC. APCs were purified by negative selection using anti-CD4 (L3T4; BD Pharmingen), YTS.177 (American Type Culture Collection), and anti-CD8 by magnetic bead depletion, as described above. Cells were irradiated (2500 rad) before use.
All cells were cultured in RPMI 1640 (Irvine Scientific) plus 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mmol/L glutamine, 25 mmol/L HEPES, and 5 × 10−5 mmol/L 2-ME (complete medium). T cells (2 × 105 cells/well) were incubated with APC (6 × 105 cells/well) plus 0.5–10 μg/ml SFE peptide (influenza hemagglutinin peptide 110–119; SFERFEIFPK) in 96-well plates at 37°C in 5% CO2 for 4–48 h. In other experiments, cultures were supplemented with 50 U/ml human rIL-2 (rhIL-2) (PeproTech).
Secondary challenge assays
CD4 T cells were stimulated with Db APC plus SFE peptide for 72 h (primary stimulation). Viable cells were collected via density gradient centrifugation (Lympholyte-M; Cedarlane Laboratories). Recovered cells were stimulated with either Dd or Db APC plus SFE peptide for 4–72 h (secondary stimulation).
Cells were incubated with FITC- or PE-conjugated anti-CD4 and biotinylated anti-Vβ8 plus one of the following Abs: anti-CD25 PE, anti-CD40L (CD154) PE, anti-CD62L FITC, or anti-CD69 FITC (BD Pharmingen). Vβ8-positive cells were visualized with streptavidin-allophycocyanin (BD Biosciences). Three-color immunofluorescence analyses were performed using a FACSCalibur and CellQuest software (BD Biosciences). In some assays, cells were treated with annexin V FITC using the Apoptosis Detection kit I (BD Pharmingen), according to manufacturer’s instructions.
Supernatants were collected after 24, 48, or 72 h of culture and assayed for IL-2 and IFN-γ by sandwich ELISA using appropriate Ab pairs (BD Pharmingen) and peroxidase-conjugated streptavidin (Jackson ImmunoResearch Laboratories) plus 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (Sigma-Aldrich). All incubations were for 30 min at 37°C. Plates were washed with PBS containing 0.05% Tween 20 after each incubation. Cytokine concentrations were interpolated from standard curves obtained from an ELx800 Microplate reader with KC Junior software (Bio-Tek Instruments).
T cells (2 × 105 cells/well) were incubated with APC (5 × 105 cells/well) plus 0.5–10 μg/ml SFE peptide in complete medium for 72 or 96 h. All samples were plated in triplicate. Plates were pulsed with 1 μCi/well [3H]thymidine (Amersham) during the last 24 h. Samples were harvested and counted on a Beckman LS600SC scintillation counter.
Adoptive transfer diabetes model
Bone marrow cells were collected from the long bones of Rag-2ko/B10.D2 mice and injected i.v. (1–3 × 107 cells/recipient) into lethally irradiated (1100 rad) Ins-HA/B10.HTG recipients. Rag-2ko/B10.D2→Ins-HA/B10.HTG bone marrow chimeric mice were allowed to recover for 6 wk before use as recipients in adoptive transfer experiments.
For adoptive transfers, 2 × 106 sorted CD4 Ly-49Ahigh/TCR-SFE or TCR-SFE T cells were injected i.v. into sublethally irradiated (700 rad) Rag-2ko/B10.D2→Ins-HA/B10.HTG recipients. Blood glucose levels were monitored at 1- to 4-day intervals to determine onset of diabetes. Mice were euthanized, and tissues were harvested for immunohistochemistry when blood glucose levels reached 16.6 mmol/L (>299 mg/dl) or when two values ≥13.8 mmol/L (250 mg/dl) were obtained. Donor chimerism was assessed for each Rag-2ko/B10.D2→Ins-HA/B10.HTG recipient using flow cytometry to determine the percentage of CD11b splenocytes that coexpressed Dd (% chimerism = (% CD11b+Dd+ cells/total % CD11b+ cells) × 100). Mean donor chimerism was 93.4 ± 5.0%.
In some cases, Ins-HA/B10.HTG mice were used as recipients. In these experiments, splenocytes from Ly-49A/TCR-SFE or TCR-SFE mice were injected i.v. into sublethally irradiated (700 rad) Ins-HA/B10.HTG recipients. The number of CD4 cells injected per recipient was kept constant at 2 × 106, as previously described (7). Disease onset was monitored, as described above.
Monitoring diabetes onset
Blood glucose levels (diabetes ≥16.6 mmol/L or two values ≥13.8 mmol/L) were measured using OneTouch Ultra test strips with a OneTouch Ultra (LifeScan) blood glucose monitor. Disease incidence was graphed as Kaplan-Meier cumulative survival plots, and statistical analysis was performed using the log-rank (Mantel-Cox) test with StatView software (SAS Institute).
Spleen, pancreas, and lymph nodes were harvested and flash frozen in cryopreservation medium (Fisher Scientific). Sections were fixed for 1 min in ice-cold acetone and stained with anti-CD4, M5/114, or F4/80. HRP-conjugated secondary Abs and 3-amino-9-ethylcarbazole were used to visualize stained cells. Sections were counterstained with hematoxylin and visualized by light microscopy for the presence or absence, intensity, and location of infiltrates.
Ly-49A signaling inhibits T cell IL-2 production and proliferation
To determine whether Ly-49A ligation inhibits CD4 cell activation, we measured IL-2 production and proliferation of CD4 T cells from double-transgenic Ly-49A/TCR-SFE mice following in vitro Ag stimulation. T cells isolated from these mice constitutively express Ly-49A and an I-Ed-restricted TCR (TCR-SFE) specific for the influenza hemagglutinin peptide 110–119 (SFE). Because Dd is a high affinity ligand for Ly-49A, SFE peptide was presented to Ly-49A/TCR-SFE CD4 cells using irradiated T-depleted splenocytes obtained from B10.D2 mice, thereby providing both the ligand for Ly-49A (Dd) and restricting element for the transgenic TCR (I-Ed). IL-2 was nearly undetectable in supernatants obtained from cultures stimulated for 24 h (Fig. 1⇓A). Consistent with a lack in IL-2 production, T cell proliferation also remained at background levels (Fig. 1⇓B). Ly-49A ligation did not simply delay the kinetics of activation because IL-2 production and proliferation remained unchanged even after 48 h of stimulation (Fig. 1⇓, A and B). These results are in contrast to those obtained from parallel cultures of TCR-SFE single-transgenic T cells stimulated in an identical manner (Fig. 1⇓, A and B). As expected for an Ag-specific response, control T cells secreted IL-2 and proliferated robustly. Given the known rapid kinetics of IL-2 production, a detectable decline in IL-2 level was noted in control culture supernatants collected at 48 h.
To determine whether simple expression of Ly-49A on naive T cells was responsible for the failure in Ag-specific response, Ly-49A/TCR-SFE CD4 cells were stimulated in the absence of Ly-49A ligation. SFE peptide was presented by irradiated T-depleted splenocytes from B10.HTG mice, an intra-MHC congenic strain that expresses I-Ed for proper presentation of SFE peptide, but lacks Dd, expressing Db instead. In this case, Ag-stimulated Ly-49A/TCR-SFE CD4 cells were able to produce IL-2 and proliferate at levels similar to those observed with TCR-SFE CD4 cells lacking Ly-49A expression. Our results demonstrate that Ly-49A ligation inhibits the ability of CD4 T cells to produce IL-2 and proliferate in response to Ag-specific activation in vitro, and are consistent with studies performed on hybridomas by Duplay and colleagues (41). However, contrary to the finding that Ly-49A protects transfected T cell hybridomas from activation-induced cell death (41); apoptosis, as measured by annexin V staining, remained similar among all groups of stimulated T cells (Fig. 1⇑D).
Ly-49A signaling blocks T cell activation-induced changes in cell surface receptor expression and IFN-γ production
Because suboptimal signaling through the TCR can lead to partial activation including up-regulation of the high affinity IL-2R (CD25) in the absence of increased IL-2 production or proliferation (42, 43), we examined the effect of Ly-49A ligation on Ag-driven CD25 expression. Following Ag exposure for 24 h with simultaneous Ly-49A ligation (Dd APC), Ly-49A/TCR-SFE CD4 cells maintained low levels of CD25 that were similar to those seen on cells directly ex vivo (Fig. 2⇓A and Table I⇓). As expected, cells stimulated with Db APC as well as T cells lacking Ly-49A demonstrated significant increases in CD25 on day 1 (Fig. 2⇓A and Table I⇓). CD25 expression is transient, and so decreased levels are detected on these cells by day 2. Greater variability in CD25 expression was observed for Ly-49A/TCR-SFE CD4 cells stimulated with Dd APC on day 2; however, levels were not significantly different from those seen on day 0 or 1 (Table I⇓).
Because expression of adhesion molecules and other costimulatory receptors may be regulated independent of CD25, we also explored the expression of CD40L, CD62L, and CD69. Consistent with maintenance of a naive cell phenotype, CD62L levels remained high, while CD25, CD40L, and CD69 were low following Ag stimulation of Ly-49A/TCR-SFE CD4 cells using Dd APC (Fig. 2⇑A and Table I⇑). As anticipated, Ag-specific stimulation of Ly-49A/TCR-SFE CD4 cells in the absence of Ly-49A ligation resulted in down-regulation of CD62L and up-regulation of CD25, CD40L, and CD69. These changes were nearly identical to those observed following antigenic stimulation of CD4 cells lacking Ly-49A (Fig. 2⇑ and Table I⇑).
Ly-49A ligation also inhibited IFN-γ production following 48 and 72 h of stimulation (Fig. 1⇑C). Although there appeared to be some IFN-γ production in cultures of Ly-49A T cells stimulated in the presence of Dd APC, this was most likely produced by APC because cultures containing only APC produced equivalent levels of IFN-γ (Fig. 1⇑C).
Ly-49A inhibition is not reversed by exogenous IL-2
IL-2 is a pleiotropic cytokine important for T cell activation and proliferation (42). Clonal anergy results from a strong TCR signal in the absence of costimulation or, in the case of memory cells, low avidity TCR engagement with costimulation (44, 45). The resulting anergic cells fail to proliferate or secrete cytokines, including IL-2, IFN-γ, or IL-4. Because these characteristics are similar to those observed in vitro when Ly-49A is engaged on T cells simultaneously with exposure to Ag, we explored the possibility that Ly-49A creates an anergic state. Because the application of exogenous IL-2 can overcome anergy induced by incomplete T cell activation (42, 45, 46), we examined its ability to alter the unresponsive phenotype of T cells inhibited by Ly-49A. Ly-49A-mediated inhibition of T cell proliferation was still evident in the presence of 50 U/ml rhIL-2 (Fig. 3⇓A). Likewise, CD25 and CD69 were significantly lower in cultures of double-transgenic cells stimulated with Dd APC compared with control cultures (Fig. 3⇓, B and C, and Table I⇑). Together these data show that exogenous rhIL-2 has little impact on T cell activation in the presence of Ly-49A engagement.
Ly-49A effectively inhibits T cell activation during a secondary challenge
We also examined the ability of Ly-49A to influence activation of Ag-experienced T cells. In these assays, T cells were stimulated initially without Ly-49A ligation (Db APC), allowed to rest 3–5 days in the presence of 20 U/ml rhIL-2, and then restimulated in the presence of Dd or Db APC. During secondary challenge under conditions of Ly-49A ligation, a significant reduction in activation of double-transgenic T cells was observed compared with controls. This was evident in all parameters measured, including proliferation, IFN-γ secretion, and surface receptor changes (Fig. 4⇓ and Table I⇑).
Ly-49A signaling delays onset of diabetes
Using Ly-49A transgenic CD8 cells, Sentman and colleagues (47) show that high antigenic peptide concentrations can overcome the inhibitory effects of Ly-49A on CD8 cells in vitro. This raises the possibility that Ly-49A-mediated inhibition of CD4 cell function may not be maintained in vivo under conditions of continuous Ag presence as occurs during development of autoimmunity. To assess the ability of Ly-49A to inhibit T cell-mediated autoimmunity, we developed a novel model of T1D based on adoptive transfer of TCR-SFE/B10.D2 T cells into Ins-HA/B10.D2 recipient mice (6). In previous studies, we demonstrated that 100% of sublethally irradiated Ins-HA/B10.D2 recipient mice become hyperglycemic within 4 wk following transfer of TCR-SFE/B10.D2 transgenic donor T cells (6, 7). Consistent with autoimmunity, extensive insulitis occurs before diabetes onset and is characterized by accumulation of CD4 cells and APCs in islet tissue. Several studies, using similar CD4-mediated adoptive transfer models of T1D, indicate that diabetes arises due to activation of islet-specific CD4 cells. Islet β cells do not appear to up-regulate MHC class II molecules, but rather islet Ag is presented to CD4 cells indirectly via bone marrow-derived APCs (48, 49, 50). Once activated, islet-specific CD4 cells destroy insulin-producing β cells possibly via cytokine-dependent mechanisms (50, 51). To permit adoptive transfer of cells from Ly-49A/TCR-SFE or TCR-SFE mice on the B10.HTG background, we modified the TCR-SFE adoptive transfer model by using bone marrow chimeric mice as recipients. To create recipient mice, donor bone marrow cells were harvested from Rag-2ko/B10.D2 mice and transferred into lethally irradiated Ins-HA/B10.HTG mice. As a result, recipient mice (Rag-2ko/B10.D2→Ins-HA/B10.HTG) lacked peripheral T cells, were tolerant to B10.D2 and B10.HTG cells, expressed Dd on bone marrow-derived APCs, and expressed hemagglutinin on islet cells.
The complexity of this model is driven in part by the fact that thymic development in Ly-49A transgenic mice on a B10.D2 (Dd) background is abnormal, and thus, these mice are unsuitable as T cell donors for studies aimed at determining the influence of Ly-49A on naive CD4 cell function in vivo (39). On the congenic B10.HTG (Db) background, Ly-49A transgenic mice appear to have normal thymic development. Unfortunately, T cells from Ly-49A/B10.HTG mice are rejected upon transfer into mice that express the Ly-49A ligand Dd due to the MHC class I mismatch between donor (B10.D2) and recipient (B10.HTG). In our modified model, recipient mice (Rag-2ko/B10.D2→Ins-HA/B10.HTG) were tolerant of T cells transferred from B10.HTG mice, yet provided the necessary elements to test the ability of Ly-49A to inhibit CD4 T cell-mediated disease.
Rag-2ko/B10.D2→Ins-HA/B10.HTG chimeric mice receiving 106 sorted Ly-49Ahigh/TCR-SFE CD4 T cells typically did not exhibit hyperglycemia during the course of the study (Fig. 5⇓). Under similar conditions, 106 TCR-SFE cells were capable of causing disease within 10 days (Fig. 5⇓). Furthermore, transfer of Ly-49Ahigh/TCR-SFE T cells was associated with less extensive infiltration of islet tissue compared with mice receiving the same number of TCR-SFE T cells (Fig. 5⇓ and Table II⇓).
To determine whether Dd expression by bone marrow-derived cells is required for Ly-49A-mediated disease attenuation, we performed additional studies in which Ly-49A/TCR-SFE donor cells were transferred into sublethally irradiated Ins-HA/B10.HTG recipient mice. Diabetes onset among B10.HTG recipients that received Ly-49A/TCR-SFE T cells was nearly identical to mice receiving TCR-SFE donor T cells (Fig. 6⇓). This demonstrates that Ly-49A/TCR-SFE T cells are capable of eliciting autoimmune destruction in the absence of Ly-49A engagement by Dd. Failure of bone marrow chimeric recipients receiving Ly-49Ahigh/TCR-SFE cells to become hyperglycemic (Fig. 5⇑) was not due to an inherent inability of the transferred CD4 cells to mediate β cell destruction. Collectively, these data support a critical role for Ly-49A ligand expression on bone marrow-derived APCs. When bone marrow-derived APCs do not express Dd, autoimmune diabetes occurs following transfer of Ly-49A/TCR-SFE cells with incidence and kinetics similar to control mice receiving TCR-SFE cells. When bone marrow-derived APCs do express Dd, diabetes is attenuated.
Autoreactive T cells accumulate at sites of cognate Ag presentation, where they receive the necessary cell-cell interactions, cytokines, and growth factors to remain viable and expand (52, 53, 54). In the case of T1D, evidence of T cell proliferation and accumulation occurs first in the lymph nodes draining the pancreas, then later in other secondary lymphoid organs and the islet tissue itself (52, 53, 54). Therapy directed at these sites is attractive because of the potential to disrupt ongoing immune responses in an Ag-specific manner. Ironically, due to their exquisite homing properties and ability to interact with APCs, T cells may be among the most useful tools for delivery of immune therapy. Unresponsive T cells (similar to those made inactive through engagement of Ly-49A, as described in this work) may disrupt autoreactive T cell proliferation simply by competing for available space and nutrients, as previously proposed for some anergic T cells (55), or by promoting development of tolerogenic dendritic cells (55, 56, 57). Additional modifications may be made to these unresponsive T cells to enhance their effectiveness. For example, they may be designed to promote interactions with APC that result in up-regulation of Ig-like transcript-3 or -4, thereby driving development of tolerogenic APC populations (58). Alternatively, delivery of IL-10 by otherwise unresponsive T cells to sites of cognate Ag presentation may condition APC pools (59). TCR specificity of the unresponsive T cells may serve to direct therapy to select APC populations (e.g., those that present specific autoantigens) with the aim of converting conventional to tolerogenic APCs in an Ag-specific manner. As a first step in developing T cells as a tool for such therapies, we explored the ability of Ly-49A-expressing CD4 cells to remain refractory to antigenic stimulation both in vitro and in vivo.
Our data demonstrate that the inhibitory functions of Ly-49A are sufficient to disrupt many aspects of Ag-specific CD4 cell activation. Upon engagement of Ly-49A, CD4 cells do not proliferate, secrete IL-2, nor up-regulate the high affinity IL-2R, CD25, in response to Ag. These results confirm and extend previous observations made using transfected CD4 T cell hybridomas in which Ly-49A ligation inhibited IL-2 production and proliferation (41). Our observation that CD25 levels do not increase suggests insufficient IL-2R signaling may be, in part, responsible for the limited proliferation observed.
Previous studies show that Ly-49A engagement results in a failure in Fas ligand up-regulation in CD4 T cell hybridomas leading to protection from activation-induced apoptosis (41). This is in contrast to our studies in which the percentage of apoptotic cells did not change with respect to Ly-49A ligation (Fig. 1⇑D). Differences in the function of hybridomas vs fresh naive CD4 cells might explain this discrepancy. Allison and colleagues (60) found higher basal levels of apoptosis among Ly-49A transgenic CTLA-4-deficient CD8 cells from mice that also expressed Dd compared with nontransgenic CD8 control cells. Interestingly, annexin V staining among the Ly-49A transgenic CD4 cells from Ly-49A/CTLA-4ko/Dd mice did not differ significantly from controls. The degree of protection from apoptosis afforded by Ly-49A engagement may vary with the cell type and conditions of T cell stimulation. Our in vitro studies combined with the ability to detect Ly-49A TCR-SFE T cells in vivo 65 days posttransfer (data not shown) suggest that Ly-49A ligation of CD4 T cells does not significantly increase their sensitivity to apoptosis. Together these results suggest that Ly-49A engagement profoundly inhibits Ag-driven T cell expansion, but the impact on some consequences of TCR engagement, such as apoptosis, may be stimulus and cell type specific.
Lack of Ag-induced changes in CD40L, CD62L, and CD69 expression also suggests that Ly-49A inhibits full T cell activation. These results are similar to, yet more profound than, those observed among CD4 cells from Ly-49A/CTLA-4ko/Dd mice (60). In the latter case, receptor modulation appeared to be responsible for the incomplete rescue of a naive T cell phenotype. Use of in vitro culture systems may account for the more profound Ly-49A-mediated inhibition in our system; however, this cannot explain the inability of Ly-49Ahigh/TCR-SFE T cells to mediate disease following adoptive transfer. Notably, the use of Ly-49Ahigh T cells most likely contributed significantly to the profound protection from disease that we observed. Receptor modulation on cells that originally contain high levels of Ly-49A may still allow sufficient residual Ly-49A, even after some degree of down-modulation has occurred, to mediate protection from autoantigen-induced activation. Hence, Ly-49A is not only capable of compensating for some functions of the T cell inhibitory receptor CTLA-4, but also can effectively prevent potent cognate Ag-driven responses in vitro and in vivo.
The biochemical interaction between Ly-49A and TCR signaling is not entirely clear; however, Ly-49A can recruit SHP-1 in T cells (33). This is most likely an important interaction because SHP-1 is capable of dephosphorylating Lck and ZAP-70 (35, 36), and may also interact with other more distal signaling molecules such as PI3K (61).
The inability of IL-2 to rescue CD4 T cell inhibition mediated by Ly-49A suggests that alterations in TCR signaling may be less like those observed for clonal anergy than like those of in vivo anergy (also known as adaptive tolerance) (45, 62). A blockade in the Ras/MAPK pathway occurs in clonally anergic T cells, although calcium mobilization remains intact (63, 64, 65). In contrast, adaptive tolerance is associated with decreased phosphorylation of CD3 ζ-chain (p23), ZAP-70, linker for activation of T cells, and phospholipase C-γ1 (45, 66, 67). Calcium mobilization is also reduced (68, 69). Reduction in phosphorylation of the ITAM motifs of CD3 ζ and ERK2 as well as reduced calcium flux have been documented in Ly-49A-expressing T cells and T cell lines (60, 70). From the limited data available currently, it appears that Ly-49A ligation may affect biochemical alterations both proximal and distal to the TCR, but further analysis is necessary to gain a complete understanding of the limits of TCR-mediated inhibition afforded by Ly-49A.
That Ly-49A is a potent inhibitory receptor is exemplified by the inability of otherwise diabetogenic T cells to cause hyperglycemia in our adoptive transfer disease model. In the model used in this study, transfer of two million TCR-SFE CD4 cells mediates rapid disease onset (<2 wk) in 100% of recipient mice. T cell-mediated disease is both highly penetrant and aggressive, yet when the disease-causing CD4 cells also express Ly-49A and recipient mice express Dd, hyperglycemia is prevented for at least 75 days posttransfer. Because recipients are chimeric, most cells do not express Dd. That hemopoietic cell expression of Dd (resulting from transfer of Rag-2ko/B10.D2 bone marrow) is sufficient to engender protection from disease suggests that APC are a likely source of ligand for Ly-49A in our in vivo system. This is consistent with other in vitro studies showing that inhibition of CD4 T cell hybridoma function did not occur with anti-Ly-49A and anti-CD3 coated on separate beads (41).
From immunohistologic analyses of pancreas tissue, it appears that Ly-49A/TCR-SFE T cells do not accumulate in pancreas tissue to the same extent as control cells. Whether Ly-49A simply inhibits proliferation, thereby keeping the total progeny of transferred T cells low, or more directly affects trafficking awaits further analysis. Future studies aimed at determining T cell proliferation and homing patterns will be of great use in this regard.
In summary, our results show that the response of naive CD4 T cells to Ag exposure can be effectively altered by simultaneous engagement of the inhibitory receptor Ly-49A. Under these circumstances, CD4 cells effectively remain unresponsive and cannot be provoked to an activated state even with application of exogenous IL-2. More importantly, unresponsiveness is maintained in vivo in the presence of sustained autoantigen expression.
The authors have no financial conflict of interest.
We thank Dr. David Lo for helpful comments, Anna Travelstead for acquisition of flow cytometry data, and Bryan Tully for preliminary adoptive transfer studies.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported by grants from the American Diabetes Association and the Illinois Health Improvement Foundation (to M.E.P.).
↵2 Address correspondence and reprint requests to Dr. Mary E. Pauza, Department of Department of Medical Microbiology, Immunology and Cell Biology, Southern Illinois University School of Medicine, 801 North Rutledge, P.O. Box 19626, Springfield, IL 62794. E-mail address:
↵3 Abbreviations used in this paper: T1D, type 1 diabetes; ko, knockout; rhIL, human rIL; SFE, influenza hemagglutinin peptide; SHP-1, Src homology 2 domain-containing tyrosine phosphatase-1.
- Received February 5, 2004.
- Accepted January 19, 2005.
- Copyright © 2005 by The American Association of Immunologists