|Home | About | Journals | Submit | Contact Us | Français|
While the pathogenic role of B cells and CD4 T cells has been studied extensively, less is known about the role of CD8 T cells in autoimmunity and self-tolerance. To evaluate the role of CD8 T cells in autoimmunity and its modulation using self-peptides, we utilized mice expressing soluble ovalbumin (sOVA) under control of the keratin-14 promoter. Spontaneous autoimmunity occurred when sOVA-mice were crossed with OT-I mice, whose CD8 T cells carry a Vα2/Vβ5-transgenic T cell receptor with specificity for the OVA257-264-peptide (OVAp). 83% of OVA/OT-I mice died during the first two weeks of life due to multiple-organ inflammation. In contrast, preventive or therapeutic OVAp injections induced a dose-dependent increase in survival. Healthy survivors exhibited reductions in peripheral CD8 T cells, CD8-coreceptor- and Vα2-expression. Furthermore, CD8 T cells from healthy mice were anergic and could not be activated by exogenous IL-2. A block in IL-2/IL-7 signaling via the STAT5-pathway provided the basis for low surface expression of the CD8-coreceptor and failure of IL-2 to break CD8 T cell anergy. Thus, soluble T cell receptor ligand triggered multiple tolerance mechanisms in these sOVA/OT-I mice, making this treatment approach a potential paradigm for modulating human autoimmune diseases.
Autoimmune diseases are heterogeneous and affect about 5% of the populations in the US and Europe (1, 2). A common hallmark of autoimmunity is the activation of self-reactive T or B lymphocytes, which leads to the destruction of single (e.g. in Type I diabetes and pemphigus vulgaris) or multiple target tissues (e.g. in systemic lupus erythematosus and scleroderma). Well-established pathomechanisms of autoimmunity include the production of autoantibodies by autoreactive B cells, CD4 T effector cells and a lack of regulatory T cells (Tregs) (3-6). CD8 T cells are also important in the pathogenesis of autoimmune diseases (e.g. multiple sclerosis, Type I diabetes), and can mediate extensive tissue damage (7, 8). However, CD8 T cells have been less well studied in the pathogenesis of autoimmune diseases in the past because 1) the genetic susceptibility to autoimmune diseases is closely linked to MHC class II genes, and 2) there were technical difficulties in expanding and maintaining autoreactive CD8 T cells in vitro (7, 8).
To date, treatment options for autoimmune diseases include general immunosuppression or use of newly emerging therapies that target constituents of the immune system, such as anti-CD20 antibodies for depletion of B cells (9). Despite their benefits in controlling autoimmunity, these treatments make patients vulnerable to serious infections and increased risk of cancer due to reduced immunosurveillance (as well as to drug-specific side effects) (10).
In light of the limited treatment options for autoimmune diseases, specific immunotherapies, particularly peptide-specific therapies, are appealing alternatives due to their ability to target antigen-specific pathogenic T or B cells while sparing the immune system from general immunosuppression. The rationale for this approach has been developed using murine models, where peptide-specific therapy has been shown to be preventive as well as potentially curative in CD4 T cell models of diabetes (11). To date, only one murine CD8 mediated model of autoimmune disease (Type I diabetes) has been described where the use of cognate ligands of autoreactive CD8 T cells showed a limited effect on disease prevention (12).
Specific immunotherapy has also been used as a prophylactic approach for human Type I diabetes in high-risk individuals (with limited benefit) (13, 14). In addition, therapeutic approaches to multiple sclerosis (MS) involved altered peptide ligands or T-cell receptor (TCR) peptides (15, 16). Thus, although specific immunotherapy is well established as an effective treatment for IgE-mediated allergies (10, 17), it is still in its infancy for the treatment of autoimmune diseases (18).
Because CD8 T cells have been identified as 1) important mediators of self-tissue damage, and 2) potential targets for treatment, we utilized a spontaneous T cell receptor transgenic model of CD8-mediated autoimmunity and self-peptide as a therapeutic option to dampen CD8 T cell responses. A soluble fragment of chicken ovalbumin was expressed as a self-antigen under the control of the keratin-14-promotor (K14) in the epithelia of the tongue, esophagus, skin and thymus of C57BL/6 mice (sOVA mice). After crossing sOVA mice with OT-I mice, 83% of sOVA/OT-I pups developed a lethal CD8 T cell mediated inflammatory disease. Preventive injections of OVA257-264-peptides (SIINFEKL, OVAp) during the first two weeks of life as well as therapeutic injections into clinically sick mice prolonged survival of sOVA/OT-I mice into adulthood. Analysis of tolerance mechanisms present in surviving adult sOVA/OT-I mice (either spontaneous or OVAp treated) demonstrated deletion of autoreactive CD8 T cells, but a small number persisted in peripheral lymphoid organs, exhibiting either downregulation of the CD8-coreceptor or the TCR a-chain. These potentially autoreactive cells were anergic, but in contrast to classical anergy, IL-2 did not restore CD8 T cell proliferation. Also IL-7 failed to upregulate surface expression of the CD8-coreceptor. Analysis of the common IL-2/IL-7 signaling pathway via STAT5 revealed impaired STAT5-phosphorylation.
In summary self-peptides, used either prophylactically or therapeutically, rescued sOVA/OT-I mice. This treatment triggered newly implicated tolerance mechanisms involving TCR subunits and pSTAT5 mediated cytokine signaling. These findings contribute to the understanding of CD8 T cell tolerance and the mechanisms of peptide-specific immunotherapies.
All mice were housed in a clean animal facility that tested negative for MHV and pinworms and were bred and used in accordance with institutional guidelines. The generation of K14-OVA mice on C57BL/6 background has been described (19). Briefly, the K14-sOVA transgene (representing a fragment of OVA that includes the aminoacids 200-1167, (encompassing OVA257-264) was generated similarly to the membrane-bound K14-mOVA transgene previously reported, but without the platelet derived growth factor receptor (PDGFR) transmembrane domain, myc and HA sequences. OVA257-264-specific class I restricted TCR Tg. mice (OT-I) were obtained from Dr. Judy Kapp (Emory University, Atlanta, Georgia). OT-I Rag-/- mice were purchased from Taconic (Hudson, NY).
Both Southern blot and PCR analysis of genomic DNA confirmed the integration of the transgene (data not shown). Total RNA was extracted from various tissues using TRIZOL (Invitrogen, Carlsbad, CA) and RNeasy® Mini Kit (QIAGEN, Valencia, CA) according to the manufacturer's protocol. Reverse transcription was performed with StrataScript™ First-Strand Synthesis System (Stratagene, La Jolla, CA) and the resulting cDNAs were used for real-time PCR using SYBR-Green PCR Master Mix (Applied Biosystems, Foster City, CA). 5′ (GGCATCAATGGCTTCTGAGAA) and 3′ (CCAACATGCTCATTGTCCCA) primers were used to amplify the OVA fragment. PCR and data collection were performed on ABIS PRISM™ 7700 Sequence Detector (Perkin Elmer, Waltham, MA). The resulting data was normalized using actin.
Mouse tongues were harvested and the epithelium was separated from the submucosa by subepithelial injection with dispase (BD Biosciences, Bedford, MA). Tongue preparations were then placed into petri dishes containing 5000U of dispase (BD Biosciences, Bedford, MA) and incubated at 37°C for 20 minutes. Tongue epithelia were peeled off, placed in a Dounce homogenizer and disrupted in lysis buffer (mammalian cell lysis kit, Sigma, St. Louis, MO). Cell lysates were run in NuPAGE® 10% Bis-Tris gel (Invitrogen, Carlsbad, CA) and blotted to a nitrocellulose membrane (Invitrogen, Carlsbad, CA). Rabbit anti-chicken ovalbumin IgG (Polysciences, Inc., Warrington, PA) was used as a probe after blocking with 0.5% porcine gelatin (Sigma, St. Louis, MO) in TBS-T (TBS, 0.1% Tween) followed by incubation with HRP-conjugated anti-rabbit antibody (Amersham, Piscataway, NJ). Blots were visualized using the ECL plus™ Western Blotting Detection Systems (Amersham, Piscataway, NJ).
Tissue samples were fixed in 10% neutral-buffered formalin. Paraffin embedded tissues were sectioned and stained with hematoxylin/eosin using standard techniques (American Histolabs, Rockville, MD).
To deplete CD8+ T cells, 50 μg of the monoclonal antibody YTS169.4 was injected intraperitoneally in sOVA/OT-I pups on days 5, 7 and 9 after birth.
sOVA/OT-I pups were graded clinically into grade 0 (no signs of disease); 1 (sluggish weight gain); 2 (seizures inducible by handling); 3 (moribund, spontaneous seizure)
The monomeric peptide OVAp (SIINFEKL; Ser-Ile-Ile-Asn-Phe-Glu-Lys-Leu) was synthesized by NeoMPS (Strassbourg, France) and the purity was confirmed by HPLC analyzed to be 99.4%. To determine whether OVAp would obviate disease in the sOVA/OT-I mice, 1, 10 or 100 μg OVAp were dissolved in 30 – 100 μl of 1× PBS (Gibco, Carlsbad, CA) and injected i.p. on both days 5 and 9 of life. To determine whether OVAp could reverse disease in sOVA/OT-I mice that were losing weight, 10 or 100 μg were injected i.p. one time on day 9 or day 10 respectively.
The following FITC-, PE-, Biotin- and APC-conjugated monoclonal antibodies (mAb) were used: CD8, Vα2, Vβ5, pSTAT5-ALEXA488, Streptavidin-PE or –Pacific Blue (BD Pharmingen, San Jose, CA). PE-labeled SIINFEKL-specific pentamers were obtained from ProImmune Inc. (Bradenton, FL). Analysis was performed using a FACScalibur flow cytometer (Becton Dickinson, Mountain View, CA). Data was analyzed by FlowJo software (Treestar, Ashland, OR). Phosphorylated STAT5 (pSTAT5) was analyzed after 30 minutes of incubation with indicated concentrations of IL-7 or IL-2 (Peprotech, Rocky Hill, NJ), followed by 15 minutes fixation with 2% paraformaldehyde at 37°C. Subsequently, cells were permeabilized on ice, using 90% methanol (Sigma, St. Louis, MO) in PBS. After washing twice with 10% FCS, cells were resuspended in staining buffer (0.5% FCS/PBS) and stained for 1 hour with anti-pSTAT5-ALEXA488 and anti-CD8. pSTAT5 induction was calculated according to a previous report (20): pSTAT5-positive fraction of CD8-T cells × 100 × pSTAT5-fluorescence intensity.
Single cell suspensions were obtained from thymi and red cells lysed using ACK lysis buffer. Total live cell counts were obtained using a Neubauer hemocytometer and trypan blue exclusion. Cell numbers were calculated by multiplying the total cell count with the percentage of thymocte subset cells determined by flow cytometric analysis (dead cells excluded by 7-AAD uptake).
Single cell suspensions from skin draining lymph node cells (pooled cervical, axillary and inguinal regions) were cleared from red blood cells by ACK lysing buffer (Invitrogen, Carlsbad, CA) and subsequently stained with 1 μM Carboxy-fluorescein-succimidylesther (Invitrogen, Carlsbad, CA) at 37°C for 10 minutes in PBS, followed by 2 washes with 10% FCS (37°). OVAp-pulsed splenocytes (4 hours, 10μg/ml OVAp) from Rag-/- mice were used as stimulators because of their lack of T cells. 100.000 stimulators and 150.000 cells from responder lymph nodes (OT-I cells, spontaneous healthy survivors and sick pups) were cultured for 48 hours at 37°C with 5% CO2.
Unpaired student's t test was used to compare thymocyte numbers and CD8 T cell numbers in skin draining lymph nodes. A p value of less than 0.05 was considered significant.
These mice were generated in a manner similar to previously described K14 –Tg mice which express a membrane bound form of the OVA-transgene (19, 21). Expression patterns of mRNA in tissues from sOVA mice were analyzed by RT-PCR, revealing OVA-mRNA in skin, tongue, esophagus, thymus, liver, and to a lesser extent in lymph nodes and spleen (Figure 1A). OVA protein production was confirmed by Western blot analysis of lysates from tongue epithelium (Figure 1B).
To test the susceptibility and the threshold level for disease induction by autoantigen-specific (OVA257-264) Vα2+Vβ5+ OT-I cells in sOVA mice, we assessed the effect of these cells in a dose-dependent adoptive transfer model: Intravenous transfer of 1×105 naive OT-I cells induced weight loss in 100% and death in 60% of sOVA mice. While doses of ≥ 3×105 cells induced weight loss more rapidly and inevitably led to death of sOVA mice, no weight loss or pathology was observed upon transfer of 3×104 OT-I cells (Figure 1C). Four to seven days after adoptive transfer, sOVA mice exhibited severe epithelial inflammation of the tongue and esophagus, thus restricting adequate intake of water and food (Figure 1D), while no pathology was observed in C57BL/6 controls after adoptive transfer of OT-I cells.
In order to generate a spontaneous model of autoimmunity, sOVA mice were crossed with homozygous OT-I Tg mice. The resulting double transgenic sOVA/OT-I F1 progeny were phenotypically normal at birth, but 83% of these sOVA/OT-I mice developed a spontaneous lethal disease 10 to 16 days after birth (Figure 2A). Clinically, an arrest of weight gain and growth preceded death of sick sOVA/OT-I pups (Figure 2B), which was preceded by impaired milk intake (detected by lack of stomach content in sick pups, data not shown). Histology of dead pups revealed extensive inflammatory infiltrates in the K14-expressing organs- tongue, esophagus, skin, and liver as well as in the lung (Figure 2C). Marked inflammation of the tongue and esophagus might explain the impaired uptake of milk or food and arrest of growth in sick pups, while the pattern of multi organ inflammation coincides with the expression pattern of the OVA-transgene in K14-expressing tissues (Figure 2C).
Three doses of a CD8+ depleting antibody (clone YTS169.4) administered i.p. into sOVA/OT-I mice on days 5,7 and 9 after birth delayed the appearance of symptoms and lethality by three to four weeks (Figure 2D). Disease at this later time point coincided with the expansion of CD8+ T cells in the periphery (data not shown). Diluent injections (PBS) reproducibly reduced the survival of the control sOVA/OT-I mice from 17% to 0%, perhaps due to the injection itself. Clinical grading of disease revealed a steady increase of the percentage of CD8 T cells in skin draining lymph nodes of healthy sOVA/OT-I mice (grade 0) to those in severely ill pups (grade 3) (Figure 2E). FACS analysis of OT-I cells derived from skin draining lymph nodes of sick sOVA/OT-I pups revealed an activated phenotype with upregulated surface activation markers CD25 and CD69, and downregulated CD62L (Figure 2F).
To determine whether CD4 Tregs mediated the increased survival of sOVA/OT-I, we backcrossed sOVA mice to the recombinase-deficient background (Rag-/-) and subsequently crossed them with OT-I Rag-/- mice. These sOVA/OT-I Rag-/- mice showed a survival curve similar to sOVA/OT-I mice, but disease onset and death were significantly delayed. In addition, neither depletion of CD4 T cells nor of CD25+ Tregs using a CD4 depleting mAB (GK1.5) or anti-CD25 mAB (PC61) induced disease in surviving sOVA/OT-I mice (data not shown). Together, these findings argue against a major role for CD4 Tregs in survival of sOVA/OT-I mice in our model. Furthermore, CD8 FoxP3+ cells or CD8aa regulatory T cells were not detected in the sOVA/OT-I Rag+/+ mice (data not shown).
We evaluated the impact of specifically targeting autoreactive OT-I cells using soluble OVAp in the course of disease in sOVA/OT-I mice. A dose dependent prophylactic effect was observed following repeated i.p. injections of the self-peptide on both days 5 and 9 of life (Figure 3A). The highest dose, 100 μg OVAp on days 5 and 9, led to 98% survival for longer than 21 days and provided a disease free survival of ≥ 42 days. A lower dose of peptide (10 μg) on days 5 and 9 rescued 50-60% of pups compared to diluent (PBS) only injections, which reduced survival to 0% by day 21 (Figure 3B). To address the issue of specificity, we ruled out non-specific immunosuppression via self-peptide injections by administering the H2-Kb-restricted Vesicular-Stomatitis-Virus derived peptide NP52-59 (VSV NP52-59) on both days 5 and 9, and observed no increase of sOVA/OT survival compared to diluent injections. In contrast, single injections of 10 μg OVAp on day 5 (data not shown) or day 9 alone to overtly sick mice (Figure 3B) did not provide significant survival benefit for sOVA/OT-I mice. Moreover, treatment of mice that had overt signs of disease (reduced body weight, Figure 3C; reduced motility and hunched appearance) on day 10 of life using 100 μg OVAp rescued 60% of sOVA/OT-I mice by day 21.
Subsequently, mechanistic studies were performed using healthy and sick sOVA/OT-I mice and controls. The nomenclature used to define the various mice is described in Table 1. Because death occurred mainly between days 11-16, sOVA/OT-I mice living beyond day 21 were considered “survivors”.
To determine the reason for survival of sOVA/OT-I mice (with or without peptide-treatment) beyond d21, we analyzed their thymic and skin draining lymph node CD8 T cell compositions. A reduction of CD8-single positive T cells was observed in the thymi of healthy spontaneous- or healthy peptide-treated survivors compared to OT-I littermates (Table 2). This reduction in CD8 T cells was more dramatic in the skin draining lymph nodes of healthy spontaneous- or healthy peptide-treated survivors compared to sick survivors (Table 3).
Although the number and percentage of CD8 T cells in the periphery of healthy spontaneous- or healthy peptide-treated survivors was low compared to sick survivors or OT-I controls (Table 2), both were dramatically higher than the frequency of naive epitope-specific precursor CD8 T cells in C57BL/6 wild-type mice (estimated 100-200 naive CD8 T cells per mouse) (22). Moreover, the CD8 T cell numbers in skin draining lymph nodes alone of healthy spontaneous- or healthy peptide-treated survivors were considerably higher than the 100.000 naive OT-I cells which induced weight loss and death in sOVA mice (Figure 1B). The persistence of substantial numbers of potentially autoreactive CD8 T cells in lymph nodes of healthy spontaneous- or healthy peptide-treated survivors prompted us to investigate the potential tolerance mechanisms silencing these normally highly reactive CD8 T cells.
CD8 T cells from healthy sOVA/OT-I mice (spontaneous and peptide-treated) which expressed both, the TCR-Vα2 and -Vβ5 chains exhibited extensive downregulation of the CD8-coreceptor compared to sick pups or OT-I littermates (Figure 4A). Furthermore, about 50% of peripheral CD8 T cells displayed complete loss of the Vα2-chain while no regulation of the Vβ5-chain was observed (Figure 4A, B and Table 3).
In light of extensively regulated TCR components in peripheral OT-I cells of healthy spontaneous- and healthy peptide-treated survivors, we determined whether these cells could respond to their cognate ligand. While naive control CD8 T cells from OT-I mice and OT-I cells from sick sOVA/OT-I pups readily proliferated when cultured with OVAp-pulsed splenocytes, Vα2/Vβ5-positive CD8 T cells from healthy spontaneous survivors responded poorly, even upon stimulation with supramaximally pulsed stimulator cells (10μg/ml OVAp for four hours, Figure 4C). This anergy could not be broken by the addition of IL-2 to the culture media (data not shown), which is in contrast to classical CD8 T cell anergy, that normally can be overcome by IL-2 (23). As expected, no proliferation was observed in CD8+ Vα2-negative T cells (data not shown). Bypass of peptide-TCR ligation via stimulation with anti-CD3 mAB or direct activation of protein kinase C (PKC) by PMA/ionomycin led to proliferation of Vα2/Vβ5-positive CD8 T cells from healthy spontaneous survivors (4C) and Vα2-negative CD8 T cells (data not shown), thus suggesting that the regulation of autoreactive CD8 T cells in our model seems to take place at the T cell receptor while the downstream pathways in these tolerized cells remain mainly intact.
To test whether Vα2+ CD8low OT-I cells retain their antigen-specificity we analyzed their binding capacity for OVA257-264-pentamers. In accordance with the low proliferative response of Vα2-positive OT-I cells, only a small fraction of these CD8 cells in healthy spontaneous- and healthy peptide-treated survivors was able to bind OVAp (Figure 5), and Vα2-negative CD8 cells did not bind any OVA257-264-pentamers at all (data not shown). In contrast, Vα2-positive OT-I CD8 T cells from sick pups or OT-I controls bound larger amounts of OVA257-264-pentamers.
CD8 is an essential coreceptor of the TCR and increases its peptide sensitivity as well as mediates signal transduction upon ligation, thus contributing critically to CD8 T cell activation (24). Park and Singer et al. have repeatedly shown that under steady state conditions the CD8 expression levels are maintained by the homeostatic cytokine IL-7 (20, 25). Additionally, they have shown that other common gamma-chain cytokines (IL-2, IL-4, IL-15) also induce CD8-upregulation via the STAT5-signaling pathway, and that a block in STAT5 signaling can reduce CD8 expression (20).
To determine a basis for the low expression of the CD8-coreceptor and the failure of IL-2 to overcome the anergy of CD8 cells in healthy spontaneous survivors, we analyzed the effect of IL-2 and IL-7 on cell surface expression of the CD8-coreceptor. An overnight period with IL-7 deprivation of skin-draining lymph node cells led to the expected reduction of CD8 expression in OT-I mice (Figure 6A and B) (20). Both IL-2 and IL-7 readily restored CD8 expression in OT-I cells from OT-I littermates and sick survivors, but only marginally affected CD8 expression in healthy spontaneous survivors (Figure 6A and B). To determine whether a block in the STAT5-pathway is the basis for this non-reactivity of tolerized CD8 T cells to IL-2 and IL-7, we incubated skin-draining lymph node cells with IL-2 or IL-7 and analyzed signal transduction by pSTAT5 on a single cell level in CD8 T cells. Indeed, CD8 T cells from OT-I littermates and sick survivors responded to IL-7 readily with STAT5 phosphorylation, while CD8 T cells from healthy spontaneous survivors only marginally increased STAT5-phosphorylation (Figure 6C). Furthermore, sick survivors responded with strong STAT5-phosphorylation in response to IL-2 stimulation as compared to naive OT-I littermates and healthy spontaneous survivors, who showed almost completely abrogated pSTAT5 induction (Figure 6D).
Similar to healthy spontaneous or healthy peptide-treated (on both days 5 and 9) survivors, overtly sick mice treated on day 10 of life showed a reduction of CD8 single-positive thymocytes as well as diminished CD8 T cells in skin draining lymph nodes (7A). These peripheral T cells exhibited either downregulation of the CD8-coreceptor (7B) or of the Vα2 subunit (7C) of the TCR.
In addition, peripheral CD8 T cells in sick mice rescued by a single OVAp injection did not upregulate CD8 in response to culture in the presence of IL-2 or IL-7 (7D, 7E).
Self-peptide administration rescued sOVA/OT-I mice, both prophylactically and therapeutically, from lethal spontaneous CD8 T cell-mediated autoimmunity. Besides clonal deletion, four additional tolerance mechanisms may be operating in the survival induced by self-peptide injection in the model presented here: (i) downregulation of the CD8-coreceptor; (ii) downregulation and re-editing of the TCR Vα chain; (iii) reduced ligand binding of the Tg TCR complex, and (iv) non-responsiveness to common-γ-chain cytokines IL-2 or IL-7 mediated due to a defective STAT5 signaling pathway.
To date, peptide-therapy has been used successfully in prophylactic settings of experimental CD8 T cell-mediated diabetes, and provided us a rationale for attempting to attenuate the disease in our model with an epithelial self-antigen target (11, 12). Similar to the findings of Bercovici and colleagues (12), peptide-therapy was well tolerated in our study. An important difference between our study and that of Bercovici et al. is our high survival rate using only two prophylactic OVAp-injections that rescued almost 100% of sOVA/OT-I mice and lasted for 6 weeks and longer, whereas in Bercovici et al.'s study a survival rate of 40% was achieved by repeated injections from day 3 to day 5 after birth (a time point when neonatal tolerance may be induced and maintained in newborn mice). In addition, when we used a single therapeutic injection in sick sOVA/OT-I mice, 60% survived. Bercovici et al. noted deletion of autoreactive T cells in peptide-treated mice, a finding we confirmed, but did not provide further mechanistic details. In a similar model to our sOVA/OT-I mice (which express a truncated version of the OVA-protein), McGargill et al. generated K14-OVA257-264 × OT-I mice, which express membrane bound OVA257-264-peptide. These authors observed a similar clinical phenotype associated with Vα-chain downregulation as an indicator for TCR editing in healthy mice. This observation which had been described earlier by this group was concluded to be a tolerance mechanism initiated during thymic T cell development (26).
CD8-coreceptor downregulation has been predicted to be a potential mechanism of T cell tolerance in basic studies of thymic selection (27), and has been proposed as a factor in determining CD8 T cell reactivity in a very recent computer model (28). There is CD8-coreceptor tuning in mice that are spared from clinically detected autoimmunity, whereas high levels of CD8-expression characterize sick mice. Support for this concept comes from a recent study in the H-Y model, where HY TCR transgenic CD8 T cells from male mice recognize a male antigen expressed in the thymus and periphery. The authors demonstrate that defective STAT5 signaling in response to common γ-chain cytokines (IL-2, -4, -7, 15) results in low CD8 surface levels in silenced peripheral T cells (20).
IL-7-dependence for CD8-expression has been shown in previous studies (20, 25). Thus, the finding of a STAT5-signaling block in healthy spontaneous survivors explains the non-reactivity to IL-7 with impaired CD8 expression in healthy mice, while OT-I cells in sick survivors were very sensitive to this cytokine. Because of its critical role in CD8 T cell homeostasis and survival, in addition to thymic deletion, lack of IL-7 responsiveness by CD8 T cells may also contribute to the low numbers of peripheral CD8 T cells (29, 30).
IL-2 is critical for breaking tolerance mediated by antigen induced non-responsiveness (31). In this context, defective signaling via pSTAT5 can explain the failure of IL-2 to overcome anergy of CD8 T cells in our model. Furthermore, IL-2 mediates CD8 transcription and therefore, IL-2 non-responsiveness can contribute to low CD8 surface expression along with IL-7 non-responsiveness.
CD8 engagement critically enhances peptide-sensitivity of the TCR and mediates stable interactions between MHC-I on target cells and the TCR on CD8 T cells (24, 32). Thus, the lower CD8 expression on tolerized OT-I cells compared with naive OT- cells might explain the low binding of OVAp by the OT-I TCR. We are not aware of reports where defective ligand binding is a feature of self-tolerance. CD8 tuning has been observed during CD8 differentiation to memory cells in experimental murine responses to listeria or vaccinia (33). This CD8 downregulation was dependent on type I interferon signaling and therefore seems to be different than the defective STAT5-signaling in our model. Because this defect is only apparent in sOVA/OT-I mice, but not in their OT-I littermates, we assume that thymic development and selection are critical factors for this phenomenon.
In contrast to other models of autoimmunity and peptide therapy, CD4 or CD8 Tregs (reviewed in (34)) do not play a critical role in survival of sOVA/OT-I pups, since depletion of CD4+ 25+ Tregs or total CD4 cells does not induce disease in healthy sOVA/OT-I mice. In addition, no CD8 Foxp3+ or CD8αα cells were detected. Furthermore, the appearance of a similar disease observed in sOVA/OT-I mice on a Rag-/- background speaks against a critical role of CD4 Tregs in our model. In addition, the similar disease observed in sOVA/OT-I Rag-/- rules out a critical role for dual receptor specific CD8 T cells, which contributed to pathology in other studies (35).
This report of successful antigen-specific immunotherapy in a lethal model of autoimmunity should provide further impetus to attempt to develop antigen-specific treatments for human diseases. Although these approaches are still in the developmental stage and setbacks have been experienced for MS and Type I diabetes, the continuous dissection of the antigenic structures down to the molecular level of antigenic peptides in autoimmune diseases [such as myelin basic protein derived peptides in MS (e.g. MBP84-102, MBP83-99) or desmoglein 3 in pemphigus vulgaris] will facilitate tailoring of self-peptide immunotherapy to the relevant T cell populations (36, 37). At the same time discovering appropriate and specific peptides for autoimmune diseases may obviate the adverse effects that are associated with general immunosuppression. Since our model as well as the study of Bercovici et al. involve peptide administration to mouse pups between 1 to 2 weeks of age (12), the approach of developing a vaccine for young high-risk children for development of Type I diabetes is promising and is the subject of a current study (14).
Taken together, prophylactic or therapeutic self-peptide injections attenuated disease in sOVA/OT-I mice, making this treatment approach a potential paradigm for modulating autoimmune diseases. This paper describes three mechanisms that have previously not been implicated in preventing active autoimmune disease, including “CD8-coreceptor tuning”, defective STAT5-signaling leading to lack of IL-2/IL-7 responsiveness and desensitization of the TCR for its ligand. This model therefore can be used to further dissect the protective mechanisms triggered by peptide-specific immunotherapy, their potential interactions, and the relative contribution to CD8 T cell tolerance. Finally this model can be used to evaluate other novel treatment options.
We thank Jay Linton for excellent technical assistance and Mark Udey and Al Singer for insightful discussions and technical advice. We are indebted to William Telford (Flow Cytometry Core Facility, NIH, Bethesda) for multi-color flow cytometry.
This work was supported by the NCI Intramural Research Center for Cancer Research and an educational grant of the Deutsche Forschungsgemeinschaft (to J. G. [DFG 806-2]).
Disclosures: The authors have declared that no conflicts of interest exist.