| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
+TCR
+ T Cells1
* Laboratory of Autoimmunity, Torrey Pines Institute for Molecular Studies, San Diego, CA 92121; and
Department of Pathology, School of Medicine, University of Utah, Salt Lake City, UT 84132
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
+ Tregs that control activated V8.2+ CD4 T cells mediating experimental autoimmune encephalomyelitis. They express exclusively the CD8 homodimer and recognize a peptide from a conserved region of the TCR V8.2 chain in the context of the Qa-1a (CD8 Tregs). They secrete type 1 cytokines but not IL-2. CD8 Tregs kill activated V8.2+ but not V8.2– or naive T cells. The CD8 Tregs prevent autoimmunity upon adoptive transfer or following in vivo activation. These findings reveal an important negative feedback regulatory mechanism targeting activated T cells and have implications in the development of therapeutic strategies for autoimmune diseases and transplantation. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Previous data generated using various animal models have indicated an important regulatory role for CD8+ lymphocytes in autoimmune diseases, transplant tolerance, neonatal tolerance, and in homeostasis of cellular and humoral immune responses (5, 7, 8, 9, 10, 11). CD8 Tregs have also been implicated in different conditions in humans as well, for example, their role in the survival of transplants (12), in the prevention of inflammatory bowel disease (13) and in the treatment of multiple sclerosis patients with either glatiramer acetate or following T cell vaccination (TCV) with irradiated autologous myelin basic protein (MBP)-reactive CD4+ T cells (14, 15). Thus, knowledge of the phenotype, Ag specificity, and mechanisms used by CD8 Tregs is important not only for elucidating the biology of immune homeostasis but also for the development of strategies to manipulate immune responses.
Mice genetically deficient in or depleted of CD8+ T cells by treatment with anti-CD8 mAb have clearly demonstrated a crucial role for CD8 T cells in regulating EAE and conferring resistance to reinduction (7, 8). Earlier studies of experimental autoimmune encephalomyelitis (EAE) in rodents, a model of Th1 CD4 T cell-mediated autoimmune disease or graft-vs-host disease and host-vs-graft disease in allotransplantation, revealed the existence of Tregs that target AgRs on T cells (16, 17). Additionally, vaccination with attenuated pathogenic CD4 T cells (TCV) or peptides derived from their AgRs provides protection from EAE and graft-vs-host disease, suggesting that the Tregs may target TCR or its peptides (18, 19, 20, 21, 22). Experiments using Qa-1 knockout or anti-Qa-1 sera have indicated that the Qa-1 molecules may play an important role in the function of the CD8 Tregs (23, 24). Qa-1 in mouse (HLA-E in humans), a nonclassical MHC class Ib molecule, forms a heterodimer with 2-microglobulin (2m) and can present peptides from both self- and foreign Ags (25, 26, 27, 28). Although the Qa-1 molecules and AgRs expressed on pathogenic CD4+ T cells have both been proposed to play a role in CD8 Treg-mediated regulation, identification of a Qa-1-binding peptide recognized by the CD8+ T cells and description of the regulatory mechanisms remain unknown.
One of the major difficulties in understanding the biology of regulatory CD8 T cells has been the inability to establish functional CD8 Treg clones in a well-defined system. We have studied Tregs in a model system in which MBP-reactive V8.2+ CD4 T cells mediate EAE in H-2u mice (6). In this model, paralytic disease is generally monophasic, and once recovered, mice resist further induction of the disease (29, 30). Previous data from ourselves and others (7, 8, 31) have suggested a crucial role for the CD8 Tregs in the control of pathogenic V8.2+ CD4 T cells. Recent studies have further indicated a crucial role for CD8+ T cells in the apoptotic depletion of MBP-reactive V8.2+CD4+ T cells following this regulation (32). In this present study, we describe the identification of a peptide from a conserved region of the TCR V8.2 chain that is targeted by the CD8 Treg population. Vaccination with this peptide prevents MBP-induced EAE in CD8+/+ but not in CD8–/– mice. We also demonstrate the Ag specificity, phenotype, MHC restriction, and regulatory mechanism used by this novel CD8+TCR+ Treg (CD8 Treg) population using functional T cell clones and lines. These findings have important implications not only for understanding immune homeostasis mediated by the CD8 Tregs, but also for the development of therapies for T cell-mediated diseases.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
B10.PL/73NS/SnJ, PL/J, C57BL/6J (B6), BALB/cJ, NOD/Lt, NZB/BINJ, B10.BR/SgSnJ, SWR/J, and SJL/J mice were purchased from The Jackson Laboratory. MBP-specific V8.2 TCR transgenic (Tg) PL/J (T/R+) (33), OT-1 TCR Tg RAG-1–/– (34), CD8–/– PL/J (8), and B6.Tlaa (B6.A-H2-T18a/BoyEg) (35) mice were provided by Drs. J. Lafaille (New York University Medical Center, New York, NY), S. P. Schoenberger (La Jolla Institute for Allergy and Immunology, San Diego, CA), T. Mak (University of Toronto, Toronto, Ontario, Canada), and P. E. Jensen (Emory University, Atlanta, GA), respectively. Mice were bred under specific pathogen-free conditions in our own colony at Torrey Pines Institute for Molecular Studies. Age-matched female mice (from 6 to 14 wk of age) were used in all experiments. Experiments involving animals were performed in compliance with federal and institutional guidelines and have been approved by Torrey Pines Institute for Molecular Studies Institute Animal Care and Use Committee.
Induction of EAE
Mice were immunized s.c. with 150 µg of MBPAc1–9 (AcASQKRPSQR) emulsified in CFA. Pertussis toxin (PT, 0.15 µg) was injected in PBS on the same day and 48 h later. Mice were observed for clinical symptoms of paralysis daily. Disease was scored on a five-point scale, as described earlier (22): 1, flaccid tail; 2, hind limb weakness; 3, hind limb paralysis; 4, whole body paralysis; and 5, death.
Proliferation assay
To examine proliferative response, draining lymph node cells from immunized mice, CD8 T cell lines or clones (50,000 cells/well) were stimulated with relevant peptides in the presence of 500,000 APCs. For the last 18 h, cultures were pulsed with [3H]TdR (1 µCi/well), harvested, and counted on a Trilux Scintillation beta counter.
Cytotoxicity assay
Cytotoxicity of CD8 Tregs was tested using a standard 4-h 51Cr release assay with minor modifications. Target cells (Con A-activated blasts or Ag-activated CD4 clones) were labeled with 51Cr (MP Biomedicals) at 37°C for 45 min. A total of 10,000/well 51Cr-labeled targets and 200 ng/well peptides was added into round-bottom 96-well plates. Plates were then incubated at 37°C, 10% CO2 for 1 h. Effector cells were added in triplicate to the wells at the indicated E:T cell ratios to a total volume of 200 µl. Spontaneous 51Cr release was measured in control wells containing target cells with medium alone. Maximum release values were obtained by lysis of target cells with 2.5% Triton X-100 (Sigma-Aldrich). The plates were then incubated for 4 h before 50 µl of supernatants was transferred into sample plates. A total of 150 µl/well scintillation liquid (Optiphase Supermix; PerkinElmer) was added, and the plates were read on a Trilux Scintillation gamma counter. The percent-specific release was calculated as (experimental release – spontaneous release)/(maximal release – spontaneous release) x 100%.
Generation of CD8 T cell lines and clones
PL/J mice were immunized s.c. with p42–50 in CFA at a dose of 20 µg/mouse. Ten days later, draining lymph node cells were harvested and stimulated in vitro as follows: 30 x 106 irradiated splenocytes in 5 ml of DMEM complete medium (20% FBS, 100 U/ml penicillin and streptomycin, 3.125 x 10–5 M 2-ME, 1 mM sodium pyruvate, 0.1 mM nonessential amino acid, and 10 mM HEPES) were pulsed with p42–50 (10 µg/ml) at 37°C, 10% CO2 for 2 h. A total of 30 x 106 draining lymph node cells in a 5-ml volume was added into the flask and mixed. Cells were cultured vertically at 37°C, 10% CO2 for 7 days. The resulting cells were then cloned at 100 or 1000 cells/well in DMEM containing 10 IU IL-2 and 1 x 106 p42–50-pulsed irradiated syngeneic splenocytes. The cells were supplemented with fresh medium (2% conditioned medium and 50 IU/ml IL-2) every 3–4 days and were restimulated with p42–50 every 2 wk. Cells from proliferating wells were then transferred to 24-well plates and expanded using the procedure as described above. To generate p42–50-reactive lines, the restimulated draining lymph node cells were expanded using 50 IU/ml IL-2 and 2% conditioned medium in 24-well plates. The lines were supplemented with fresh medium every 3–4 days and restimulated with p42–50-pulsed irradiated splenocytes every 2 wk.
To generate SIINFEKL (OVAp257–264)-specific CTL clones, splenocytes from OT-1 Tg mice were stimulated with an engineered fibroblast cell line MEC.B7.SigOVA. Briefly, the adherent fibroblast APCs were seeded at 100,000 cells/well in 24-well plates and cultured overnight. The next day, the plates were irradiated with 7000 rad and washed three times with medium to remove any nonadherent cells or cell debris. A total of 5 x 105 OT-1 Tg cells was seeded into wells and then replenished with new medium containing 50 IU/ml IL-2 and 2% conditioned medium every 3–4 days and restimulated every 2 wk. Clonality was confirmed by flow cytometry, which showed all cells expressed the V5 and V2 Tg TCR.
Peptide binding to recombinant Qa-1a molecules
Purified Qa-1a/Qdm complexes were generated from Escherichia coli-derived proteins as described previously (25). Qa-1a/Qdm complexes were incubated with the indicated concentrations of biotin peptides (b-Qdm or b-p42–50) alone or with competing unlabeled peptides (Qdm, p42–50, or QdmM2K) overnight at room temperature to allow for peptide exchange in PBS containing 0.01% Nonidet P-40 and 200 mM citrate phosphate buffer (pH 7.0). The next day, the complexes were separated from unbound peptides using G-25 Sephadex Quick Spin Protein Columns (Roche). The amount of biotin-peptide bound to Qa-1a was measured by a Europium-based fluorescence immunoassay modified for measuring MHC class I complexes using an anti-2m capture Ab.
Real-time PCR
Total mRNA was extracted from cells using the RNeasy mini kit (Qiagen) and was subjected to cDNA synthesis with an oligo dT12–18 primer. Real-time PCR was performed using Brilliant SYBR Green QPCR kit (Stratagene) on a Stratagene Mx3000p machine. The calculation of comparative mRNA expression was performed by the Stratagene software and was designated as relative quantity after normalization against internal control genes (L32 and cyclophilin) and after consideration of amplification efficiency of individual genes. The primers used are as follows (36): L32, sense, 5'-GAAACTGGCGGAAACCCA-3', and antisense, 5'-GGATCTGGCCCTTGAACCTT-3'; cyclophilin, sense, 5'-GGCCGATGACGAGCCC-3', and antisense, 5'-TGTCTTTGGAACTTTGTCTGCAA-3'; Qa-1a, sense, 5'-GCGGTATTTCCACACTGCCA-3', and antisense, 5'-TCTG TGAGGCAAAGTCAGTC-3'; and Qa-1b, sense, 5'-CCTGGACCGCGAATGACATA-3', and Qa-1b, antisense, 5'-CACCACAGCTCCAAGGA TGAT-3'.
ELISA
IL-2, IL-4, IL-5, IL-6, IL-10, IL-12, IL-13, TNF-, and IFN-
levels were measured by a sandwich ELISA using supernatants obtained from peptide-stimulated CD8 T cell clone cultures as described earlier (37). Briefly, Nunc Immuno Plates MaxiSorp F96 (Roskilde) were coated with a capture Ab at 4°C overnight. After blocking with PBS containing 10% FBS, 50 µl of supernatants was added, and the plates were incubated overnight at 4°C. Plates were extensively washed with PBS 0.05% Tween 20 and incubated with a biotin-conjugated detection Ab. Finally, plates were washed and developed using avidin-peroxidase and 2-2'-azino-bis(3-ethyl-benzthiazoline-6-sulfonic acid) substrate (Sigma-Aldrich). OD405 was measured. All cytokine capture and detection Ab pairs were purchased from BD Pharmingen.
Flow cytometry
Cell surface molecules were detected by fluorescence conjugated mAbs. Briefly, 100,000–400,000 cells/100 µl were incubated with relevant Abs at 4°C for 30 min. Cells were then washed twice with FACS buffer containing 1% FCS and 0.05% sodium azide and analyzed on a FACSCalibur (BD Biosciences). Abs purchased from BD Pharmingen include 145-2C11 (anti-CD3), 53-6-7 (anti-CD8), 53-5.8 (anti-CD8.2), 7D4 (anti-CD25), H1.2F3 (anti-CD69), IM7 (anti-CD44), MEL-14 (anti-CD62L), HM40-3 (anti-CD40), 37.51 (anti-CD28), PK136 (anti-NK1.1), GL3 (anti-TCR
), A1 (anti-Ly49A), Dx5 (CD49b), GL-7 (anti-GL-7), 6A8.6F10.1A6 (anti-Qa-1b), RR4-7 (anti-TCR V6), and MR5-2 (anti-TCR V8.1/8.2). Abs purchased from Biolegend include 18d3 (anti-CD94) and C7 (anti-NKG2D). Anti-CD122, anti-IL7R, anti-thymic leukemia (TL)-HD168, and anti-TL-18/20 were gifts from Dr. H. Cheroutre (La Jolla Institute for Allergy and Immunology, San Diego, CA).
TL-tetramer staining of the CD8 T cell clones
TL monomers were provided by Dr. H. Cheroutre’s laboratory. To conjugate the monomers, streptavidin-PE was added into biotinylated TL-monomer at 1:1 ratio. Mixtures were incubated on ice for 15 min in dark. To stain the CD8 clones, 1 µl of the above conjugated TL-tetramer was added into 100,000 cells/100 µl in FACS buffer. Cells were incubated for 30 min at 4°C and were then washed twice with buffer before being analyzed on the FACSCalibur (BD Biosciences).
Statistical analysis
Data are expressed as mean ± SEM for each group. Statistical analyses were performed using SPSS software. Two independent samples were tested by independent t test. Otherwise, the ANOVA test was used. A value of p < 0.05 was considered statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
In this study, we sought to characterize the Ag-specificity of the CD8 Tregs that control MBP-reactive V8.2+ T cells in a well-defined model of T cell-mediated autoimmune disease in B10.PL mice. Helper CD4 Tregs in this system recognize a determinant from the conserved framework 3 region of the V8.2 TCR, and because, in general, helper determinants are in close proximity to the determinants for CD8+ T cells (22), we focused on the TCR V8.2 chain as the target for the CD8 Treg response. We synthesized a series of 9 or 10-mer peptides (Table I) from the entire sequence of the TCR V8.2 chain (29) ending at hydrophobic residues, a requirement for binding to the transporter associated with Ag processing (TAP) (38). Groups of B10.PL mice were vaccinated with individual peptides and 1 wk later challenged with MBP/CFA/PT for the induction of EAE. Data in Table I show that animals vaccinated with p42–50 or longer peptides containing these residues (p41–70) but not with others are significantly protected from disease. To determine an appropriate dose, mice were vaccinated with 5 or 50 µg of p42–50 before EAE induction. As shown in Fig. 1A, although mice vaccinated with the lower dose contracted a much milder paralytic disease with delayed onset followed by a quick spontaneous recovery, the higher dose provided complete protection from EAE. To determine whether the protective effect is dependent upon the presence of CD8+ cells in vivo, CD8+/+ or CD8–/– PL/J mice were vaccinated with p41–50, p41–70, or an irrelevant TCR peptide before the induction of disease. The data in Fig. 1B show that p41–50, as well as the longer peptide p41–70, provides significant protection from EAE in CD8+/+ but not in CD8–/– mice, clearly demonstrating that the p42–50-induced regulation requires CD8+ T cells. In addition, PLP139–151-induced disease in SJL/J mice, in which the pathogenic T cells do not predominantly use the TCR V 8 gene segment, is not prevented following vaccination with the p41–50 (Table I).
|
|
+/+ or CD8–/– mice. As shown in Fig. 1C, p41–70 induces a proliferative response in CD8+/+ but not in CD8–/– PL/J mice. In contrast, another TCR peptide p76–101, containing a determinant for the CD4 Tregs, induces a similar response in both CD8+/+ and CD8–/– mice. Similarly, anti-CD8 Abs are able to completely block an in vitro recall response to p42–50 in the draining lymph node cells of mice immunized with p42–50 (Fig. 1D). In contrast, anti-CD4 Abs had no effect on the response to p42–50. These data collectively indicate that p42–50-reactive CD8 Tregs are important in the control of V8.2+ T cell-mediated disease. Generation of p42–50-specific CD8 Treg lines and clones
To further characterize the CD8 Treg response, CD8+ T cell clones were generated from p42–50-immunized mice using the limiting dilution cloning technique. Despite the difficulty of maintaining long-term in vitro culture of the p42–50-reactive CD8 T cells, three CD8 Treg clones from separate animals were generated. A representative clone, 2D11, was selected for detailed characterization. To eliminate possible artifacts resulting from long-term in vitro culture, properties of these clones were verified at the bulk level using freshly generated p42–50-reactive CD8 T cell lines. OVA-specific CD8 T cell lines and clones were generated and cultured in parallel as controls.
To examine Ag specificity, the CD8 Treg clones were assayed for proliferation and cytokine secretion in response to in vitro challenge with a panel of peptides derived from the TCR V8.2 chain in the presence of irradiated syngeneic APCs. Fig. 2, A and B, shows that 2D11 proliferates and secretes IFN- in response to p42–50 but not to another peptide, p80–94, derived from the TCR V8.2 chain. Similarly a short-term CD8 T cell line responds to p42–50 but not to other V8.2 chain-derived peptides (Fig. 2C). The cytotoxic capacity of the CD8 Treg clones was determined using a standard 4-h 51Cr release assay. Fig. 2D shows the killing of syngeneic Con A blasts pulsed with p42–50 but not an irrelevant peptide. The killing of p42–50-pulsed targets by CD8 Treg clones and lines is MHC-restricted as it does not occur when allogeneic (BL/6) blasts are used as targets (Fig. 2D and data not shown).
|
+ and express the CD8 homodimer (CD8 Treg)
The cell surface phenotype of the CD8 Treg clones was determined by staining 2D11 with a panel of mAbs and analyzed by flow cytometry. In parallel, a conventional CD8 T cell clone (OT-1) specific for a peptide of OVA (SIINFEKL) and propagated under equivalent conditions was used as a control. As shown in Fig. 3, the CD8 Treg clone is TCR+, –, NK1.1–, and B220–. Both clones express CD25 (IL-2R chain), CD122 (IL-2R chain), and IL-7R, suggesting an activated/memory phenotype. Interestingly, the CD8 Tregs maintain a low-level expression of CD69 even after a prolonged resting period in vitro in the absence of exogenous TCR peptide. This may indicate low-level cross-presentation of p42–50 because of the presence of V8.2+ T cells in the irradiated splenocytes that were used as APCs. The CD8 Treg clones are negative for CD62L, Ly49A, and GL-7 expression and positive for CD28, NKG2, and TL Ag expression.
|
One of the most striking features of the CD8 Treg clones is the absence of surface CD8 chain expression (see Fig. 3). As TL Ag displays a 10-fold higher binding affinity for CD8 homodimers compared with CD8 heterodimers (41), we examined the binding of TL tetramers to 2D11 cells. As shown in Fig. 3, only the CD8 Treg clone binds the TL-tetramer, confirming the expression of CD8 by the Treg clone. In fact, cloning of CD8+TCR+ intraepithelial lymphocytes (IELs) (CD8 IELs) have not been reported so far and which may explain the difficulty in generating CD8+ T cell clones. The CD8 Tregs also express TL Ags that can be recognized by two different Abs, HD-168 and 18/20, specific for two different epitopes on TL Ags (42, 43).
CD8 Tregs secrete Tc1-like cytokines
The cytokine secretion profile of 2D11, as well as an OT-1 CD8 clone, was examined in cell cultures after stimulation with peptide-pulsed APCs. As shown in Fig. 4A, 2D11 secretes IFN- and TNF- (Tc1-like) but no detectable levels of IL-2, IL-4, IL-5, IL-10, IL-12, and IL-13. A very low level of IL-6 secretion was detected. In contrast, the OT-1 clone secretes IFN-, TNF-, IL-2, IL-5, IL-10, and IL-13 (Fig. 4B). It is interesting that, like CD25+CD4+ Tregs, CD8 Tregs do not secrete detectable amounts of IL-2. The Tc1 phenotype of the CD8 Treg clone is not an artifact of long-term culture because short-term p42–50-reactive T cell lines also secrete IFN- but not Tc2-like cytokines (X. Tang and V. Kumar, data not shown).
|
Tregs specifically kill activated V8.2+ but not V8.2– cells
We have demonstrated recently the CD8 T cell-dependent depletion of activated but not resting MBP-reactive V8.2+ CD4 T cells following induction of regulation in vivo (32). To determine whether CD8 Treg clones can specifically kill V8.2+ CD4 T cells, a MBPAc1–9-reactive pathogenic V8.2+ T cell clone (33) was used as a target in an in vitro cytotoxicity assay. In parallel, a V14+ CD4 T cell clone was used as a V8.2– target (22). As shown in Fig. 4C, 2D11 kills Ag-activated V8.2+ but not V14+ T cells (upper panel). To determine whether V8.2+ T cell activation was required for the killing, we rested the CD4 T cell clones until they had the small rounded appearance of naive T cells. No detectable cytotoxicity toward the resting CD4 T cell targets was found (lower panel in Fig. 4C).
To further define Ag specificity of CD8 Tregs, a number of V8.2+ and V8.2– T cell lines, clones, and hybridomas were used as targets in cytotoxicity assays. CD8 Tregs showed specific cytotoxicity toward MBPAc1–9-reactive TCR V8+ CD4 T cell line (PL4/V8+) or clones (3C10/V8.2V4.2 and 2C6/V8.2V2.3) but not TCR V8– CD4 T cell line (PL9/non-V8) or clone (A25/V13V2.3) (Table II). It is clear that V8.2+ T cells using different V chains are efficiently killed by the CD8 Tregs. We also examined the ability of the CD8 Tregs to kill MBPAc1–9-reactive TCR V8+ or V8– CD4+ T cell hybridomas. Interestingly, CD8 Tregs did not show detectable cytotoxicity toward either TCR V8+ (172.10 generated from the clone 2C6) or TCR V8- (I-12, L-12, and L-15) hybridomas. However, CD8 Tregs do kill p42–50-pulsed TCR V8– (I-12) or V8+ (172.10 or 1934.4) hybridomas (Table III), suggesting poor display of the TCR-derived peptide Qa-1a complexes on the target T cell hybridomas (see below).
|
|
Tregs are restricted by a class Ib MHC molecule, Qa-1a
Genetic, biochemical, and immunological approaches were used to determine the MHC restriction of the CD8 Treg clones. First, to determine whether presentation of the p42–50 requires 2m, which pairs with a H chain to form MHC class I molecules, we compared the ability of APCs from 2m+/+ and 2m–/– mice to stimulate 2D11 in vitro. As shown in the upper panel of Fig. 5A, 2m+/+ but not 2m–/– APCs are able to present p42–50 and stimulate the 2D11 clone. These data clearly suggest that 2D11 clone is restricted by a class I MHC molecule. Next, in vitro proliferation assays were performed to examine the response of 2D11 to p42–50-pulsed APCs derived from mice of several MHC haplotypes. As shown in Table IV, 2D11 responds to p42–50-pulsed APCs from PL/J (H-2u), B10.PL (H-2u), NZB (H-2d), B10.BR (H-2k), SWR/J (H-2q), and NOD (H-2g7) mice but not to p42–50-pulsed APCs from C57BL/6 (H-2b), BALB/c (H-2d), and SJL/J (H-2s) mice. Since a particular MHC class Ia gene did not correlate with the 2D11 response, we reasoned that MHC class Ib molecules might be involved in the presentation of p42–50. Indeed, those APCs expressing various class Ia alleles capable of stimulating 2D11 all express Qa-1a whereas those incapable of presentation express Qa-1b. These data indicate that Qa-1a molecules may present p42–50 to the CD8 Tregs. This is consistent with the finding that B10.PL and PL/J mice from which the CD8 Treg clones were isolated express Qa-1a but not Qa-1b molecules (data not shown). To further validate Qa-1a presentation, we next compared presentation of p42–50 by Qa-1 congenic mice B6.Tlaa (Qa-1a) and B6 (Qa-1b). Data in the lower panel of Fig. 5A show that APCs from B6.Tlaa, but not B6, can present peptide p42–50. Likewise, the 2D11 clone showed specific cytotoxicity toward p42–50-pulsed T2 Qa-1a transfectants but not T2 cells (data not shown).
|
|
) but not in the presence of an irrelevant class II-binding peptide MBPAc1–9 (
) or a class Ia-binding OVA peptide (data not shown). In contrast, responses of the OT-1 clone (class Ia restricted, middle panel) and an MBPAc1–9-reactive CD4 T cell clone (class II restricted, lower panel) are not blocked by the Qdm peptide. To exclude the possibility that Qa-1 restriction was not just a property of this particular 2D11 clone, we examined whether the Qdm peptide could competitively inhibit a polyclonal response to p42–50. As shown in Fig. 5C, Qdm, but not an irrelevant peptide, blocks the p42–50 response of draining lymph node cells derived from p42–50-immunized mice.
The above experiments strongly indicate that p42–50 is presented by the Qa-1a molecule to the CD8 Tregs. To directly demonstrate binding of p42–50 to Qa-1a molecules, a binding assay was performed using purified recombinant Qa-1a molecules. Both peptides p42–50 and Qdm were biotinylated (b-p42–50 and b-Qdm), and a fluorescence tag was added for their visualization and quantification in an in vitro binding assay as described earlier (25). As shown in the upper panel of Fig. 6, both b-Qdm and b-p42–50 bind to Qa-1a, and their binding is competed by unlabeled Qdm (100 µM). Interestingly, a 10-fold higher concentration (10 µM) of p42–50, compared with Qdm, was required for comparable binding to Qa-1a. This suggests that p42–50 may have either a lower binding affinity or a higher off-rate compared with Qdm. Accordingly, a 2-fold higher concentration (200 µM) of unlabeled p42–50 was necessary for blocking the binding of b-p42–50 to Qa-1a molecules (lower panel of the Fig. 6).
|
Tregs
Inability of CD8 Tregs to kill TCR V8.2+ T cell hybridomas that were generated by fusion of MBPAc-1–9-reactive CD4+ T cells (Qa-1a+) with a tumor cell line BW5147 (Qa-1b+) led us to investigate whether this is due to an inefficient expression of Qa-1 on T cell hybridomas in comparison to freshly activated T cells or clones. Therefore, we examined the expression of Qa-1a and Qa-1b in the hybridomas by real-time PCR (Qa-1a and Qa-1b) and FACS analysis (Qa-1b only). Although the TCR V8.2+ CD4 hybridomas (172.10 and 1934.4) express mRNA for both Qa-1a and Qa-1b as expected, the expression is decreased greatly compared with freshly activated polyclonal splenocytes or V8.2+ CD4 T cell clone (Fig. 7A). Using anti-Qa-1b mAb, we further analyzed the surface Qa-1b expression. As shown in Fig. 7B, the Qa-1b expression on the two TCR V8.2+ hybridomas (172.10 and 1934.4) was reduced greatly compared with naive splenocytes and Con A blasts.
|
Treg clones provides protection from EAE
To determine the in vivo regulatory potential of the CD8 Tregs, 2D11 cells were adoptively transferred i.v. into syngeneic mice. The following day recipients were immunized with MBPAc1–9/CFA/PT for the induction of EAE, and clinical symptoms were monitored daily and scored on a scale from 1 to 5 as described previously (22). As shown in the right panel of Fig. 8, mice injected with 1 million 2D11 cells recover more rapidly from a milder disease than those in the control group. To rule out the possibility that the ability to control disease is an artifact of the long-term culture of a particular clone, short-term p42–50-reactive T cell lines were generated and used in similar adoptive transfer experiments. As shown in the left panel of the Fig. 8, mice that receive 5 million cells of a p42–50-reactive T cell line are completely protected from MBP-induced EAE, and transfer of only 1 million cells enables a more rapid recovery from a milder paralysis in comparison to animals in the control group. Adoptive transfer of short-term T cell lines raised against two irrelevant TCR peptides had no effect on EAE (Ref. 22 and data not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
+TCR+ T cells (CD8 Tregs) that are involved in a negative feedback mechanism for the regulation of pathogenic CD4+ T cells reactive to a self-Ag. Our findings suggest a mechanism of active immune regulation wherein CD8 Tregs kill only activated T lymphocytes through specific recognition of a TCR-derived peptide bound to class Ib MHC molecules. The CD8 Tregs described in this study are CD8+ and TCR + and recognize a TCR peptide from the conserved CDR2 region of the V8.2 chain in the context of the class Ib molecule Qa-1a. They are cytotoxic, secrete Tc1-like cytokines, and are able to kill activated but not resting V8.2+ CD4 T cells. Adoptive transfer of p42–50-reactive CD8 Tregs, as well as in vivo activation of the CD8 Tregs following immunization with p42–50, protects mice from Ag-induced EAE. Since this immune regulatory mechanism operates following activation and expansion of CD4+ T cells, it is likely that CD8 Tregs of the type characterized in this study participate in the contraction phase following a productive immune response to a self- or foreign Ag, thereby contributing to immune homeostasis.
The degradation of a cytosolic protein into peptides for MHC class I presentation is a process that includes cleavage of the protein by proteosomes to produce a series of peptides that can be transported into the endoplasmic reticulum by TAP. Mouse TAP preferably transports peptides with hydrophobic carboxy termini (38). These peptides are further trimmed by aminopeptidase in the endoplasmic reticulum before binding to the MHC class I molecules (45, 46). Upon stimulation with IFN- or TNF-, cells synthesize three proteosome subunits, low-molecular-mass polypepticle2 (1i), low-molecular-mass polypepticle7 (5i), and multicatalytic endopeptidase complex-like 1 (2i), inducing the assembly of immunoproteosomes that generate new antigenic determinant (47). We have shown earlier that IFN- secretion by CD4 Tregs is required for the recruitment and expansion of the CD8 Tregs that ultimately kill the V8.2+ pathogenic T cell population and protect mice from EAE (48). It is noteworthy that one of the optimal cleavage sites within the TCR V8.2 chain for immunoproteosomes (49) lies immediately following the tyrosine at position 9 in p42–50. We postulate that IFN- secreted by the CD4 Tregs would enhance the presentation of p42–50 in both the target V8.2+ T cells, increasing their sensitivity to killing by the CD8 Tregs, as well as in dendritic cells promoting Ag cross-presentation.
Although earlier studies have implicated either class Ia or Ib MHC restriction of the CD8 Tregs, they have remained inconclusive because of the lack of a defined peptide and nonavailability of functional CD8 Treg clones (15, 23, 24). Our data clearly indicate that a TCR-derived peptide p42–50 (GLRLIHYSY) binds to the Qa-1a class Ib molecule and is recognized by the CD8 Tregs. Although p42–50 does not possess the anchor residues suggested by the Qdm peptide (AMAPRTLLL) for binding to Qa-1a or b (50), it contains a number of residues present in other Qa-1-binding peptides (25). The leucine and histidine residues in p42–50 at P2 and P6, respectively, are also the predominant residues at the same positions in Qdm variant pools (25). Similarly, consistent with other Qa-1-binding peptides (28), hydrophilic and nonpolar or hydrophobic residues, at P3 (arginine) and P7/P9 (tyrosine), respectively, are present in p42–50. It is clear from our data (Fig. 6) that p42–50 binds with lower affinity to Qa-1a than the Qdm peptide. This relatively poor but still functional binding property of p42–50 is reflected in the dose needed to stimulate CD8 Treg clones, which is similar to a Qa-1b-restricted clone specific for an insulin peptide (51).
Besides the Qa-1a-restriction, another important feature of the CD8 Tregs described in this study is the expression of the CD8 homodimer as demonstrated by the binding of TL-tetramer (Fig. 3). Most CD8+TCR+ T cells reside in IELs (CD8 IELs) and are present in mice genetically deficient in class Ia MHC molecules (52, 53), suggesting their potential restriction by class Ib MHC. Our p42–50-reactive CD8 Tregs share several characteristics with CD8 IELs (54). First, our CD8 Tregs expand poorly in response to Ag stimulation both in vitro and in vivo (X. Tang and V. Kumar, unpublished observation), comparable with the poor response of CD8 IELs to self-Ags (55). Second, both CD8 Tregs and CD8 IELs display regulatory function (54). Experiments are in progress to examine whether the CD8 Tregs described here represent circulating CD8 IELs or are part of a separate lineage.
Activation-induced cell death and/or passive cell death of activated T cells in the periphery are important mechanisms for maintaining homeostasis in the immune system and peripheral tolerance (1). Fas and/or Fas ligand knockout mice have a normal contraction phase in CD8+ T cells following their activation, whereas in perforin knockout mice, a large number of activated T cells remain (56). Therefore, perforin is likely to be involved in the death of activated T cells. In fact, a perforin gene defect has been associated with familial hemophagocytic lymphohistiocytosis, which is characterized by uncontrolled activation of T cells (57), and our own preliminary experiments indicate that CD8 Treg-mediated killing of activated V8.2+ CD4 T cells involves the perforin but not the Fas-Fas ligand pathway (T. Beeston, I. Maricic, X. Tang, and V. Kumgr, manuscript in preparation). Taken together, these observations suggest that, following the expansion phase of a dominant T cell clone(s), a CD8 Treg response is cross-primed by dendritic cells that have captured apoptotic T cells and present TCR-peptides to both CD4 and CD8 T cells. Thus, primed TCR-peptide-reactive CD8 Tregs will contribute to the clearance of the activated effectors and further recruitment of Tregs (58). This is consistent with the depletion of Ag-reactive T cells and the control of autoimmunity and transplant rejection following TCV with irradiated autologous T cells (15, 17, 18). Thus, CD8 Tregs may control effector lymphocytes through both direct killing (6) or by modulation of APCs (59). Killing of CD4 T cells by the CD8 Tregs is restricted to activated but not resting V8.2+ or V8.2– CD4 T cells (Fig. 4C). Since both the V8.2+ and V8.2– CD4 T cells have identical Qa-1a alleles, it is likely that the CD8 Tregs discriminate among cells by differential recognition of TCR peptide/Qa-1a complexes on the cell surfaces. In contrast to MHC class Ia molecules, Qa-1b expression is low on the surface of resting T cells and up-regulated substantially upon activation (60) (Fig. 7B). Resting T cells may not be targeted due to poor display of TCR-peptide/Qa-1 complexes on the cell surface (Fig. 4C). There also exists a correlation between the ability of CD8 Tregs to kill and the level of expression of Qa-1 molecules on target T cells. Thus, CD8 Tregs can kill the T cell clones but not the hybridomas expressing the same TCR V/V chains (Tables II and III) but having much reduced expression of Qa-1 on their cell surface (Table III and Fig. 7). Similarly, naive V8.2+ T cells that express very low levels of Qa-1 are not subjected to killing by the CD8 Tregs. Thus, a dynamic expression of TCR peptide/Qa-1 complexes ensures that T cells are only targeted for killing by the CD8 Tregs within a narrow window of time.
Why are p42–50-reactive CD8 Tregs not negatively selected in the thymus? There are at least four possibilities. First, Qa-1 expression is generally much lower in comparison to class Ia molecules. Second, it appears that p42–50 can only be processed when immunoproteosomes are present (X. Tang and V. Kumar, unpublished observation). Third, we have found that cross-presentation of TCR-peptide to CD8 Tregs by immature DC is very inefficient (T. R. F. Smith, X. Tang, and V. Kumar, unpublished data) (58). Fourth, the CD8 Treg TCR has a low avidity for p42–50/Qa-1a complexes as manifested by a requirement for relatively higher concentrations of stimulating peptide and the relative instability of peptide MHC-Ib complexes (61). It has been shown that positive selection of MHC class Ib-restricted CD8 T cells reactive to insulin or CD8 IELs can be mediated by recognition of agonistic ligands on thymic epithelial or hemopoietic cells (51). Therefore, it will be interesting to investigate whether TCR peptides act as agonistic ligands for positive selection of the CD8 Tregs in the thymus similar to the positive selection of CD4+CD25+ Tregs following recognition of high-affinity self- ligands (2).
Since CD8 Tregs recognize a determinant from a conserved region of the TCRV8.2 chain rather than a clone-specific variable CDR3 region, this enables them to control activated T cells that use the same TCRV chain but may have different Ag specificity. Therefore, this regulation is not anti-idiotypic in the classical sense. A single CD8 Treg population could potentially regulate activated T cells of differing Ag specificities as long as they express the same TCR (the p42–50 sequence of V8.1, V8.2, or V8.3 is identical). Thus, clonal dominance among responding CD4 effectors and specificity of the CD8 Tregs for a conserved region of TCRs make these regulators highly effective and independent of minor TCR sequence variations on the target CD4+ T cells (6, 58, 62). We suggest that CD8 Tregs of the type described here may represent part of the CD8 Treg population involved in the prevention of autoimmune diseases, transplant tolerance, neonatal tolerance, oral tolerance, and homeostasis of immune responses in general (9, 10, 11, 59, 63, 64).
It is interesting to note that although CD8 Tregs secrete Tc-1-like cytokines, they do not secrete detectable level of IL-2 and thus may be dependent upon the provision of IL-2 from other cells, including activated pathogenic CD4 T cells. This dependency of the CD8 Tregs on IL-2 is reminiscent of CD25+CD4+ Tregs (2). Thus, IL-2 secreted by conventional CD4 T cells upon exposure to Ags serves to attract or even expand both Treg subsets that in turn inhibit the proliferative response of (CD25+CD4+ Tregs) or kill (CD8 Tregs) activated effector T cells in a negative feedback loop. Since the priming and expansion of the CD8 Tregs depend upon the cross-presentation of TCR peptide by activated dendritic cells and the armed CD8 Tregs act only on activated T cells, it is a perfect mechanism to limit uncontrolled T cell expansion following a productive immune response.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported by grants from the National Institutes of Health, National Multiple Sclerosis Society, Alzheimer’s and Aging Research Center, and Multiple Sclerosis National Research Institute (to V.K.). 
2 Address correspondence and reprint requests to Dr. Vipin Kumar, Laboratory of Autoimmunity, Torrey Pines Institute for Molecular Studies, 3550 General Atomics Court, San Diego, CA 92121. E-mail address: vkumar{at}tpims.org
3 Abbreviations used in this paper: Treg, regulatory T cell; TCV, T cell vaccination; MBP, myelin basic protein; EAE, experimental autoimmune encephalomyelitis; 2m, 2-microglobulin; Tg, transgenic; PT, Pertussis toxin; TL, thymic leukemia; TAP, transporter associated with Ag processing; IEL, intraepithelial lymphocyte.
Received for publication March 29, 2006. Accepted for publication September 15, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
