Autoantigen-specific regulatory T cells, a potential tool for immune-tolerance reconstitution in type-2 autoimmune hepatitis

Authors

  • Maria Serena Longhi,

    1. Institute of Liver Studies, King's College London School of Medicine at King's College Hospital, Denmark Hill, London, UK
    Search for more papers by this author
  • Munther J. Hussain,

    1. Institute of Liver Studies, King's College London School of Medicine at King's College Hospital, Denmark Hill, London, UK
    Search for more papers by this author
  • William W. Kwok,

    1. Benaroya Research Institute at Virginia Mason Medical Center, Seattle, WA
    Search for more papers by this author
  • Giorgina Mieli-Vergani,

    1. Institute of Liver Studies, King's College London School of Medicine at King's College Hospital, Denmark Hill, London, UK
    Search for more papers by this author
  • Yun Ma,

    1. Institute of Liver Studies, King's College London School of Medicine at King's College Hospital, Denmark Hill, London, UK
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Diego Vergani

    Corresponding author
    1. Institute of Liver Studies, King's College London School of Medicine at King's College Hospital, Denmark Hill, London, UK
    • Institute of Liver Studies, King's College Hospital, Denmark Hill, London SE5 9RS, UK
    Search for more papers by this author
    • These authors contributed equally to this work.

    • fax: +44 20 32993700


  • Potential conflict of interest: Nothing to report.

Abstract

Effector CD4 and CD8 T cell immune responses to cytochrome P450IID6 (CYP2D6), the autoantigen of autoimmune hepatitis type 2 (AIH-2), are permitted by a numerical and functional impairment of CD4posCD25high regulatory T cells (T-regs). We aimed to investigate whether T-regs specific for CYP2D6 immunodominant regions and restricted by the appropriate human leukocyte antigen (HLA)-DR molecule can be generated in patients with AIH-2 and can control CD4 and CD8 T cell effectors targeting identical or overlapping CYP2D6 regions. CYP2D6-specific regulatory T cells (CYP2D6 T-regs) were obtained from peptide-pulsed monocyte-depleted peripheral blood mononuclear cells of 17 patients with AIH-2, who were positive for the predisposing HLA-DR7 and/or HLA-DR3 alleles. Their antigen specificity was assessed by cytofluorimetry using HLA class II tetramers and their cytokine profile by intracellular staining. T-reg ability to suppress was ascertained by measuring reduction of CD4posCD25neg cell proliferation/effector cytokine secretion and of CD8 T cell cytotoxicity. The most efficient suppression of effector T cell proliferation, inflammatory cytokine release, and cytotoxicity was obtained by coculturing T-regs with CYP2D6-peptide-loaded semimature dendritic cells (smDCs), and smDC-CYP2D6 T-regs also expressed high levels of FOXP3 (forkhead box P3). Possession of the appropriate HLA-DR molecule and recognition of the CYP2D6 autoantigenic sequence were critical to the synergistic smDC-CYP2D6 T-reg immunoregulatory functions, and lack of either element led to poor control of responder cell proliferation and cytokine secretion. Moreover, interferon-γ neutralization significantly boosted the suppressive ability of CYP2D6 T-regs. Conclusion: T-regs generated under CYP2D6-specific conditions and cocultured with smDCs are highly effective at controlling autoreactive T cells, thus providing the basis for a powerful and tailored form of immunotherapy for AIH-2. (HEPATOLOGY 2011;53:536-547)

See Editorial on page 385.

Autoimmune hepatitis type 2 (AIH-2) is a juvenile autoimmune liver disorder associated with the possession of human leukocyte antigen (HLA)-DRB1*0701 (DR7) and HLA-DRB1*0301 (DR3) alleles, characterized by a serious course at times progressing to cirrhosis and liver failure despite immunosuppressive treatment. Cytochrome P450IID6 (CYP2D6), a 50-kDa liver enzyme,1 is the AIH-2 main autoantigen, recognized by liver-kidney-microsomal type 1 (LKM-1) autoantibody. Two CYP2D6 regions, spanning the 217-260 and 305-348 amino acid sequences,2 are targets of effector B, CD4, and CD8 T cell immune responses,2-4 the extent of which parallels biochemical degree of liver damage.2, 4 AIH-2 is also characterized by numerical and functional impairment of polyclonal CD4posCD25high regulatory T cells (T-regs),5, 6 a subset central to immune homeostasis because they are able to suppress effector T cell immune responses of diverse specificity.7-9 Such impairment, however, can be at least partially overcome, because T-regs from patients with AIH undergo polyclonal expansion upon stimulation in vitro, suggesting that adoptive transfer of these expanded T-regs could have a therapeutic role.10

There is, however, increasing evidence that induction and maintenance of organ-specific tolerance is critically dependent on antigen-specific T-regs, which not only express higher levels of forkhead box P3 (FOXP3), the lineage-specific transcription factor, but also suppress more efficiently than their non–antigen-specific counterparts.11, 12 Autoantigen-specific T-regs revert diabetes in NOD (nonobese diabetic) mice11 and prevent relapse of experimental autoimmune encephalomyelitis when administered after disease onset.12 Generation of T-regs controlling CD4 and CD8 T cells targeting a specific antigenic sequence would be highly relevant to reestablishing and maintaining immune tolerance in patients with AIH-2.

Of the various approaches available for generating antigen-specific T-regs,13, 14 the most effective at expanding Foxp3-positive cells and promoting T-reg antigen specificity in mice is coculture with peptide-pulsed semimature dendritic cells (smDCs).15-17 SmDCs represent a transitional stage of maturation between immature (iDCs) and mature DCs (mDCs), and are characterized by marked up-regulation of major histocompatibility complex (MHC) class I and MHC class II molecules and by medium levels of costimulatory molecules. The smDCs present antigens for induction of tolerance, as shown in a murine model of autoimmune encephalomyelitis where autoantigen-peptide–pulsed smDCs are able to prevent disease by inducing peptide-specific interleukin-10 (IL-10)-producing CD4+ T cells.15

The aim of the present study was to generate and characterize T-regs specific for CYP2D6 sequences identical or overlapping with those targeted by effector CD4 and CD8 T cells and restricted by the appropriate HLA molecule in AIH-2, and to explore how their suppressive function in an antigen-specific setting is influenced by the presence of smDCs.

Abbreviations

AIH, autoimmune hepatitis; APC, allophycocyanin; CYP2D6, cytochrome P4502D6; FITC, fluorescein isothiocyanate; FoxP3, forkhead box P3; HLA, human leukocyte antigen; iDC, immature dentritic cell; IL, interleukin; IFN, interferon; LKM-1, liver-kidney microsomal antibody type 1; mAb, monoclonal antibody; mDC, mature dendritic cell; MFI, mean fluorescence intensity; MHC, major histocompatibility complex; PBMC, peripheral blood mononuclear cell; PE, phycoerythrin; SEM, standard error of the mean; smDC, semimature dendritic cell; T-reg, T regulatory cell; TGF, transforming growth factor.

Patients and Methods

Patients and Controls

Peripheral blood mononuclear cells (PBMCs) were obtained from 22 patients with AIH-2 (16 females; median age: 13.6 years; range: 0.9-29.4 years). A liver biopsy performed close to diagnosis showed interface hepatitis in all. Thirteen patients were DR7pos, eight were DR3pos, including four positive for both DR7 and DR3; five patients were negative for both alleles. Three patients were studied before immunosuppressive treatment was started and again during treatment; 19 patients were studied during treatment only. Of the 22 patients studied during treatment, 18 were in remission (i.e., normal aminotransferase levels), and seven were in relapse, including three studied on both occasions. Treatment consisted of prednisolone (2.5-5 mg/daily at remission and 1-2/mg/kg/day at relapse) and azathioprine (1-2 mg/kg/day). As controls, PBMCs were obtained from six healthy subjects (all females; median age: 28; range: 22-35 years), including two who were DR7pos, one who was DR3pos, and three who were DR7neg and DR3neg. Written consent was obtained for each subject. The study was approved by the Ethics Committee of King's College Hospital, London, UK. Demographic and laboratory data at the time of study are shown in Table 1.

Table 1. Demographic and Laboratory Data
Study GroupNumber of SubjectsSex (F/M)Age (Years)HLA-DRAspartate Aminotransferase (nv <50 IU/L)Bilirubin (nv < 20 mol/L)Immunoglobulin G(nv = 6.5-17 g/L)Autoantibody Titer (Reciprocal)
DR7DR3DR7+/ DR3+DR7–/ DR3–ANASMALKM-1
  • Data are presented as range (median) unless noted otherwise; nv: normal value. ANA, antinuclear antibody; SMA, smooth muscle antibody

  • *

    Individual values of aspartate aminotransferase, bilirubin, immunoglobulin G, and LKM-1 titer are given.

  • Including three patients studied both at diagnosis and during treatment.

  • ANA autoantibody titer in two patients.

  • §

    SMA autoantibody titer in two patients.

  • Including three treated patients studied both during remission and relapse.

AIH2216/60.9-29.4 (13.6)9445 
At diagnosis3*   226, 525, 2237212, 147, 47813.7, 23.4, 15.8negneg1280,640, 160
During treatment22         
Inactive disease18   12-48 (24)3-39 (7)3.1-16.6 (9.8)10,16020,40§20-2560 (160)
Active disease7   61-608 (163)6-673 (15)6.6-16.8 (11.5)negneg40-1280 (160)

Cell Separation

PBMCs were obtained as previously described.18 Mononuclear cell viability, determined by Trypan blue exclusion, always exceeded 98%.

Generation of CD4posCD25highT-CellsUnderCYP2D6-Specific Conditions UsingsmDCs

The smDCs were generated from monocytes, isolated by positive selection from PBMCs using CD14 microbeads (Miltenyi Biotec, Bergisch-Gladbach, Germany)13 and cultured for 5 days in the presence of granulocyte-monocyte-colony-stimulating factor (GM-CSF, 20 ng/mL) and IL-4 (50 ng/mL). Cells were then treated overnight with 3 nM interferon lambda (interferon [IFN]-λ; PeproTech, Inc, Ltd., London, UK) and pulsed for the following 48 hours with individual peptides (two DR7-restricted, CYP2D6193-212 and CYP2D6305-324, and two DR3-restricted, CYP2D6313-332 and CYP2D6393-412), previously shown to elicit proliferation and IFN-γ secretion by CD4 T cells.2 The DR7-restricted hemagglutinin (HA)217-236 peptide19 was used as control. Peptide sequences are shown in Supporting Information Table 1. Flow cytometry analysis using fluorescein isothiocyanate (FITC)-conjugated anti-CD14, and phycoerythrin (PE)-conjugated anti-CD80, anti-CD86 (BD Bioscience, San Jose, CA) and anti–blood dendritic cell antigen-1 (BDCA-1) monoclonal antibodies (mAbs) (Miltenyi Biotec), confirmed the partially mature phenotype of smDCs,13 i.e., CD14low, CD80low, CD86med, and BDCA-1pos, successfully obtained from all patients with AIH-2, regardless of disease stage.

CYP2D6 T-regs and CYP2D6-CD25neg target cells (CYP2D6-responders) were obtained from the CD14-depleted cell fraction that contains B lymphocytes capable of presenting antigen to CD4 T cells. CD14-depleted cells were cultured for 10 days with 10 μmol CYP2D6 or HA217-236 peptide,2 recombinant IL-2 (rIL-2, 300 U/mL; Eurocetus Amsterdam, Netherlands) and anti-CD3/anti-CD28 T cell expander (Dynal Invitrogen, Oslo, Norway; four beads/cell) to yield sufficient numbers of T-regs for investigation. At day 10, T-regs and responder cells were isolated using immunomagnetic beads.5, 6, 18 In all subsequent experiments CYP2D6 T-regs were used either alone or after 2-day coculture with smDCs (smDC-CYP2D6 T-regs, see below). A schematic representation of how smDCs, CYP2D6 T-regs, smDC-CYP2D6 T-regs, and CYP2D6-responders were obtained is provided in Supporting Information Fig. 1.

Figure 1.

Tetramerpos T-regs. Mean ± SEM frequency of Tetpos cells within (A) PBMCs and (B) CYP2D6 T-regs. CYP2D6 T-regs were purified from CD14neg cells cultured for 10 days in the presence of CYP2D6 peptide, high-dose IL-2, and T cell expander in: DR7pos and/or DR3pos patients (black bars), DR7neg/DR3neg patients (gray bars), and DR7pos or DR3pos healthy controls (white bars). HS, healthy subjects. (C) Dot plots of CYP2D6-tetramer-PE (y axis) versus anti-CD25-APC (x axis). Cells are gated on CD4 lymphocytes. (D) Frequency of Tetpos FOXP3pos cells within CYP2D6 T-regs from a DR7pos patient. Dot plot of CYP2D6-tetramer-PE (y axis) versus FOXP3-FITC (x axis). Asterisk (*) denotes after culture with CYP2D6 peptide, high-dose IL-2 and T cell expander; pilcrow symbol (¶) denotes after culture with high-dose IL-2 and T cell expander; section symbol (§) denotes after culture with HA peptide, high-dose IL-2 and T cell expander.

Phenotype of CYP2D6 T-Regs

T-reg phenotype was assessed by flow cytometry (FACSCalibur with CellQuest; BD Bioscience, San Jose, CA) using FITC-conjugated anti-CD4, PE-conjugated anti-CD25, anti-CD45RO, anti-CD62L, and anti-CD127 mAbs (BD Bioscience). Expression of chemokine (C-X-C motif) receptor 3 (CXCR3), a chemokine receptor which mediates the adhesion and transendothelial migration of T cells across the hepatic endothelium20 and whose ligands are expressed at high level by hepatocytes in patients with AIH,21 was investigated using allophycocyanin (APC)-conjugated anti-CD183 (CXCR3) mAbs (BD Bioscience) and expressed as mean fluorescence intensity (MFI). Intracellular FOXP3 staining was performed as previously reported.10 The frequency of IFN-γ, IL-2, IL-4, IL-10, transforming growth factor-beta (TGF-β), and IL-17–producing cells within CYP2D6 T-reg subsets was determined by intracellular cytokine staining,6 after staining with FITC- or PE-Cy5 conjugated anti-CD4, PE-conjugated or APC-conjugated anti-CD25, and APC-conjugated anti-IFN-γ (IQ Products, Groningen, the Netherlands), anti–IL-2, anti–IL-10 (BD Bioscience), FITC-conjugated anti–IL-4, anti–IL-17 (eBioscience), PE-conjugated anti–IL-17 (eBioscience) and PerCP-conjugated anti-human LAP (TGF-β1) (R&D Systems) mAbs. The frequency of T-reg binding to CYP2D6 peptides was assessed using PE-conjugated DR7-restricted and DR3-restricted tetramers of CYP2D6193-212, CYP2D6305-324, and CYP2D6313-332 or CYP2D6393-412 peptide. Tetramer construction and cell staining were carried out as described.14, 22

Inhibition of Responder Cell Proliferation

The ability of T-regs grown under CYP2D6-specific conditions to suppress the proliferation of CYP2D6-responders was tested in 17 patients (three both at disease presentation and during treatment, 14 during treatment only), 12 DR7pos, seven HLA-DR3pos, including three positive for both alleles; two DR7neg/DR3neg patients and six healthy subjects (two DR7pos, one DR3pos, and three DR7neg/DR3neg). Following purification, T-regs were rested for 48 hours in T cell-expander–free and IL-2–free medium either alone or in the presence of smDCs treated with mytomycin (Sigma-Aldrich, Ltd., Gillingham, UK), then added to autologous CYP2D6-responders (5×104 cells/well; ratio between T-regs and responders = 1:8; between smDCs and responders = 1:10). Proliferation of peptide-pulsed responder cells was induced by low-dose IL-2 and T cell expander.18 After 5 days, cocultures were pulsed for 18 hours with 0.25 μCi/well [3H]thymidine. In control experiments, iDCs and mDCs were also used, the latter obtained by treating iDCs, generated from monocytes after treatment with GM-CSF and IL-4, with IFN-1β (100 units/mL; PeproTech, Inc, Ltd.).23 In addition, the following control cocultures were performed: CYP2D6 T-regs (±smDCs) + polyclonal responders (obtained as previously described10; polyclonal T-regs + CYP2D6-responders; HA217-236 T-regs (±smDCs) + CYP2D6-responders; CYP2D6 T-regs (±smDCs) + HA217-236-responders. All coculture experiments were performed in duplicate.

CYP2D6 T-RegTetramer Sorting.

In five DR3pos patients with sufficient number of cells, Tetramerpos T-regs were sorted to a purity exceeding 90% into Tethigh and Tetlow fractions using a BD FACSAria cell sorter (BD Bioscience). The effect of Tethigh and Tetlow CYP2D6 T-regs, alone or after coculture with smDCs, on CYP2D6-responder cells was assessed by proliferation.

Neutralization Assay.

the effect of IFN-γ blockage on T-reg suppressor function was measured by adding anti–IFN-γ (10 μg/mL; R&D Systems, Abingdon, UK) mAbs to CYP2D6 T-regs (±smDCs) before coculture with CYP2D6-responder cells.

Cytokine Measurement by Enzyme-LinkedImmunosorbentAssay

IFN-γ, IL-17, and IL-4 released by responder cells following addition of CYP2D6 T-regs (±smDCs) was determined by enzyme-linked immunosorbent assay18 in 13 patients (all treated; seven DR7pos, three DR3pos, and three DR7neg/DR3neg) and six healthy controls (two DR7pos, one DR3pos, and three DR7neg/DR3neg).

Inhibition of CD8 T-CellCytotoxicity

The ability of CYP2D6 T-regs (±smDCs) to regulate CYP2D6-specific CD8 T cell cytotoxicity was assessed using a chromium release assay in eight HLA-A2pos patients (all treated), including one DR7pos, two DR3pos, two DR7pos/DR3pos, and three DR7neg/DR3neg. CD8 T cell lines specific for CYP2D6393-402 and CYP2D6306-314, immunodominant HLA-A2–restricted CYP2D6 CD8 T cell epitopes,4 were generated as described.4 Purified CYP2D6305-324 or CYP2D6393-412 T-regs were added, alone or with smDCs, to the CD8 T cell lines specific for CYP2D6306-314 or CYP2D6393-402. To test the ability of CYP2D6 T-regs (±smDCs) to exert bystander suppression, i.e., to inhibit the cytotoxic activity of CD8 T cell lines specific for nonhomologous CYP2D6 regions, CYP2D6305-324, CYP2D6313-332, and CYP2D6393-412 T-regs were added, alone or with smDCs, to CYP2D6393-402, and CYP2D6306-314 CD8 T cell lines. Controls consisted of CYP2D6 CD8 T cell lines cultured alone; or HA217-236 T-regs cultured with CD8 T cell lines specific for CYP2D6306-314 or CYP2D6393-402 peptide. Cytotoxicity of T2 cells, loaded with the corresponding HLA-A2–restricted CYP2D6 peptide (10 μmol), was carried out as described.4

Statistical Analysis

Comparison was performed by paired and unpaired Student t tests if values were normally distributed and by Wilcoxon rank sum test or Mann-Whitney test if values did not follow a normal distribution. A value of P< 0.05 was considered significant. Results are expressed as mean ± standard error of the mean (SEM).

Results

T-RegsGenerated Under CYP2D6-Specific Conditions

The mean percentage of CD25high cells within the total CD4 T cell pool was similar in DR7pos and/or DR3pos patients (23.6 ± 2.09), and in the two control groups, being 22.4 ± 3.6 in DR7neg/DR3neg patients and 21.2 ± 2.7 in DR7pos/DR3pos healthy subjects.

In preliminary experiments, CYP2D6-responder cell proliferation was unaffected by addition of CYP2D6 T-regs cocultured with iDCs, whereas it was inhibited by 20% by addition of CYP2D6 T-regs cocultured with mDCs, and by 54.8% by addition of CYP2D6 T-regs cocultured with smDCs (P< 0.001, compared to baseline proliferation; P = 0.025, compared to proliferation in the presence of iDCs; P = 0.05, compared to proliferation in the presence of mDCs). Based on these results, subsequent experiments were carried out using smDCs. CYP2D6 T-regs (±smDC) retained T-reg markers, being CD45ROhigh, CD62Lhigh, and CD127low.

Frequency of CYP2D6 tetramerpos (Tetpos) T-regs within PBMCs was higher in DR7pos and/or DR3pos patients than in controls (Fig. 1A). Frequency of Tetpos cells within the CYP2D6 T-regs was higher than among PBMCs in DR7pos and/or DR3pos patients (P< 0.001) and in controls (DR7neg/DR3neg patients: P = 0.005; DR7pos or DR3pos healthy controls: P = 0.036), remaining higher in the former than in the latter (Fig. 1B). Figure 1C shows the frequency of Tetpos cells within PBMCs and within purified CYP2D6 T-regs after culture with antigenic peptide in representative DR7pos and DR3pos patients and controls. Figure 1D shows the frequency of Tetpos FOXP3pos cells within CYP2D6 T-regs cultured with antigenic peptide from a DR7pos patient.

Suppressor Function of CYP2D6 T-Regs and smDC-CYP2D6 T-Regs

In DR7pos patients during treatment, the proliferation of CYP2D6193-212 and CYP2D6305-324 responder cells was reduced by 30.6% and 21.1%, respectively, after addition of CYP2D6193-212 and CYP2D6305-324 T-regs on their own and by 53.9% and 56.2%, respectively, when smDCs were also added (Fig. 2A).

Figure 2.

Suppression of CYP2D6-responder cell proliferation by CYP2D6 T-regs (±smDCs). (A) DR7pos and (B) DR3pos patients studied during treatment (individual CYP2D6 peptide results). (C) Combined data from three DR7pos and/or DR3pos patients studied both at disease presentation (left) and during treatment (right) (results for the four CYP2D6 peptides are pooled). Mean ± SEM proliferation of CYP2D6-responder cells before (black bars) and after addition of CYP2D6 T-regs (gray bars) and smDC-CYP2D6 T-regs (white bars).

In DR3pos patients during treatment, the proliferation of CYP2D6313-332 and CYP2D6393-412 responder cells was reduced by 18% and 11%, respectively, following coculture with CYP2D6313-332 and CYP2D6393-412 T-regs on their own and by 38.3% and 81%, respectively, when smDCs were also added (Fig. 2B).

In three DR7pos patients (including two who were also DR3pos) studied at disease presentation, inhibition of proliferation was observed only when responder cells were cocultured with smDC-CYP2D6 T-regs (30.9%); in the same patients studied during treatment, responder cell proliferation was reduced by 12.3% after addition of CYP2D6 T-regs and by 46.6% after addition of smDC-CYP2D6 T-regs (Fig. 2C).

Among patients studied during treatment, no difference in inhibition of responder cell proliferation was observed between those with active or inactive disease, with smDC-CYP2D6 T-regs exerting stronger suppressor activity than CYP2D6 T-regs in both groups (P = 0.008 in inactive disease; P = 0.03 in active disease).

In DR7neg/DR3neg patients, inhibition of responder cell proliferation was negligible and significantly lower than in DR7pos and/or DR3pos patients (P = 0.01 when comparing the inhibition values obtained after addition of CYP2D6 T-regs; P = 0.0004 when comparing the inhibition values obtained after addition of smDC-CYP2D6 T-regs). No effect on responder cell proliferation was observed in DR7pos or DR3pos healthy controls after addition of CYP2D6 T-regs (±smDCs). Similar results were obtained in DR7neg/DR3neg healthy controls.

In five DR3pos patients, addition of Tetlow or Tethigh CYP2D6 T-regs led to a decrease of responder cell proliferation by 30.7% and 88%, respectively; these inhibition values increased to 71.8% and 95.6%, respectively, when Tetlow and Tethigh CYP2D6 T-regs were cocultured with smDCs. In these patients, addition of CYP2D6 T-regs and smDC-CYP2D6 T-regs before sorting decreased responder cell proliferation by 44.8% and 69.1%, respectively (Fig. 3).

Figure 3.

Suppression of CYP2D6-responder cell proliferation by Tetlow and Tethigh CYP2D6 T-regs. Mean ± SEM proliferation of CYP2D6-responder cells alone (black bar) and after addition of: CYP2D6 T-regs (gray bar), smDC-CYP2D6 T-regs (white bar), TetlowCYP2D6 T-regs (black diagonal bar), smDC-TetlowCYP2D6 T-regs (black horizontal bar), TethighCYP2D6 T-regs (gray diagonal bar), and smDC-TethighCYP2D6 T-regs (gray horizontal bar). Data refers to five DR3pos LKM-1pos patients. cpm, counts per minute.

Additional Control Experiments.

Negligible or no inhibition of responder cell proliferation was observed in control experiments (see Supporting Information Material).

Regulation of Cytokine Secretion

In DR7pos and/or DR3pos patients, the amount of IFN-γ and IL-17 released by CYP2D6-responder cells was reduced by 34.1% and 19%, respectively, following addition of CYP2D6 T-regs and by 52.7% and 76%, respectively, following addition of smDC-CYP2D6 T-regs (Fig. 4A,B), whereas the amount of IL-4 increased respectively by 22.8% and 98.5% (Fig. 4C). There was no difference in regulation of cytokine secretion between patients with inactive or active disease. Neither decrease in IFN-γ and IL-17 nor increase in IL-4 secretion was observed in pathological or healthy controls following addition of CYP2D6 T-regs in the absence or presence of smDCs and in all other control experiments (see Supporting Information Material).

Figure 4.

Regulation of cytokine secretion by CYP2D6 T-regs (±smDC). Plots refer to DR7pos and/or DR3pos patients. Mean ± SEM amount of (A) IFN-γ, (B) IL-17, and (C) IL-4 secreted by CYP2D6-responder cells before (black bars) and after addition of T-regs of the same CYP2D6 specificity alone (gray bars) or in the presence of smDCs (white bars).

Inhibition of CYP2D6-Specific CD8 T-CellCytotoxicity

In the HLA-A2pos patients studied, cytotoxicity of CD8 T cells specific for CYP2D6393-402 decreased by 57%, and 39% after addition of T-regs specific for CYP2D6393-412 or for the nonhomologous CYP2D6313-332 sequence, but it remained unchanged after addition of T-regs specific for the nonhomologous CYP2D6305-324 sequence. Addition of CYP2D6393-412 T-regs, CYP2D6313-332 T-regs, or CYP2D6305-324 T-regs after coculture with smDCs reduced the cytotoxicity of CYP2D6393-402-specific CD8 T cells by 70%, 44%, and 11.5%, respectively (Fig. 5A).

Figure 5.

Regulation of CD8 T cell cytotoxicity by CYP2D6 T-regs (±smDC). Cytotoxicity of CYP2D6393-402-specific and CYP2D6306-314-specific CD8 T cell lines is shown in (A) and (B). Mean ± SEM CD8 T cell cytotoxicity before T-reg addition (black bars) and after addition of CYP2D6 T-regs (dark pattern bars) or smDC-CYP2D6 T-regs (light pattern bars). CD8 T cell cytotoxicity in the presence of CYP2D6393-412 T-regs (square pattern bars), CYP2D6313-332 T-regs (vertical stripe bars) and CYP2D6305-324 T-regs (diagonal stripe bars) T-regs is shown. BS, bystander suppression. *P ≤ 0.05; **P ≤ 0.01; #P denoting a trend.

Cytotoxicity of CD8 T cells specific for CYP2D6306-314 remained unchanged following addition of CYP2D6305-324 T-regs but was reduced by 74% following addition of T-regs specific for CYP2D6313-332 and by 46% after addition of T-regs specific for the nonhomologous CYP2D6393-412 sequence. Addition of CYP2D6305-324 T-regs, CYP2D6313-332 T-regs, or CYP2D6393-412 T-regs after coculture with smDCs decreased the cytotoxicity of CYP2D6306-314-specific CD8 T cells by 19%, 80%, and 54%, respectively (Fig. 5B).

No inhibition of CD8 T cell cytotoxicity was observed in DR7neg/DR3neg patients after addition of CYP2D6 T-regs, alone or after coculture with smDCs and in DR7pos and/or DR3pos patients after addition of HA217-236 T-regs, alone or after coculture with smDCs.

Cytokine Profile of CYP2D6- and smDC-CYP2D6 T-Regs

CYP2D6 T-regs, with or without smDCs, contain a higher percentage of TGF-β-producing cells and a lower percentage of IFN-γ-, IL-2-, IL-17-, IL-4-, and IL-10–producing cells than CYP2D6-responder cells (P< 0.001 for all); the frequency of IFN-γ and IL-17–producing cells is lower within T-regs grown under CYP2D6-specific conditions than within polyclonally expanded T-regs. The lowest frequency of IFN-γ–producing cells was observed in the presence of smDC-CYP2D6 T-regs (Fig. 6). On the basis of these results, neutralization experiments were performed to evaluate whether IFN-γ blockage enhances CYP2D6 T-reg suppressor function. After IFN-γ block, the frequency of IFN-γ–producing cells decreased significantly in CYP2D6 T-regs (Fig. 7A). Anti-IFN-γ–treated CYP2D6 T-regs reduced the proliferation of CYP2D6-responder cells by 45.6%, with this inhibition value being higher than in the absence of neutralizing antibodies (23%) and similar to that obtained after addition of smDC-CYP2D6 T-regs (50%) (Fig. 7B). Blockage of IFN-γ did not further decrease the frequency of IFN-γ–producing cells in smDC-CYP2D6 T-regs and had no effect on their suppressor function (Fig. 7A,B).

Figure 6.

Cytokine profile of T-regs generated under CYP2D6-specific conditions. Mean ± SEM frequency of cytokine-producing cells within: CYP2D6 T-regs (vertical stripe bars), smDC-CYP2D6 T-regs (diagonal stripe bars), and polyclonal T-regs (gray bars) from DR7pos and/or DR3pos patients. Data from the four different CYP2D6 peptides are pooled.

Figure 7.

Effect of IFN-γ inhibition on CYP2D6 T-reg IFN-γ production and suppressor function. (A) Mean ± SEM frequency of IFN-γ-producing cells within: CYP2D6 T-regs (gray bar), smDC-CYP2D6 T-regs (white bar), anti-IFN-γ–treated CYP2D6 T-regs (white dot bar) and anti-IFN-γ–treated smDC-CYP2D6 T-regs (white square bar). (B) Proliferation of CYP2D6-responder cells alone (black bar) and following addition of the different T-reg subsets (bar pattern as in panel A). Data from the four CYP2D6 peptides are pooled.

Phenotype of CYP2D6 T-Regsand smDC-CYP2D6 T-Regs

Levels (MFI) of FOXP3 expression in DR7pos and/or DR3pos patients were higher in smDC-CYP2D6 T-regs (54.6 ± 6.2) than in CYP2D6 T-regs (40.9 ± 2; P = 0.016) and polyclonal T-regs (43.4 ± 5.2; P = 0.05), with levels in CYP2D6 T-regs and smDC-CYP2D6 T-regs being higher than in controls (CYP2D6 T-regs: 28.9 ± 2.9, P = 0.013 in DR7neg/DR3neg patients; 28.8 ± 1.1, P = 0.003 in healthy subjects; smDC-CYP2D6 T-regs: 34.1 ± 4.1, P = 0.009 in DR7neg/DR3neg patients; 32.9 ± 2.1, P< 0.001 in healthy subjects). The frequency of FOXP3pos cells and FOXP3 MFI in CYP2D6 T-regs (±smDCs) from a representative DR3pos patient are shown in Supporting Information Fig. 2.

In DR7pos and/or DR3pos patients, CXCR3 MFI was higher among smDCs-CYP2D6 T-regs (5581 ± 2238) than among CYP2D6 T-regs (3039 ± 342; P = 0.005) and polyclonally expanded T-regs (2744 ± 401; P = 0.01). No difference between CYP2D6 T-regs and polyclonally expanded T-regs cells was observed.

Discussion

This study provides evidence that regulatory T cells, generated under antigen-specific conditions, suppress more efficiently than polyclonal T-regs.

The role of the autoantigen in enhancing T-reg suppressor function is emphasized by the finding that T-regs with the highest affinity to the peptide/MHC class II complex were also the most effective at controlling target cell proliferation. Notably, we found that the suppressor function of the antigen-specific T-regs, including Tethigh and Tetlow cell fractions, is substantially boosted by coculture with smDCs loaded with homologous antigenic sequences. In our AIH-2 disease model, CD25high cells generated in the presence of the autoantigenic CYP2D6 peptides and exposed to CYP2D6-pulsed smDCs exert a strong control over proliferation, inflammatory cytokine production, and cytotoxicity, effector functions executed by CD4 and CD8 T cells targeting overlapping antigenic sequences. The strength of this immunoregulatory effect depends on the concurrence of two factors: possession of the appropriate HLA-DR molecule and recognition of the CYP2D6 autoantigenic sequence restricted by the same HLA-DR, lack of either factor leading to poor control of effector cell function in HLA-DR mismatched patients and in HLA-DR matched healthy subjects.

Cross-talk between HLA-DR and autoantigen does require the intervention of smDCs, which, as reported in a number of studies, enhance suppression while maintaining their ability to present autoantigen to T-regs.15-17 The ability of smDCs to promote antigen-specific T-reg suppressor function observed in the present investigation recalls an in vivo experimental study demonstrating that injections of peptide-pulsed smDCs induces the activation of peptide-specific Tr1 regulatory T cells and protects from the development of autoimmune encephalomyelitis.15 These findings were subsequently echoed in another murine study showing that thyroglobulin-pulsed smDCs block the development of autoimmune thyroiditis by activating thyroglobulin-specific T-regs.16

The T-reg/smDCs immunoregulatory system operates more effectively when cells are isolated from patients on prednisolone/azathioprine treatment. In this context, we have previously shown that successful immunosuppression reduces T cell effector activity and favors an anti-inflammatory over a proinflammatory milieu.2 Impairment of the T-reg/smDCs immunoregulatory system may derive from a defect in either T-reg or smDC activity. In this study, when cells were obtained before starting immunosuppressive treatment, T-reg suppressor activity was observed only following their coculture with smDCs. The extent of this suppression, however, was lower than that observed when T-regs and smDCs were obtained from patients during treatment, implicating a functional impairment involving not only T-regs but also smDCs at the time of disease presentation. That immunosuppressive treatment has an effect on smDCs was highlighted in a recent study on patients with myasthenia gravis, showing that prednisolone leads to a down-regulation of costimulatory molecules on monocyte-derived DCs, resulting in augmented T-reg suppressor function and prevention of DCs maturation.24

The results of cytotoxicity inhibition experiments show that another property of the T-reg/smDCs immunoregulatory system we describe is its ability to suppress in a bystander fashion, targeting effectors reacting against nonhomologous CYP2D6 regions (Fig. 5). T-reg bystander suppression has been reported also in human type 1 diabetes14 and is regarded as an important feature of regulatory T cells, especially in view of their potential immunotherapeutic application, allowing inhibition of immune reactions to different epitopes on the same autoantigens, a phenomenon frequently described in autoimmune disease.25, 26

When T-regs were cultured with smDCs, a significantly lower number of IFN-γ–producing cells was observed, implicating a role for IFN-γ in offsetting T-reg inhibitory activity. A marked increase in suppressor activity was in fact observed after treatment of T-regs with anti–IFN-γ antibodies, which was not further increased after addition of smDCs, suggesting that smDCs shape the regulatory potential of T-regs by switching off IFN-γ production.

The high regulatory potential of CYP2D6 T-regs in the presence of smDCs renders this cell combination particularly attractive for immunotherapy in our disease model (AIH-2), in which the CYP2D6 autoantigen and its immunodominant regions have been identified.1-4 This approach may also prove useful in other autoimmune diseases, in which a defect in immunoregulation has been demonstrated and the target autoantigen is known, such as primary biliary cirrhosis27 and type 1 diabetes,28 which are conditions unresponsive to conventional immunosuppression. This strategy may also extend to allotransplantation, as suggested by an experimental islet transplant model in which graft tolerance depends on alloantigen-specific Tr1 cells.29

Although our data showing that enhanced smDC-CYP2D6 T-reg suppressor function is associated with a regulatory cytokine profile and with high level of FOXP3 suggest that these cells would be suitable for therapeutic intervention, their phenotypic and functional stability over time needs to be addressed before clinical application. Autoantigen-specific T-regs offer a clear advantage over polyclonal T-regs for the treatment/cure of autoimmune disorders, because they achieve similar levels of suppression at much lower numbers, as illustrated by the work from Tarbell et al. who were able to significantly delay diabetes development in NOD mice by injection of 5000 antigen-specific T-regs, whereas 100 times more polyclonal T-regs were required to obtain the same effect.30

The finding that T-regs cocultured with CYP2D6-pulsed smDCs up-regulate CXCR3, a chemokine receptor present on lymphocytes trafficking to the liver,20, 31 suggests a role for the autoantigen in determining the homing of regulatory T cells into the target organ. Whether T-reg organ specificity relates to the expression of homing molecules or to the autoantigen expressed in the target organ is unclear. Recently, Lennon et al. reported that in a murine model of diabetes, migration and accumulation of antigen-specific T lymphocytes in the target organ is dependent on the expression of the autoantigen in the islets.32

Because antigen-specific T-regs control specifically unwanted effector immune responses, they may represent a tailored form of immunosuppression that reconstitutes tolerance to a specific autoantigen, a feature that renders them preferable to pharmacological immunosuppression which indiscriminately thwarts immune responses without providing tolerance restoration.

Ancillary