Published online before print August 7, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
King’s College London, MRC and Asthma UK Centre in Allergic Mechanisms of Asthma, London, United Kingdom
2 Correspondence: Department of Asthma, Allergy, and Respiratory Science, King’s College London, 5th Floor Thomas Guy House, Guy’s Hospital Campus, London SE1 9RT, UK. E-mail: alistair.noble{at}kcl.ac.uk
ABSTRACT
CD4+CD25+ regulatory T cells (Tregs) are well known to suppress immunopathology induced in lymphopenic animals following T cell reconstitution, including acute graft-versus-host disease (GVHD) post-bone marrow transplantation. The regulatory potential of this subset in nonlymphopenic hosts and in chronic, Th2-mediated GVHD is less clear. We have generated alloantigen-specific cells from CD4+CD25+ populations stimulated with MHC-disparate dendritic cells and found them to express a stable Treg forkhead box p3+ phenotype with enhanced suppressive activity mediated by cell contact. When transferred into nonlymphopenic F1 hosts, nonspecific Tregs proliferated as rapidly as CD4+CD25– cells but displayed distinct growth kinetics in vitro. Tregs, expanded in response to alloantigen in vitro, displayed greatly enhanced suppressive activity, which was partially antigen-specific. They were effective inhibitors of chronic GVHD, blocking donor cell engraftment, splenomegaly, autoantibody production, and glomerulonephritis. CD25+ and CD25– cells were equally susceptible to inhibition by immunosuppressive drugs targeting TCR signaling and rapamycin, but Tregs were resistant to inhibition by dexamethasone. The data indicate that alloantigen-driven expansion, rather than homeostatic proliferation, is key to the effectiveness of CD4+CD25+ Tregs in GVHD and suggest that cellular therapy with alloantigen-induced Tregs in combination with glucocorticoid treatment would be effective in prevention of chronic GVHD after immune reconstitution.
Key Words: alloreactivity • suppression • dexamethasone
INTRODUCTION
Graft-versus-host disease (GVHD) is a frequent and potentially fatal complication of stem cell transplantation procedures and results from donor T cell reactivity against host tissue alloantigens. The pathophysiology of GVHD is incompletely understood, and treatments for GVHD often fail. Studies of GVHD in murine parent-into-F1 models showed that two distinct phenotypes of GVHD exist: acute and chronic disease. Acute disease involves depletion of host hemopoietic cells, hypo-
-globulinemia [1
, 2
], and immunosuppression with progressive damage to target organs such as skin, liver, and the gastrointestinal tract. Chronic GVHD involves production of a variety of autoantibodies (e.g., anti-DNA, erythrocyte, and thymocyte antibodies), which result in immune complex glomerulonephritis [3
4
5
]. Both types of disease show an absolute requirement for the presence of mature T cells in the donor graft. It is known that acute GVHD is caused by cytotoxic CD8 T cell responses, and chronic GVHD is associated with Th2 CD4 T cells, which stimulate host B lymphocytes to secrete autoantibodies [6
]. Although T cell depletion of donor bone marrow cells prevents GVHD, the presence of mature CD4 and especially CD8 T cells in donor cell populations prevents graft rejection and is responsible for the graft-versus-leukemia effect [7
, 8
]. Investigators have therefore sought other techniques for preventing GVHD, including delayed infusion of donor leukocytes [9
, 10
]. Knowledge of the immunoregulatory mechanisms, which might suppress GVHD and preserve graft survival and anti-tumor activity, is essential for such strategies.
Functionally specialized regulatory T cells (Tregs), characterized by a CD4+CD25+ phenotype and expression of the forkhead box p3 (Foxp3) transcription factor, exist as part of the normal repertoire of thymic and peripheral T cells and act to prevent development of pathogenic responses to self and intestinal antigens [11 , 12 ]. The mechanism of suppression is not entirely clear but appears to require cell contact and APCs and may involve down-regulation of costimulatory molecules on dendritic cells (DCs) [13 ] and events triggered by CTLA-4 ligation on Tregs [14 ]. Their activity results in decreased IL-2 production, anergy [15 ], or increased IL-10 production in CD25– cells [16 ]. CD25+ Tregs are also anergic but can be expanded in the presence of IL-2 [15 ]. The "altered negative selection" model proposes that cells with intermediate affinity escape negative selection but are induced to become CD25+ anergic/suppressive cells or cells that produce suppressive cytokines. It is now also apparent that such cells are capable of mediating transplantation tolerance [17 ].
Most studies of CD25+ Treg-mediated suppression have used polyclonal CD25+ populations, which have limited suppressive capacity in vitro, generally requiring large numbers of Tregs for effective inhibition. Nonspecific expansion of Tregs using anti-CD3 has been described, but resulting populations are not antigen-specific. In vivo, most studies have used lymphopenic mice to demonstrate that transfer of polyclonal CD25+ cells prevents immunopathology, for example, in the classic studies of autoimmune-wasting disease [18 ] and in acute GVHD after bone marrow transplantation [19 ]. CD25+ cell therapy prevents GVHD pathology but not the graft-versus-leukemia effect [20 ]. It is now clear that alloantigen-specific Treg populations can emerge from CD25+ populations via clonal expansion, and these cells suppress acute GVHD in an antigen-specific manner [21 , 22 ], but the potency of such cells in control of chronic, Th2-mediated disease in nonlymphopenic hosts is not known. Chronic GVHD has been inhibited using Tregs induced from conventional, alloreactive T cells treated with IL-2 and TGF-β [23 ]. It has been suggested that Treg suppression might be enhanced by homeostatic proliferation in lymphopenic mice as a result of their enhanced ability to compete for available IL-2 through high CD25 expression [24 ]. In clinical situations, it is unclear how immunosuppressive drug therapy would affect Treg engraftment and function.
Here, we have studied the allo-response of CD25+ Tregs, examined their susceptibility to immunosuppressive drugs, and determined the effectiveness of alloantigen-induced Tregs in chronic GVHD. The results suggest that generation of such cells would be a useful, cellular therapy when used as a donor lymphocyte infusion to prevent GVHD after immune reconstitution.
MATERIALS AND METHODS
Mice
Female BALB/c, C57BL/6 (B6), C3H, and [BALB/cxC57BL/6]F1 (CB6F1) mice (6–10 weeks) were purchased from Harlan (Bicester, UK). Experimental protocols were approved by our institutional committee and performed under UK Home Office regulation (London, UK).
Cell separations and MLRs
CD8 cells were depleted from lymph nodes and spleens (LN/SP) of BALB/c mice using MACSTM (Miltenyi Biotec, Bisley, UK; <0.1% CD8+). CD4+CD25+ cells were then purified by labeling with 0.1 µg/106 cells, anti-CD25-PE (Caltag, Towcester, UK), washing, and labeling with anti-PE microbeads, followed by two rounds of autoMACSTM positive selection. The resulting positive fraction was >90% CD4+CD25+. CD4 cells were then purified from the CD25-depleted fraction and were >95% CD4+CD25–. In some experiments, CD25-depleted populations were obtained by labeling as above followed by stringent autoMACSTM negative selection (<0.1% CD25+). DCs were purified from B6 splenocytes by labeling with anti-CD11c–biotin (0.5 µg/106 cells), followed by anti-biotin microbeads. CD11c+ DCs were selected by two rounds of autoMACSTM selection and were >80% CD11c+ MHC class IIhi CD14–. B6 DCs (1x105/ml) were cultured with BALB/c T cells (1x106/ml) in DMEM + 10% FBS + L-glutamine (2 mM), nonessential amino acids (1 mM each), gentamicin (50 µg/ml), and 2-ME (50 µM). Recombinant murine cytokines (Peprotech, London, UK) were added as indicated. For MLR proliferation assays, BALB/c T cells (106/ml) were stimulated with DCs (5x104/ml) in microtiter plates. For cell titration experiments, responding T cells were serially diluted as far as 3 x 103/ml, maintaining the same DC concentration. Inhibitors of TCR signaling and rapamycin (all from Calbiochem, Nottingham, UK) or dexamethasone (Sigma, Poole, UK) were added. Wells were pulsed with 0.5 µCi 3H-thymidine on the indicated day, and incorporation was measured by scintillation counting after overnight incubation. For generation of allospecific Tregs, CD4+CD25+ or control CD4+CD25– T cells were stimulated with B6 DC + IL-2 (10 ng/ml), IL-4 (10 ng/ml), and IL-7 (10 ng/ml) in six-well plates for 7 days.
CFSE assays
Cells were washed twice in PBS and labeled with CFSE (2.5 µM, Cambridge Bioscience, Cambridge, UK) for 10 min at 37°C, washed, and cultured, with or without unlabeled cells in MLRs as above. For transwell experiments, 0.4 µm cell culture inserts (BD Biosciences, Oxford, UK) were used with targets placed below the membrane. CD25+ cells were then placed above or below the membrane. After 4 days, cells were analyzed by flow cytometry for CFSE dilution. For the in vivo assay, 1 x 107 CFSE+ BALB/c CD4+CD25– or CD4+CD25+ cells were injected i.v. into CB6F1 or BALB/c mice. After 4 days, splenocytes were analyzed.
GVHD induction
GVHD was induced by i.p. injection of 7 x 107 BALB/c LN/SP cells into CB6F1 recipients, with a second injection of cultured CD4+CD25+ or CD4+CD25– cells or PBS i.p. Recipient mice received no preconditioning or treatment. This protocol results in chronic GVHD as previously shown [25
]. Mice showed no signs of ill health at any stage. They were killed after 7 weeks, and splenomegaly was measured by weighing spleens. Kidneys were snap-frozen in liquid nitrogen for assessment of glomerulonephritis by immunohistology. Cryostat sections (5 µM) were collected on polylysine-coated slides, air-dried, and fixed in acetone, dried, and incubated with anti-mouse IgG1-FITC (Serotec, Oxford, UK; 1/100 in PBS/FCS) before washing in PBS/FCS and examination by fluorescence microscopy. Anti-DNA IgG2a autoantibody production was measured in serum by ELISA as described [25
].
Flow cytometric analysis
Cells were stained with antibodies to CD4, CD62L, and CD45RB (Caltag) or CD134 (OX40, Serotec) and analyzed using a FACSCaliburTM (BD Biosciences). To determine engraftment of donor T cells in GVHD, LN/SP cells were stained with anti-H-2KbDb-FITC (VHBio, Newcastle, UK) and anti-CD4-APC (Caltag). For intracellular cytokine staining, cells from MLRs were washed, restimulated with immobilized anti-CD3 + anti-CD28 (BD Biosciences, Bedford, MA, USA; 1 µg/ml) + 3 µM monensin for 5 h, fixed with 4% formaldehyde, permeabilized in 0.1% saponin/0.5% BSA, and stained for CD4, IL-4, and IFN- (PE-anti-IL-4 BVD4-1D11, FITC-anti-IFN- XMG1.2, BD Biosciences). For Foxp3 intranuclear staining, CD4-labeled cells were fixed/permeabilized in 0.1% Triton X-100/0.5% BSA/4% formaldehyde in PBS for 30 min before two washes with 0.1% Triton X-100/0.5% BSA in PBS and staining with 0.1 µg anti-Foxp3-Alexa 647 (FJK-16 s, Insight Biotech, Wembley, UK) for 30 min. Cells were washed twice in permeabilizing buffer before analysis.
Statistical analysis
Differences between experimental groups were analyzed using unpaired t-tests. P values less than 0.05 were considered significant and are indicated by *, P < 0.05; **, P < 0.005; and ***, P < 0.0005.
RESULTS
CD25+ Tregs divide rapidly in vivo in response to alloantigen but suppress donor CD4 cell engraftment
As CD25+ cells are anergic in vitro, we determined whether freshly isolated CD4+CD25 Tregs could proliferate in vivo in a GVHD response and thus, be capable of developing into alloantigen-specific Tregs. BALB/c CD25+ or control CD25– CD4 cells were labeled with CFSE and transferred into CB6F1, semiallogeneic hosts, or syngeneic BALB/c hosts. CD25+ Tregs were able to divide as rapidly as control cells, undergoing as many as seven divisions in 4 days (Fig. 1A
). Furthermore, this proliferation was alloantigen-specific, as CD25+ cells did not divide in BALB/c mice. Thus, CD25+ cells, although highly anergic in vitro (not shown), are not anergic in vivo.
 View larger version (55K): [in a new window] |
Figure 1. Freshly isolated CD4+CD25+ cells proliferate in response to alloantigen in vivo but suppress donor engraftment. (A) CD4+CD25– and CD4+CD25+ populations were purified from untreated BALB/c mice and labeled with CFSE, and 1 x 107 were injected i.v. into CB6F1 or control BALB/c recipients. After 4 days, the splenocytes from recipient mice were analyzed by flow cytometry. Percent divided cells in CFSE+-gated events are shown. Similar data were obtained in four independent experiments. (B) CD25+ cells in the donor inoculum suppress donor T cell engraftment in chronic GVHD. BALB/c donor LN/SP cells were depleted of CD25+ cells or left unfractionated. Equal numbers were injected i.p. into CB6F1 recipients, and donor T cell engraftment was determined in spleen cells after 14 days by H-2KbDb staining. Percent donor CD4 T cell engraftment data from groups of seven mice in two independent experiments are shown.
|
CD4+CD25– and CD4+CD25+ populations exhibit similar alloantigen-specific precursor frequencies but distinct growth kinetics
To confirm that Tregs rather than contaminating non-Tregs within the CD25+ population were responding to alloantigen, we analyzed intranuclear Foxp3 expression and CFSE dilution simultaneously in MLRs (Fig. 2A
). After 4 days, it was clear that Foxp3+ cells had divided in response to alloantigen + IL-2 to a similar extent to the Foxp3– population. This proliferation was driven by alloantigen.
 View larger version (30K): [in a new window] |
Figure 2. (A) CD25+ Treg populations contain similar alloantigen-specific precursor frequencies to CD25– cells but exhibit distinct growth kinetics. (A) Foxp3-expressing Tregs proliferate in MLRs. BALB/c CD4+ cells were labeled with CFSE and stimulated with C57BL/6 DC + IL-2 (5 ng/ml) or IL-2 alone. Cells cultured without IL-2 did not survive (not shown). After 4 days, cells were washed and stained for intranuclear Foxp3 expression. Data in boxes refer to the proportion of Foxp3+ or Foxp3– populations, which had divided. (B) CD25+ Tregs have accelerated growth kinetics. BALB/c CD4+CD25+ and CD4+CD25– populations were cultured at a range of concentrations in the presence of C57BL/6 DC + IL-2 (5 ng/ml). Replicate cultures were pulsed with 3H-thymidine after 2, 3, 4, or 5 days, as indicated, and harvested the following day for measurement of proliferation. Similar data were obtained in three independent experiments. (C) Expansion of CD25+ Tregs in response to alloantigen can be prolonged by additional cytokines IL-4 and IL-7. MLRs were set up as in A, with or without addition of IL-2 (5 ng/ml), IL-4 (10 ng/ml), and IL-7 (10 ng/ml) or combinations of cytokines as indicated. Proliferation was assessed on Day 8.
|
Generation of potent alloantigen-induced Tregs in vitro
Bulk cultures of CD25+ and CD25– BALB/c CD4 cells, stimulated with B6 (H-2b+) DCs as above, were added at increasing doses to MLRs containing freshly isolated BALB/c CD25– cells stimulated with B6 DCs or DCs from a third party strain (C3H, H-2k+). Their suppressive capacity was compared directly with freshly isolated BALB/c CD25+ populations. Suppressive activity of CD25+ populations toward alloantigen-stimulated proliferation was greatly enhanced during the culture period (Fig. 3
). Alloantigen-primed CD25+ cells (designated "allo-Treg") were up to 100-fold more potent than fresh CD25+ Tregs, retaining suppressive function at 1:100 ratios to target cells. This suppression appeared to be alloantigen-specific only at the 1:27 Treg:target ratio; H-2b-primed allo-Tregs were less potent inhibitors of MLRs against H-2k+ DCs at this dose. By contrast, fresh CD25+ populations were entirely nonantigen-specific and suppressed at doses of 1:3 or greater. The data indicate that in vitro priming greatly enriched Treg populations for alloantigen-specific T cells and/or activated them to become more highly suppressive. However, the strong suppression mediated against third-party MLRs suggests that once activated in vitro, Treg effector function is not antigen-specific.
 View larger version (21K): [in a new window] |
Figure 3. Alloantigen-expanded CD25+ cells display highly enriched, suppressive activity in secondary MLRs, which is partially antigen-specific. BALB/c CD4+CD25– or CD4+CD25+ populations were stimulated with B6 DCs + IL-2 for 7 days. These "allospecific" cells were washed and titrated into secondary MLRs consisting of freshly isolated BALB/c CD4+CD25– cells and B6 or third-party C3H DCs. Freshly isolated BALB/c CD25– and CD25+ populations were also titrated into parallel cultures for comparison. Proliferation was assessed after 4 days. Graph shows mean ± SD 3H-thymidine incorporation from triplicate cultures. Similar results were obtained in four independent experiments.
|
was detectable at a low level (5%) on cultured CD25+ and CD25– cells but not on freshly isolated cells (not shown). Together, the data indicate that alloantigen-expanded Tregs retain their characteristic phenotype but also express additional activation markers. Freshly isolated CD25+ Tregs were >90% Foxp3-positive, as assessed by intranuclear staining, and CD25– cells contained a minor population of Foxp3+ cells (6%; Fig. 4A
). After expansion, CD25+ cells retained high Foxp3 expression, and CD25– populations remained Foxp3–.
 View larger version (45K): [in a new window] |
Figure 4. In vitro-generated allo-Tregs retain a characteristic Treg phenotype, fail to produce IL-10, and suppress via a cell contact-dependent mechanism. (A) BALB/c CD4+CD25– (filled histograms) or CD4+CD25+ (open histograms) populations were separated and stained before (left panels) or after (right panels) 7 days of culture with B6 DC + IL-2 for cell surface markers or intranuclear Foxp3 as indicated. Flow cytometry profiles shown are typical of three experiments. (B) Alloantigen-expanded cells from CD25– or CD25+ populations as in A were washed after 7 days and restimulated with immobilized anti-CD3 + anti-CD28 before intracellular cytokine (IL-10, IL-4, and IFN-) analysis. Typical staining from three independent experiments is shown. (C) Alloantigen-expanded CD25+ cells were cultured with CFSE-labeled fresh BALB/c CD4 cells at a 1:1 ratio, and stimulated with B6 DCs. Cells were mixed (contact) or separated on either side of a transwell insert (no contact). CFSE-labeled cells were then analyzed by flow cytometry after 4 days culture. Percent of CFSE+-gated cells, which have divided, is indicated.
|
. In contrast, CD25+ populations failed to produce these cytokines, including IL-10. Furthermore, target cell division in a MLR was suppressed by allo-Treg when cells were in contact but not when targets and suppressors were separated by a transwell membrane (Fig. 4C)
. These data indicated that expanded allo-Tregs were contact-dependent suppressors, at least in vitro.
Allo-Tregs induced in vitro are potent suppressors of chronic GVHD in nonlymphopenic hosts
CB6F1 mice were injected with BALB/c LN/SP cells to induce GVHD, along with expanded CD25+ or CD25– populations from BALB/c versus B6 MLRs, generated as above (Fig. 5
). We and others [25
, 28
] have shown previously that GVHD induced by this strain combination results in Th1- and Th2-associated, immunopathological features early in the disease (2–4 weeks), which progresses to a polarized Th2 phenotype at later time-points (6–8 weeks). We therefore measured IgG2a autoantibodies (Th1-associated) in serum at 4 weeks and IgG1 autoantibody/immune complex deposition and splenomegaly (Th2-associated) at 7 weeks. Th1- and Th2-associated immunopathology was strongly inhibited by the adoptive transfer of in vitro-generated allo-Treg cells but was unaffected by expanded CD4+CD25– cells. The data clearly show that in vivo suppression can be achieved in an intact immune system by adoptive transfer of relatively small numbers of antigen-primed CD25+ Tregs.
 View larger version (74K): [in a new window] |
Figure 5. In vitro-generated allo-Tregs potently inhibit chronic GVHD in nonlymphopenic hosts. CB6F1 mice were injected i.p. with 7 x 107 BALB/c parental cells, with or without 4 x 106 in vitro-generated, allospecific (BALB/c vs. C57BL/6) cells from CD4+CD25– (allo-CD25–) or CD4+CD25+ (allo-CD25+) populations as indicated. (A) Splenomegaly induced by GVHD at 7 weeks post-transfer. Means ± SEM from groups of four mice; *, P < 0.05, compared with CD25– cells. Similar results were obtained in two independent experiments. (B) IgG2a autoantibody levels to single-stranded DNA induced by GVHD, measured in sera at 4 weeks after transfer. Means ± SEM from groups of four mice; *, P < 0.05, compared with the CD25– group. Similar results were obtained in two independent experiments. (C) IgG1 immune complex deposition in kidneys of Week 7 GVHD mice, measured by staining kidney sections with anti-mouse IgG1-FITC. Results are representative of groups of four mice; similar data were obtained in two independent experiments.
|
/β inhibitor), PD98059 (MEK inhibitor), and PP1, which targets lck/fyn protein tyrosine kinases, inhibited alloantigen-driven growth in a dose-dependent manner. There was no difference between CD25– and CD25+ populations in their susceptibility to these drugs, and similar IC50 values were apparent in all cases. We also tested the effects of rapamycin, an inhibitor of p70 S6 kinase used in transplantation, which is reported to favor development of Foxp3+ Tregs [31
]. Our results showed no difference between CD25– and CD25+ populations in susceptibility to rapamycin. However, dexamethasone (glucocorticoid) inhibition curves differed consistently between control and Treg populations, and Treg proliferation showed a biphasic curve. This suggested that the CD25+ population contains "genuine" Tregs, which are glucocorticoid-resistant [32
], and contaminating non-Tregs, which are glucocorticoid-sensitive. To test this, we set up MLRs ± dexamethasone, stained cultured cells for CD4 and nuclear Foxp3 after 4 days, and counted the total recovered cells by flow cytometry (Fig. 6B)
. Yields of Foxp3+ populations from CD25+ cultures were not greatly inhibited, even at high concentrations of dexamethasone. By contrast, development of Foxp3– cells in CD25– cultures was suppressed dramatically by similar concentrations of the glucocorticoid to those that blocked proliferation. Foxp3+ Treg growth is therefore resistant to dexamethasone but not other immunosuppressive drugs.
 View larger version (23K): [in a new window] |
Figure 6. Alloantigen-driven Tregs are susceptible to inhibition by TCR signaling inhibitors and rapamycin but resistant to glucocorticoid. (A) Freshly isolated CD4+CD25– and CD4+CD25+ cells from BALB/c mice were stimulated with C57BL/6 DCs and IL-2 as in Figure 2
. Immunosuppressive drugs cyclosporin A (calcineurin inhibitor), Go6976 [protein kinase C (PKC)/β inhibitor], PD98059 (MEK inhibitor), protein phosphatase 1 (PP1; lck/fyn tyrosine kinase inhibitor), rapamycin, and dexamethasone (glucocorticoid) were titrated into the cultures at concentrations indicated. Proliferation was assessed after 4 days. Mean data from triplicate cultures were expressed as percent maximum 3H-thymidine incorporation for each inhibition curve. (B) Cultures as in A were set up with dexamethasone, and cells were harvested at Day 4, fixed, and stained for intranuclear Foxp3, and total yields of Foxp3+ and Foxp3– CD4 cells from each culture were determined by flow cytometry. Similar results were obtained in three independent experiments.
|
CD25+ Tregs represent a lineage of thymic emigrants, which is entirely distinct from conventional T cells. A large number of studies have shown how this subset is distinct in terms of cell surface phenotype, expression of transcription factor Foxp3, cellular responsiveness, and susceptibility to death, in addition to their specialized, suppressive function [11 ]. It has been less clear whether this subset develops normal antigen-specific populations in the periphery via clonal expansion or whether it retains self-specificity. Studies of Foxp3 regulation have not provided a definitive answer to this, as it appears that some peripherally induced Tregs express Foxp3, and others do not [33 34 35 ]. Here, we have shown how in the presence of IL-2, CD25+ Tregs respond normally to the alloantigen and develop into activated cells, which retain potent, suppressive activity and Foxp3 expression. Related findings have shown that use of DCs as APCs is critical for maintenance of Foxp3 expression [36 ]. Our data suggest that unlike other methods for in vitro generation of antigen-specific Tregs, such as vitamin D3/glucocorticoid treatment [34 ], CD25+-derived Tregs are not IL-10-associated. Both types of Tregs are capable of in vivo suppression of immunopathology. It is likely that CD25+-derived Foxp3+ Tregs act through cell contact but also by skewing responding cells toward an IL-10+ phenotype [16 ]. Studies of transplantation tolerance have suggested that Tregs use IL-10 as a dominant mechanism in vivo [17 ]. Inhibition of acute GVHD by high doses of CD25+ cells was also found to be partially IL-10-dependent [19 ]. However, treatment of acute GVHD mice with IL-10 suppresses engraftment but does not alter disease progression, indicating that other suppressive mechanisms are important [37 ]. Allospecific Tregs have also been generated from human CD25+ populations and shown to be contact, not cytokine-dependent, suppressors [38 ]. A recent study also shows that contact with Tregs induces expression of a set of genes associated with growth inhibition but not including Foxp3 or IL-10 [39 ]. Thus, allospecific CD25+ Tregs may have short-term, direct, suppressive function and long-term, indirect effects, which promote development of tolerance. Our data suggest that in allo-Tregs, contact-dependent suppression predominates in vitro, but whether this is true in vivo, during chronic GVHD in nonlymphopenic hosts, is not known.
CD25+ Tregs are highly dependent on IL-2 for their survival and growth and remain anergic in vitro in the absence of IL-2. Our data suggest that in vivo, this anergic phenotype has little consequence for initial clonal expansion of CD25+ cells in response to antigen, which may rely on the availablity of other cytokines in vivo. CD25+ growth was short-lived in IL-2, unless further -chain cytokines (IL-4 or IL-7) were included. Responsiveness to IL-4 would suggest that Tregs might be sustained by the Th2 populations, which develop in chronic GVHD. Another explanation could be that contaminating non-Tregs within the CD25+ population provides sufficient IL-2 for Treg proliferation in vivo but not in vitro as a result of dilution or lack of costimulatory cytokines. In any case, our results indicate that the anergic phenotype of Foxp3+CD25+ Tregs does not prevent them from participating in responses to foreign antigen in the periphery. However, the dynamics and longevity of their response may differ to allow separate regulation of the effector and suppressive lineages of antigen-specific T cells. This is reflected in our observation that CD25 depletion of donor populations results in greater donor CD4 cell engraftment, suggesting that Tregs do not expand to the same extent as CD25– cells.
The remarkable potency of allo-Treg is likely to be a combination of two factors: enrichment of allospecific clonal populations in culture and differentiation of Tregs into a more activated or recently primed state. Although allo-Tregs were more potent suppressors of MLRs, raised to the MHC molecules with which they were primed, they also suppressed third party MLRs to a considerable extent. Other investigators have shown that antigen-specific Tregs can exert "bystander" suppression if activated recently [40 ]. This is consistent with our proposal that suppression is contact-dependent, as release of cytokines such as IL-10 generally requires priming and restimulation with a specific antigen. We did not determine the degree of antigen specificity of in vivo suppression, which could differ markedly, as longer-term outcomes would be affected by issues of survival and indirect suppressive effects. The engraftment of Tregs in F1 hosts was dependent on alloantigen (Fig. 1A) , so it is unlikely that nonspecific Tregs would be effective in our model or in immune-sufficient hosts generally. In irradiated hosts, expanded Tregs display a degree of antigen specificity [36 ]. The allo-Tregs we generated retained a characteristic cell-surface phenotype in vitro. CD134 (OX40), a molecule associated with Tregs [41 ], was expressed at a higher level on allo-Treg post-expansion, perhaps reflecting a greater dependence on costimulatory factors for Treg persistence [42 ]. Allo-Treg also retained high Foxp3 expression, and CD25– cells contained fewer Foxp3+ cells after culture, again suggesting less-sustained growth or survival in Tregs. There was no indication of increasing Foxp3 expression as cells divided (Fig. 2A) . However, it is possible that Foxp3– cells converted to Foxp3+ during proliferation.
Immunosuppressive drug treatment is essential for the prevention and control of GVHD after stem cell transplantation. How such drugs affect the development and activity of Tregs is not well understood. Although IL-2 is critical for CD25+ Treg function, inhibitors of IL-2 signaling do not prevent suppression by human Tregs in vitro [43 ] or tolerance induction in vivo [44 ]. We examined how inhibitors of TCR signaling pathways might affect development of allospecific Tregs, as they may develop in the thymus via a distinct pattern of TCR signaling [26 , 27 ]. Early TCR signaling through linker for activated T cells differentially affects development of CD25+ Tregs [45 ]. Programmed death 1, a suppressive cell surface molecule, possibly used in CD25+ Tregs [46 ], also alters TCR signals [47 ]. We showed previously that polarized types 1 and 2 T cell subsets exhibit distinct patterns of TCR signaling, which render them more or less susceptible to immunosuppressive drugs [30 ]. Our data comparing CD25– and CD25+ populations indicate that no such imbalance in TCR signals exists in endogenous Tregs, as both populations were equally susceptible to inhibitors of calcium/calcineurin, PKC, and MAPK. Likewise, PP1-mediated inhibition of upstream lck/fyn tyrosine kinase activities appeared equivalent in CD25+ Tregs. Cyclosporin is widely used for prevention of GVHD, but it has been suggested that it may enhance autoimmunity [48 ] or interfere with tolerance induction by preventing activation-induced cell death [49 ]. Our data indicate that Tregs are no more susceptible to cyclosporin than effector T cells, consistent with earlier observations [50 ]. Rapamycin has been reported to select for development of CD25+ Tregs [31 ], but our data show no difference between the dose:inhibition curves of CD25– and CD25+ populations. However, the effect of rapamycin was seen after long-term culture of Tregs, and our data show that rapamycin inhibits T cell proliferation only partially, even at high doses. The effects of rapamycin may be mediated through induction of suppressive function in CD25– cells, as indicated by recent studies [51 , 52 ].
Glucocorticoids are also believed to enhance development of peripheral Tregs and are used for GVHD treatment. We found that proliferation of CD25+ but not CD25– populations showed a biphasic dose:inhibition curve with dexamethasone, suggesting that glucocorticoid could select for Tregs by preferentially blocking CD25– cell proliferation. Our Foxp3 analysis confirmed this was the case, as development of Foxp3+ cells was resistant to inhibition by dexamethasone. Indeed, Chen et al. [32 ] showed that CD25+ Tregs express high levels of glucocorticoid receptor and are resistant to dexamethasone-induced apoptosis. Preferential expansion of Tregs might therefore be achieved clinically after immune reconstitution by glucocorticoid treatment or with the use of adjuvants including TLR2 agonists, which stimulate Treg proliferation [53 , 54 ]. However, the latter strategy abrogates the suppressive function of Tregs during the proliferative phase. In addition to GVHD, development of stem cell therapies may benefit from allospecific immunosuppression to prevent rejection. Translation of Treg research into clinical practice should reduce the need for long-term therapy with nonselective, immunosuppressive drugs.
ACKNOWLEDGEMENTS
This work was funded by the Biotechnology and Biological Sciences Research Council, UK.
FOOTNOTES
1 Current address: Department of Nephrology and Transplantation, King’s College London, London, UK. 
Received April 16, 2007; revised June 22, 2007; accepted July 3, 2007.
REFERENCES
signalosome and downstream signaling to PKC
FEBS Lett. 574,37-41[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  M. M. Imanguli, W. D. Swaim, S. C. League, R. E. Gress, S. Z. Pavletic, and F. T. Hakim Increased T-bet+ cytotoxic effectors and type I interferon-mediated processes in chronic graft-versus-host disease of the oral mucosa Blood, April 9, 2009; 113(15): 3620 - 3630. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Zhao, C. Zhang, T. Yi, C.-L. Lin, I. Todorov, F. Kandeel, S. Forman, and D. Zeng In vivo-activated CD103+CD4+ regulatory T cells ameliorate ongoing chronic graft-versus-host disease Blood, September 1, 2008; 112(5): 2129 - 2138. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
