IL-10–differentiated dendritic cells (DC10) induce allergen tolerance in asthmatic mice, during which their lung Th2 effector T cells (Teffs) are displaced by activated CD4+CD25hiFoxp3+ T cells. Intestinal DCs promote oral tolerance by inducing Ag-naive T cells to differentiate into CD4+CD25+Foxp3+ regulatory T cells (Tregs), but whether DCs can induce Teffs to differentiate into Tregs remains uncertain. In this study, we addressed this question in OVA-asthmatic mice that were treated with DC10. OVA-presenting DC10 treatment maximally activated lung Tregs in these animals at 3 wk posttreatment, as determined by upregulation of activation markers (ICOS, programmed cell death-1, glucocorticoid-induced TNFR-related protein, LAG3, and CTLA-4) and in functional assays. This in vitro regulatory activity was ≥90% reduced by treatment with anti–IL-10 but not anti–TGF-β Abs. In parallel cultures, OVA- but not house dust mite (HDM)-presenting DC10 induced ≈43% of CFSE-labeled CD25−/loFoxp3− Teffs from asthmatic OVA–TCR transgenic mice to differentiate into tolerogenic CD25hiFoxp3+ Tregs. We recapitulated this in vivo using OVA-asthmatic mice that were coinjected with OVA- or HDM-presenting DC10 (i.p.) and CFSE-labeled CD4+CD25-/loFoxp3− Teffs (i.v.) from the lungs of asthmatic DO11.10 mice. From ≈7 to 21% of the activated (i.e., dividing) DO11.10 Teffs that were recovered from the lungs, lung-draining lymph nodes, or spleens of the OVA–DC10 recipients had differentiated into CD4+CD25hiFoxp3+ Tregs, whereas no CFSE-positive Tregs were recovered from the HDM–DC10-treated animals. These data indicate that DC10 treatments induce tolerance at least in part by inducing Teffs to differentiate into CD4+CD25hiFoxp3+ Tregs.
Multiple laboratories have reported on the tolerogenic activities of IL-10–differentiated dendritic cells (DC10) in mouse models or ex vivo with human T cells (1–8). Thus, DC10 can protect against the development of OVA-induced asthma (4) or reverse the asthmatic phenotype in OVA (9) or house dust mite (HDM) allergen-sensitized mice, reducing airway hyperresponsiveness (AHR) to methacholine, eosinophilia, and Th2 responses to allergen challenge and circulating levels of allergen-specific IgE and IgG1. DC10-mediated asthma tolerance is allergen-specific (4, 8) and IL-10–dependent (4). In our hands, tolerance is first discernible at 2 wk following i.p. delivery of allergen-presenting DC10, and by 3 wk, the asthmatic animal’s AHR disappears entirely. The Th2 reactivity of pulmonary T cells wanes progressively from 2 wk forward, such that by 8 mo, their responsiveness to recall allergen challenge in vivo is near background (10). Four DC10 treatments at 2-wk intervals bring the asthma phenotype to near background within 8 wk (10). Studies employing DCs that have been transfected with an IL-10–expressing lentivirus, and which thus express exceptionally high levels of IL-10, indicated that endogenous IL-10 expression (e.g., by T cells) is also critical to asthma tolerance in that model (11). In inducing a robust asthma tolerance, treatments with these virally transfected DCs increase the numbers of CD4+CD25+Foxp3+ cells present in the lung-draining (mediastinal) lymph nodes (11). This occurs also in the lungs and mediastinal lymph nodes of DC10-treated HDM-asthmatic mice, and adoptive transfer of pulmonary CD4+ T cells from these mice into asthmatic recipients induces full HDM tolerance in the recipients (M. Lu, W. Dawicki, X. Zhang, H. Huang, and J.R. Gordon, submitted for publication).
Intestinal DCs that present innocuous environmental (e.g., commensal bacterial) Ags to naive T cells in the mesenteric lymph nodes induce their differentiation to CD4+CD25+Foxp3+ regulatory T cells (Tregs), and evidence indicates that expression of TGF-β is central to this process (12). Naive T cells can be readily converted to a Treg phenotype by culture with either CTLA-4–Ig (13) or TGF-β (14–16), but there has been no hard data to date regarding the conversion of Ag-experienced (i.e., effector) T cells (Teffs). It has been reported that tuberculin purified protein derivative (PPD)-specific IFN-γ–producing CD4+ cell lines from PPD-sensitive donors become anergized poststimulation with immobilized anti-CD3 and begin to express high levels of Foxp3 mRNA (17), but Foxp3 expression by itself does not confer activated regulatory cell status on CD4+CD25+ T cells (18) (M. Lu et al., submitted for publication). Coculture of specific allergen-presenting DC10 from atopic asthmatic donors with autologous peripheral blood Th2 phenotype cells also induces allergen-tolerance associated with the outgrowth of activated CD4+CD25+Foxp3+ Tregs (8). Nevertheless, whereas DC10 induce asthma tolerance through activation of Tregs, it is not known whether activated Teffs differentiate into Tregs in treated subjects. In this study, we examined the mechanisms by which DC10 induce Tregs in asthmatic mice. Our data indicate that DC10 do activate Tregs in the lungs of asthmatic mice and that these cells do differentiate from CD4+CD25−/loFoxp3− Teffs.
Materials and Methods
Reagents and mice
8, 19). Female BALB/c and C57BL/6 mice (6–8 wk old) were purchased from Charles River Laboratories (Saint Constant, Quebec, Canada). DO11.10 OVA-specific TCR-transgenic mice were obtained from The Jackson Laboratory (Bar Harbor, ME). C57BL/6 mice that expressed GFP under the control of the Foxp3 promoter were kindly provided through Dr. S. Rudensky (University of Washington, Seattle, WA) and bred in our institutional animal care unit. All mice were treated in accord with the guidelines of the Canadian Council for Animal Care.+CD25+ regulatory T cell isolation kits were purchased from Miltenyi Biotec (Auburn, CA). The lipid dye DiI was purchased from Molecular Probes (Carlsbad, CA). The sources of all other reagents have been reported previously (
Generation of tolerogenic DCs
Bone marrow DCs were generated largely as reported previously (20, 21), and differentiated to tolerogenic or immunostimulatory phenotypes by further culture with 50 ng/ml recombinant murine IL-10 (DC10) or 1 μg/ml Escherichia coli serotype 0127:B8 LPS (DC–LPS), respectively. The cells were pulsed with 50 μg/ml OVA or irrelevant control HDM allergen for 2 h, then washed extensively before use. In our hands, DC10 express low levels of costimulatory molecules (e.g., CD40, CD80) and MHC class II compared with mature DCs and secrete significantly higher levels of IL-10, but little TGF-β, relative to immature DCs (8, 10).
Establishment of the asthma mouse model and DC10 treatments
OVA-specific asthma was induced in BALB/c, C57BL/6, or OVA–TCR-transgenic DO11.10 or GFP–Foxp3-transgenic mice by sensitization with OVA–alum and airway exposure to nebulized aerosols of 1% OVA, as noted in detail previously (19, 22). The asthmatic mice were treated i.p. with 106 allergen-presenting DC10 2 wk after their last airway exposure to allergen, also as noted previously (10).
Assessments of Treg activities
We used in vitro and in vivo approaches to assess the regulatory activities of CD4+CD25hi T cells recovered from the enzymatically dispersed lung tissues (22) of asthmatic mice that had been treated with OVA-presenting DC10 1, 2, 3, or 4 wk earlier. The putative Tregs were purified by positive-selection magnetic sorting from these or untreated normal mice and assessed in functional assays for regulatory activities.
Flow cytometry for expression of activation markers.
We purified T cells from the lungs of asthmatic GFP–Foxp3-transgenic mice that had been treated 3 wk earlier with 1 × 106 OVA- or HDM-presenting DC10 (i.p.). The cells were stained with PE-Cy5-CD4 and PE-labeled isotype control or specific Ab against CD25, ICOS, PD-1, GITR, LAG3, or CTLA-4 (intracellular staining), then the CD4+ cells were gated and GFP+ (i.e., Foxp3+) cells were analyzed by FACS.
In vitro assessments.
CD4+ Th2 Teffs for the assays were magnetically sorted from the lungs of untreated asthmatic donor mice. In preliminary experiments done in 96-well plates, we titrated the numbers of Th2 cells and irradiated (3000 rad) OVA-pulsed DC–LPS required to optimally stimulate Th2 cell proliferation and Th2 cytokine secretion and the impact of similarly irradiated putative Tregs on this response. Thus, for the experiments reported in this study, we cocultured Th2 cells (105 cells/well) with half-maximal numbers of OVA-presenting DC–LPS (3.7 × 103 cells/well) and putative Tregs (105 cells/well). In some assays, we added neutralizing anti–IL-10 or anti–TGF-β Ab (each 10 μg/ml) to the cultures. After 2 d, the CD4+ Th2 cells’ proliferative and cytokine (IL-4, -5, -9, and -13 and IFN-γ) responses were assessed using standard [3H]thymidine uptake and ELISA assays, respectively.
In vivo assessments.
We injected 106, 5 × 105, 2.5 × 105, or 1.25 × 105 Treg i.v. into untreated asthmatic recipient mice and 4 wk later assessed the AHR of the mice. The following day, we challenged them for 20 min with aerosols of 1% nebulized OVA, and 2 d later, we euthanized them to assess their pulmonary immunoinflammatory responses as noted in detail (22). We did differential cell counts on their bronchoalveolar lavage (BAL) leukocytes and assessed the BAL fluid levels of IL-4, -5, -9, and -13.
Measurement of AHR
AHR was assessed in conscious animals by head-out, whole-body plethysmography, as noted in detail (19, 22). Briefly, air was supplied to the body compartment of a plethysmograph via a small animal ventilator, and changes in the airflow through the body compartment were monitored. Doubling doses of nebulized methacholine (0.75–25 mg/ml) were delivered to the head compartment, and bronchoconstriction data were gathered as running 1 s means of the airflow at the 50% point in the expiratory cycle ([email protected]%TVe1). This parameter accurately reflects bronchiolar constriction, as opposed to alveolar constriction or airway occlusion (23, 24).
ELISA for airway Th2 cytokines
Our capture cytokine ELISA protocols have been reported in detail previously (19, 22). BAL fluids were not diluted for the assays. Cytokine levels are presented as picograms per milliliter based on recombinant protein standard curves; all assays were sensitive to 5–10 pg/ml.
Assays for conversion of CD4+CD25−/loFoxp3− Teffs to CD4+CD25hi T cells
We assessed whether Tregs could differentiate from CD4+CD25−/loFoxp3− Teffs both in vitro and in vivo using CD25−/loFoxp3− cells purified by negative selection from the pulmonary CD4+ cells of asthmatic OVA-specific TCR-transgenic DO11.10 mice. These cells were stained with CSFE (2 μM; 20 min at 37°C).
In vitro assay.
We cocultured Teffs (105/well) with OVA- or HDM-presenting (i.e., nonspecific allergen) DC10 or OVA-presenting immunostimulatory DC–LPS (3 × 104/well) in culture medium that was supplemented with IL-2 (10 U/ml) as a growth factor. After 5 d, cells were fixed and permeabilized, then stained with PE-Cy5–labeled anti-Foxp3 Ab. For the FACS analysis, we gated on the CFSE-positive cells and assessed both proliferation and Foxp3 expression. We also assessed the in vitro regulatory activities of the cells arising in these cultures as noted above.
In vivo assay.
We injected 5 × 106 CFSE-stained Teffs from asthmatic DO11.10 donors i.v. into asthmatic BALB/c recipient mice and at the same time treated the recipients with 1 × 106 OVA- or HDM-presenting DC10 i.p. Two weeks later, single-cell suspensions generated from the lungs, mediastinal lymph nodes, and spleens of the recipients were stained with PE-Cy5 Foxp3 and PE-CD25 mAb and analyzed by FACS, wherein we gated on the activated (i.e., dividing) CSFE+ cells.
All data are expressed as the mean ± SEM. Multigroup comparisons were assessed by ANOVA with post hoc Fisher’s least significant difference testing, whereas AHR to methacholine was assessed by linear regression analyses. Significance was established at p ≤ 0.05.
Pulmonary CD4+CD25hiFoxp3+ cells become activated following tolerogenic DC treatment of asthmatic mice
Treatment of asthmatic mice with DC10 broadly ameliorates airway Th2 and eosinophil responses to allergen challenge within 4 wk of treatment (10). In both OVA and HDM models, tolerogenic DC treatments augment the regulatory activities of CD4+ T cells at this time (M. Lu et al., submitted for publication) (11). However, when we assessed the numbers of CD4+CD25hiFoxp3+ cells in the lungs or lung draining (i.e., mediastinal) lymph nodes of saline- or DC10-treated OVA-asthmatic mice at 4 wk following DC10 delivery, we found no differences in the proportions of these cells in the two groups (Fig. 1A). Thus, CD4+CD25hiFoxp3+ cells comprised a mean of 7.07 ± 0.88% and 8.24 ± 1.41% of the pulmonary CD4+ T cells in the saline- and DC10-treated animals, respectively, and they comprised 2.62 and 2.54% of the mediastinal lymph node CD4+ T cells in these respective groups. Others have similarly reported equivalent proportions of CD4+CD25hiFoxp3+ cells in tolerant and nontolerant animals in other model systems (18).
We used a Treg magnetic-sorting kit to purify these cells and Teffs from the lungs of asthmatic mice that had been treated 4 wk earlier with OVA-presenting DC10. The Tregs expressed high levels of both CD25 and Foxp3, whereas the magnetic column Teff flow-through population expressed low levels of CD25 and no discernible Foxp3. The median fluorescence intensity (MFI) of CD25 on the Teffs was 5.41 (isotype control MFI, 2.51), whereas the MFI of CD25 on the Tregs was 15.8 (isotype control MFI, 3.34; Fig. 1B). We next assessed the respective abilities of the purified Tregs and Teffs from DC10-treated asthmatic mice to suppress the activation of pulmonary Teffs from untreated asthmatic donors by OVA-presenting immunostimulatory DCs (DC–LPS). The CD4+CD25hi T cells from the lungs of the DC10-treated mice reduced proliferation of these Teffs by 53.8 ± 11.8% (Fig. 1C) and reduced their expression of Th2 cytokines by 35–40%, whereas the Teffs from DC10-treated mice had no discernible impact in these assays (Fig. 1D). To determine whether the regulatory activity of these cells was dependent on their expression of IL-10 or TGF-β, we also tested the impact on this Th2-proliferative response of neutralizing Abs. Anti–TGF-β Ab had no effect on T cell proliferation, but the anti–IL-10 Abs eliminated the tolerogenic activities of the CD4+CD25hiFoxp3+ cells from the DC10-treated asthmatic mice (Fig. 1E). The anti–IL-10 Abs had no impact on the Th2 cytokine responses of Teffs from asthmatic mice (Fig. 1E, a-IL-10), suggesting that these cells were not the source of the IL-10 in the Treg/Teff cocultures.
We next examined the kinetics of pulmonary Treg induction after DC10 treatment, purifying CD4+CD25hiFoxp3+ cells from the lungs of normal mice or asthmatic mice that had been treated with DC10 1–4 wk earlier. We had previously found that DiI-stained DC10 traffic from the peritoneal cavities of recipient mice to the airways and lung-draining (mediastinal) lymph nodes (M. Lu, H. Huang, and J.R. Gordon, unpublished observations). Maximal accumulation of these cells in the mediastinal nodes occurs at 3 wk postimplantation, when up to 15% of the cells recovered from the nodes were signal-positive (H. Huang and J.R. Gordon, unpublished observation). In the current study, we found that CD4+CD25hi T cells from the lungs of normal mice had modest regulatory activity in our in vitro assay, whereas pulmonary CD4+CD25hi T cells from the DC10-treated asthma phenotype mice were significantly more active (Fig. 1F). At 1 wk after DC10 delivery, there was a ≈2-fold increase in regulatory activity relative to the CD4+CD25hi T cells from normal mice, and this activity continued to increase to a maximum (≈72% inhibition) at 3 wk. Interestingly, despite the fact that tolerance in asthmatic mice is progressive well beyond 4 wk (9), the activity of the purified 4-wk Tregs was diminished relative to that of 3-wk Tregs (Fig. 1F).
To further assess the activation of Tregs in our model, we used asthmatic mice that expressed GFP under the control of the Foxp3 promoter as a means of easily tracking Tregs. Three weeks after treating these mice with OVA- or irrelevant allergen (i.e., HDM)-presenting DC10 or OVA-presenting GM-CSF–differentiated DC (DC–OVA; as a negative control population), we examined the expression of a panel of regulatory cell surface markers (i.e., ICOS, PD-1, GITR, and LAG3) as well as intracellular CTLA-4 on Tregs recovered from their lungs. We found that the OVA-pulsed DC10 treatments led to increased expression of a number of these (Fig. 2). Thus, the proportions of Foxp3+ cells expressing ICOS, PD-1, GITR, LAG3, and CTLA4 were increased ~2–10-fold relative to the analogous cells from HDM-presenting DC10-treated or DC–OVA–treated (i.e., nontolerant) asthmatic mice. In keeping with our observation of no differences in the numbers of CD4+CD25hi cells in normal, asthmatic, or tolerant mice (Fig. 1A), we did not observe any differences in CD25 expression in these mice.
Passive transfer of tolerance with DC10-induced pulmonary CD4+CD25hiFoxp3+ cells
We next sought confirmation that the CD4+CD25hiFoxp3+ cells from the lungs of DC10-treated mice could operate as regulatory cells in vivo. We titrated the numbers of Tregs required to transfer tolerance, giving asthmatic mice either saline or 1.25–10 × 105 CD4+CD25hi cells (i.v.) from mice treated 3 wk earlier with DC10. Given our demonstration that Tregs from asthmatic animals possess only ≈30% of the regulatory activity of DC10-activated Tregs, we did not also titrate these cells or Tregs from the lungs of control DC-treated asthmatic mice. We assessed AHR in the recipients weekly thereafter, then administered a recall allergen challenge (20 min of nebulized 1% OVA aerosol) on day 28 and sacrificed the animals 2 d later. We assessed airway eosinophilic inflammation and Th2 cytokine responses, as determined by BAL fluid IL-4, -5, -9, and -13 levels. The AHR of the animals given 106 DC10-induced Tregs was fully normalized by 21 d posttransfer (Supplemental Fig. 1) and remained so at 4 wk (Fig. 3A). When we transferred 5 × 105 3 wk Tregs, AHR was not discernibly affected at 2 or 3 wk (Supplemental Fig. 1), but by 4 wk, it was significantly reduced (p < 0.05 versus saline-treated asthmatic mice). Lower numbers of Tregs were without effect on AHR, at least as of 4 wk posttransplant (Fig. 3A). Passive transfer of Tregs also reduced airway eosinophilia and Th2 cytokine levels in a dose-dependent manner. Specifically, either 5 × 105 or 106 cells reduced the eosinophil responses to background, whereas 2.5 × 105 or fewer cells were without effect (Fig. 3B; p > 0.05 versus saline-treated asthmatic mice). In accord with this, Th2 responses were also amenable to Treg tolerance, such that delivery of 106 cells reduced the Th2 response to near background and 5 × 105 Tregs reduced them very markedly (p < 0.05, versus saline-treated asthmatic animals), whereas transfer of fewer cells was without discernible effect (Fig. 3C).
For a more finely tuned in vivo confirmation of the kinetics for DC10-driven pulmonary Treg induction, we transferred a limiting number (5 × 105) of pulmonary Tregs from mice treated 1–4 wk earlier with DC10 and again assessed their impact on AHR and on Th2 cytokine and eosinophilic inflammatory responses to airway recall allergen challenge (Supplemental Fig. 2). When used at these limiting numbers, only the cells from animals treated 3 wk earlier with DC10 were sufficiently activated to alter AHR or eosinophilia (for both, p < 0.05 versus saline-treated asthmatic recipients) in the asthmatic recipients. The airway Th2 responses were slightly more amenable to regulation, such that 5 × 105 Tregs from mice treated with DC10 either 3 or 4 wk earlier significantly affected airway cytokine levels (p < 0.05 for each cytokine versus saline-treated asthmatic recipients). The 1- and 2-wk cells were ineffective in reducing airway Th2 cytokine expression (p > 0.05 for each cytokine).
DC10 induce the differentiation of CD4+CD25−/loFoxp3− Teffs into CD4+CD25hiFoxp3+ Tregs
It has been reported that naive T cells are readily converted to a regulatory phenotype when costimulated with CTLA-4–Ig (13) or TGF-β (14–16). Our DC10 do not express significant levels of TGF-β relative to either immature or TNF-treated in vitro-differentiated DCs (10), but our data indicate that they are clearly tolerogenic in asthmatic animals, wherein they appear to fully reverse Th2 T cell responses. Thus, we questioned whether they induce CD25−/loFoxp3− Teffs to differentiate into Tregs. It has been shown that Teff cell lines generated from tuberculin PPD-challenge skin sites can be anergized ex vivo and that Foxp3 expression is upregulated in concert with this induction of anergy (17), but, as far as we are aware, there have been no reports demonstrating that Tregs can differentiate from Ag-experienced Teffs. To test this directly, we set up an in vitro culture system in which CD4+CD25−/loFoxp3− Teffs purified from the lungs of asthmatic OVA–TCR-transgenic DO11.10 mice were stained with CSFE and cocultured with specific (OVA) or irrelevant allergen (HDM)-presenting DC10 or OVA-presenting immunostimulatory cells (DC–LPS). After 5 d, we analyzed the CFSE+ cells from these cultures to assess their proliferation (CFSE dilution) and expression of Foxp3 (Fig. 4A), but we also used magnetic sorting to purify the CD25hi cells that were induced in these cultures and titrated their regulatory activities (Fig. 4B). In the cultures containing irrelevant allergen-presenting DC10, we observed little if any proliferation of the Teffs and no expression of Foxp3 by the CFSE-labeled cells, and this makes sense based on the known allergen specificity of DC10-induced tolerance (4, 8). The DC–LPS strongly induced Teff cell proliferation but not Foxp3 expression, but the bulk of the CD4+CD25−/loFoxp3− cells in the OVA-presenting DC10 cocultures had proliferated, and ≈45% of them expressed Foxp3 at high levels (Fig. 4A). When we magnetically sorted the induced CD25hi cells back out of these cultures and titrated their activity in vitro, we found them to be highly effective in dampening DC–LPS-induced Teff cell proliferation (Fig. 4B; p ≤ 0.05 versus CD25hi T cells sorted from DC–LPS/Teff cell cultures). These data confirmed that DC10 are fully capable of efficiently inducing CD4+CD25−/loFoxp3− Teffs from asthmatic mice to differentiate into functional regulatory cells in vitro and that this response is fully dependent on cognate allergen presentation.
To confirm the in vivo relevance of this observation, we also assessed whether delivery of DC10 to asthmatic mice would similarly induce Teffs to differentiate into CD25hiFoxp3+ Tregs. We magnetically purified CD4+CD25−/loFoxp3− Teffs from OVA-asthmatic DO11.10 mice, labeled them with CFSE, and injected them i.v. into asthmatic recipients. At the same time, we injected the recipients i.p. with either OVA- or HDM-presenting wild-type DC10. Two weeks later, we generated single-cell suspensions from the lungs, mediastinal lymph nodes, and spleens of the treated mice, stained the cells with anti-Foxp3 and anti-CD25 Abs, and analyzed the cells by FACS, gating on the dividing (i.e., activated) CFSE+ cells (Fig. 5). We found negligible numbers of CFSE+CD25hiFoxp3+ T cells in the lungs (0.21%), lung-draining lymph nodes (0.64%), or spleens (0.43%) of the animals we had treated with HDM-presenting DC10, but there were significant numbers of activated CFSE+CD25hiFoxp3+ cells in the asthmatic mice that had been treated with specific allergen-presenting DC10. We found that 7, 16, and 21% of the proliferating CFSE+ cells recovered from the lungs, mediastinal lymph nodes, and spleens, respectively, were now CD25hiFoxp3+ (Fig. 5). Taken together, our in vitro and in vivo data indicated that during DC10-mediated induction of allergen-tolerance in asthmatic mice, CD4+CD25−/loFoxp3− Teffs do differentiate into CD4+CD25hiFoxp3+ Tregs.
Multiple laboratories have reported that CD4+ cells take on regulatory activities with tolerance induction in asthma (M. Lu et al., submitted for publication) (8, 11, 18), but the mechanisms by which this occurs had not been defined. We documented in this study that in inducing tolerance in asthmatic mice, DC10 also induce CD4+CD25−/loFoxp3− Th2 Teffs to differentiate into CD4+CD25hiFoxp3+ Tregs. The kinetics with which these cells were activated in the lung correlated very well with the acquisition of asthma tolerance in mice models (M. Lu et al., submitted for publication) (10). Nevertheless, whereas peak Treg activation occurred at 3 wk after the DC10 treatment, we have found that tolerance induced by a single DC10 treatment is progressive over many months in our mouse model (10). This suggests that alternate mechanisms supplant the pulmonary Treg-dependent tolerance that sets in over the first few weeks following DC10 treatment, perhaps in the context of infectious processes that incorporate regulatory DCs and/or alternate Treg populations (25). For example, Tregs can induce myeloid DCs to adopt a regulatory phenotype (26, 27), whereas DCs can reciprocally express substantial control over Treg populations (28).
It had been shown previously that gut lamina propria or mesenteric lymph node, but not splenic, DCs can induce naive T cells to differentiate to a regulatory phenotype without need for exogenous input (i.e., TGF-β) (29), and this would be appropriate in a compartment routinely presented with commensal (i.e., nonpathogenic) bacteria (30). In the lungs, which are also under constant exposure to innocuous foreign Ags (e.g., pollens), pulmonary DCs that present such Ags to naive T cells express a semimature phenotype and IL-10 and thereby induce Treg responses that prevent development of pathogenic (e.g., allergic) responses (31). The fact that we generated Teff pools from the lungs of fully symptomatic asthmatic mice indicates that these cells would have been educated Teffs as opposed to naive T cells. Thus, our data showing that DC10 induced the differentiation of these Teffs into CD4+CD25hiFoxp3+ Tregs provides our first clear documentation that Teffs are amenable to such a phenotypic change. We have shown previously that specific allergen-presenting DC10 generated from CD14+ monocytes of asthmatic individuals similarly induce the outgrowth of CD25+Foxp3+LAG3+ CTLA-4+ Tregs from autologous Th2 Teff populations. Those Tregs subsequently suppressed autologous Teff responses in a contact-dependent manner (8), as reportedly occurs with Tregs in other model systems (32, 33). However, in the mouse model we employed in this study, our DC10-induced regulatory activity was dependent on expression of IL-10 by the T cells, a characteristic consistent with a Tr1-like phenotype regulatory cell. Others have reported that IL-10–treated DCs induce anergy among Ag-specific T cells in part via CTLA-4 (34). Nevertheless, IL-10 expression by both the treatment DC10 (4) and endogenous host cells (11) has been implicated in tolerance induction in asthmatic mice.
Interestingly, although ectopic expression of Foxp3 alone is reportedly sufficient to turn Teffs into Tregs (35–38), there were equivalent numbers of CD4+CD25hiFoxp3+ cells in the lungs of asthmatic mice irrespective of whether they were fully asthmatic or their pulmonary CD4+CD25hiFoxp3+ cells were activated and expressed a regulatory phenotype. And, although DC10 induce Th2 Teffs from atopic asthmatic individuals to take on an activated Treg phenotype (i.e., CD4+CD25+Foxp3+LAG3+CTLA-4+ IL-10–secreting T cells), the cells in these cultures do not express increased levels of Foxp3 relative to those in cultures containing immunostimulatory DC-activated Teffs (8). A similar observation has been made with CD4+CD25+Foxp3+ cells from asthmatic rats versus those rendered allergen-tolerant by chronic airway exposure to allergen (18). This provides further evidence that Foxp3 expression by itself is not sufficient for optimal induction of a regulatory phenotype in CD4+CD25hi cells. The fact that the numbers of pulmonary CD4+CD25hiFoxp3+ cells were equivalent in our asthmatic and tolerant animals, despite the observation that sizable numbers of Teffs had ostensibly converted to activated Tregs in the latter group, suggests that Treg homeostatic control mechanisms (39) were operative in their lungs. It is recognized that DCs and Treg populations control one another in a reciprocal homeostatic fashion (28), and this raises the question of whether the tolerogenic DCs we introduced into asthmatic animals may exercise homeostatic control over lung Treg numbers. Perhaps the CD4+CD25hiFoxp3+ cells that were present in the asthmatic lung prior to tolerance induction represent a subpopulation of cells that are uniquely susceptible to apoptotic (40) or other control mechanisms.
Tregs play important roles in maintaining the balance between protective and pathogenic immune responses (18, 41). To date, both naturally occurring thymic CD4+CD25+ Tregs and inducible Tregs have been recognized, with the latter cells including IL-10– and TGF-β–secreting Tr1 and Th3 cells, respectively (42). Naturally occurring Tregs constitute 1–5% of the CD4+ T cells in healthy adult mice and humans, but regulatory cells with similar surface markers and functions can also be induced in the periphery. These cells can be isolated from mice and humans based on their high-level expression of CD25 (the IL-2R α-chain). Other markers that were originally thought to be specific for naturally occurring Tregs include CTLA-4 and GITR but, like CD25, they are also expressed by activated T cells. Our data indicated that DC10 increased the expression of ICOS, PD-1, GITR, LAG3, and CTLA-4 with the acquisition of the regulatory phenotype, and this fits well with observations by others (43).
In summary, our data support the observation that these cells induced CD4+CD25−/loFoxp3− Teffs to differentiate into CD4+CD25hiFoxp3+ Tregs. We did not assess whether they also directly activated or induced proliferation of pre-existing CD4+CD25hiFoxp3+ T cells, but these cells can indeed proliferate strongly and particularly so under the influence of IL-2 (44). We also did not rigorously investigate the cellular interactions between the DC10 and Teffs in asthmatic animals, but we know that these populations do engage one another intimately and in an Ag-specific fashion (M. Lu et al., submitted for publication) (8). These will be important issues to address in the future.
We thank Mark Boyd for assistance in FACS analysis.
Disclosures The authors have no financial conflicts of interest.
This work was supported by grants from the Canadian Institutes of Health Research (to J.R.G.).
The online version of this article contains supplemental material.
Abbreviations used in this paper:
- airway hyperresponsiveness
- bronchoalveolar lavage
- dendritic cell
- IL-10–differentiated dendritic cells
- glucocorticoid-induced TNFR-related protein
- house dust mite
- median fluorescence intensity
- programmed cell death-1
- purified protein derivative
- effector T cell
- regulatory T cell.
- Received October 21, 2009.
- Accepted August 20, 2010.
- Copyright © 2010 by The American Association of Immunologists, Inc.