Increased CD127 expression on activated FOXP3+CD4+ regulatory T cells

Authors


Abstract

Regulatory T cells (Treg) are commonly identified by CD25 (IL-2Rα) surface expression and/or intracellular expression of the FOXP3 transcription factor. In addition, Treg are also characterized by low CD127 (IL-7Rα) expression when compared to conventional T cells and their biology in the periphery is considered essentially independent of IL-7. We further investigated CD127 expression on Treg and we demonstrated differential CD127 expression depending on Treg subsets considered. Notably, we observed high CD127 expression on inducible costimulatory molecule (ICOS)- and CD103-expressing Treg subsets. Since these two markers reflect activation status, we addressed whether Treg activation modulated CD127 expression. We demonstrated that in contrast to conventional T cells, Treg significantly upregulated CD127 expression during in vitro and in vivo activation using adoptive transfer and contact dermatitis models. High CD127 expression on Treg was also predominantly detected ex vivo in some specific sites, notably bone marrow and skin. Importantly, higher CD127 expression on Treg correlated with higher phosphorylation of STAT5 upon IL-7 exposure. High CD127 expression on Treg also provided survival advantage upon in vitro incubation with IL-7. We thus demonstrated that low CD127 expression is not an intrinsic characteristic of Treg and we identified activated Treg as a potential target of endogenous or therapeutic IL-7.

Introduction

CD4+CD25+FOXP3+ Treg are a critical CD4+ T-cell subset involved in the control of immune homeostasis by preventing autoimmune diseases and regulating chronic inflammation 1, 2. Treg have been shown to exert suppressive activity by various mechanisms, notably by deprivation of IL-2 3–5, inhibitory cytokine production (IL-10, TGF-β, IL-35), cytolytic function, metabolic disruption and modulation of APC function 6. They are commonly identified by CD25 (IL-2Rα) surface expression and/or intracellular expression of forkhead box P3 (FOXP3) transcription factor 7–9. In addition to high CD25 expression, Treg are also characterized by low CD127 (IL-7Rα) expression when compared to conventional T cells 10–12. Although CD127 expression and consequently IL-7/IL-7R signaling is crucial to most peripheral T cells, Treg stand a remarkable exception and exhibit different cytokine requirement profile: IL-2 appears as the main survival factor, whereas IL-7 is thought to contribute moderately to Treg homeostasis 13–15. High IL-2 requirement and low IL-7 influence on Treg are in accordance with the high level of CD25 and low level of CD127 expression on Treg. However, CD127 expression has been shown to be highly regulated on conventional T cells depending notably on IL-2 and IL-7 availability 16, 17 or TCR stimulation 18. We thus questioned whether low CD127 expression was an intrinsic hallmark of Treg or could be regulated by their activation status and/or environmental contexts. First, we determined whether Treg exhibited some heterogeneity in CD127 expression ex vivo. When addressing Treg as subsets by introducing markers of Treg heterogeneity 19, 20, high CD127 expression correlated with the expression of markers such as CD103 and inducible costimulatory molecule (ICOS). Because these two markers have been described to identify activated Treg 21–25, these results suggested that in contrast to conventional T-cell activation, CD127 expression may be more prominent during Treg activation. To confirm this hypothesis, we developed different approaches. We detected higher CD127 expression on Treg upon in vitro activation. Using adoptive transfer and contact dermatitis models, we confirmed in vivo that CD127 is highly expressed on activated Treg both in lymphoid and non-lymphoid tissues. Collectively, these data demonstrated high CD127 expression correlated with Treg activation. Finally, since Treg activation also occurs in peripheral tissues, we performed tissue-specific analysis, revealing striking diversity in CD127 expression on Treg depending on the sites considered. To formally identify activated Treg as a preferential target of IL-7, we determined whether Treg transduced intracellular signaling mediated by IL-7. We demonstrated that high levels of CD127 expression on Treg correlated with increased IL-7 signal transduction. We finally confirmed the functional relevance of CD127 expression by demonstrating enhanced survival capacity of Treg expressing high level of CD127 (CD127hi) upon in vitro incubation with IL-7.

Our results demonstrated that low CD127 expression is not an intrinsic characteristic of Treg and that differential CD127 expression on Treg depends on their localization and their activation status. They also revealed differential regulation of CD127 expression among conventional and regulatory CD4+ T cells upon activation, providing potential regulation of Treg biology depending on IL-7 availability.

Results

ICOS and/or CD103 identify CD127-expressing Treg ex vivo

Most CD4+FOXP3+ Treg have been shown to exhibit low expression of CD127 compared to conventional CD4+ and CD8+ peripheral T cells both in humans and mice (11, 12). Regarding large phenotypic heterogeneity among Treg, we questioned whether CD127 expression could differ depending on the Treg subsets considered. To obtain a comprehensive characterization of Treg phenotypic subsets, we determined on FOXP3+CD4+ T cells the expression of several markers discriminating activation status (CD62L, CD44, CD69), or known to be preferentially expressed on Treg (GITR, CTLA-4) or representing co-stimulatory (CD28, ICOS, OX40, CD154, PD-1) and adhesion molecules (ICAM-1, CD103) (data not shown). Among the various molecules studied, we identified four markers for which the Treg population showed heterogeneity: in LN, CD62L was expressed by about half of Treg (56±6%); CD69 expression identified 34±1% of Treg; CD103-expressing Treg represented 21±1% of the total Treg population; and ICOS was expressed by 16±2% of Treg (percentages are mean±SD of a representative experiment as defined on a sample in Fig. 1A). Prior to the analysis of CD127 on each subset, we confirmed the suppressive capacity of these respective subsets. CFSE-labeled sorted naïve CD4+ T cells (CD62L+CD25CD4+ T cells) were stimulated in vitro with anti-CD3 with or without Treg and proliferation was assessed at days 3–4. As shown in Fig. 1B, all Treg subsets studied exerted a suppressive activity at least comparable with that of the whole CD4+CD25+ population in a 1:1 Teff:Treg ratio. We next analyzed CD127 cell surface expression on total Treg and Treg subset expressing or not CD62L, CD69, ICOS and CD103 (Fig. 1C). As previously published, total Treg exhibited low CD127 expression compared to conventional CD4+ T cells (CD127 MFI: 68±12 and 302±23, respectively). To note, level of CD127 expression on Treg was significantly different from isotype control staining, confirming low but existing CD127 expression on Treg. CD69- or CD62L-expressing Treg presented CD127 surface levels not significantly different from their negative counterparts. Importantly, CD103 or ICOS-expressing Treg exhibited significantly higher levels of CD127 compared to their negative counterparts (CD127 MFI: 111±3 on CD103+ compared to 62±14 on CD103; 146±30 on ICOS+ compared to 67±12 on ICOS). To determine whether ICOS- and CD103-expressing Treg subsets were independent or partially overlapping, we analyzed the co-expression of ICOS and CD103 on Treg and CD127 expression depending on ICOS and CD103 co-expression (Fig. 1D). CD103ICOS Treg were predominant (73±1%) and exhibited low CD127 expression. The three subsets identified by CD103 and/or ICOS expression (i.e. CD103 or ICOS single positive, CD103/ICOS double positive) were similarly represented (9±1%, 7±1% and 9±1% of CD4+FOXP3+ cells isolated from LN, respectively) but exhibited very different CD127 profiles. Single positive cells for either ICOS or CD103 expressed CD127 levels comparable with those expressed by the whole Treg subset. Strikingly, levels of CD127 on CD103+ICOS+ cells were significantly higher when compared to total Treg (CD127 MFI: 195±12 on ICOS+CD103+ Treg fraction compared to 68±12 on total Treg). We thus demonstrated that in LN, CD103/ICOS expression identified a fraction of CD4+FOXP3+ T cells, which exhibited high CD127 expression.

Figure 1.

In lymph nodes, CD103 and ICOS expression identifies Treg presenting high levels of CD127. (A) Phenotypic heterogeneity of Treg. Dot plots show the expression of CD62L, CD69, CD103 and ICOS depending on FOXP3 expression on CD4+TCRβ+ LN cells. Percentages of cells from a representative sample are indicated in each quadrant. (B) Suppression assay of total, CD62L, CD69, CD103 or ICOS expressing CD4+CD25+ subpopulations. CFSE profiles of sorted naïve CD4+ T cells cultured 72–96 h in the presence of 0.5 μg/mL anti-CD3 with or without the indicated sorted CD4+CD25+ subsets (filled and open profile, respectively). Data are representative of three independent experiments. (C) CD127 expression on total and CD62L-, CD69-, CD103- or ICOS-expressing (grey histogram) or non-expressing (thick open line) FOXP3+ Treg compared to conventional FOXP3CD4+ T cells (open histogram). Isotype control staining for both Treg subsets are represented (dotted and shaded lines). (D) CD103 and ICOS expression on LN CD4+FOXP3+ Treg. Histograms show CD127 expression on total Treg (thick open line) or indicated Treg subsets identified based on ICOS and/or CD103 expression (gray shaded profile) compared to conventional CD4+ T cells (open histogram). Isotype of total Treg and Treg subsets are presented (dotted and dashed lines, respectively). Gating strategy is shown in Supporting Information Fig. 1A. FACS data are representative of more than five independent experiments, using at least three C57BL/6 mice per experiment.

Treg upregulated CD127 during in vitro and in vivo activation

Since CD103 and ICOS have been described as Treg activation markers, we thus speculated that the CD127hi Treg fraction could be more prominent during Treg activation. To directly address this assumption, we studied CD127 expression on Treg upon in vitro and in vivo activation. First, we determined CD127 expression simultaneously on Treg and conventional CD4+ T cells following in vitro anti-CD3 stimulation of total CD4+ T cells (Fig. 2A). At day 4, we confirmed increased ICOS expression during activation: ICOS-expressing Treg represented 92±4% compared to 23±2% ex vivo (percentages expressed are mean±SD of a representative experiment, as defined on one sample in Fig. 2A). To note, ICOS expression was also increased on non-Treg CD4+ T cells but to a lesser extent (59±9% upon activation compared to 6±1% ex vivo). Four days after stimulation, CD127 expression was increased on FOXP3+ Treg, whereas CD127 downregulation was observed on some conventional FOXP3CD4+ T cells (Fig. 2B). Interestingly, CD127 expression on activated Treg was similar to CD127 expression on activated conventional cells (CD127 MFI: 354±123 on activated Treg compared to 248±91 on activated conventional CD4+ T cells) (Fig. 2C).

Figure 2.

Treg upregulate CD127 upon in vitro activation. Total CD4+ cells isolated from LN and spleen of three to five C57BL/6 mice were activated in vitro with anti-CD3 for 4 days. (A) ICOS and FOXP3 expression on fresh (left panel) and stimulated CD4+ T cells (right panel). (B) CD127 expression on conventional FOXP3 (upper panel) and regulatory FOXP3+ (lower panel) CD4+ cells upon in vitro activation. CD127 expression on control cells (open profile) or on activated cells (gray shaded profile) is shown. Isotype controls staining of both activated and ex vivo population are represented by the dotted and dashed lines, respectively. Cells were identified by a wide gate on FSC versus SSC plot to define lymphocytes followed by gating on CD4+ cells. (C) Mean CD127 MFI±SEM of three independent experiments. *p<0.05, two-tailed Mann–Whitney U-test.

To confirm these data in vivo, we used a classical adoptive transfer model: sorted CD4+ CD25+ Treg were co-injected with CD45.1+ naïve CD4+ T cells into Rag-deficient hosts. Analysis was performed 6 wk after transfer and Treg activation was confirmed by CD103 and ICOS expression (data not shown). Transferred conventional FOXP3CD4+ T cells expressed lower levels of CD127 when compared to conventional CD4+ T cells isolated from WT mice (MFI: 196±13 and 353±45, respectively) (Fig. 3A and B). In contrast, activation of Treg induced higher CD127 expression compared to ex vivo isolated Treg (CD127 MFI: 154±33 after transfer compared to 89±10 ex vivo) (Fig. 3A and B). To determine whether higher CD127 expression relies on upregulation or expansion of pre-existing activated Treg, similar experiments were performed using “naïve-like” Treg (CD4+ CD25+CD62L+CD103CD69ICOS). CD45.2+ transferred “naïve-like” Treg expressed higher levels of CD127 than ex vivo Treg (CD127 MFI: 222±22 after transfer compared to 89±10 ex vivo) (Fig. 3C and D). High CD127 expression in “naïve-like” Treg supported upregulation rather than expansion of a pre-existing CD127hi Treg fraction. We thus demonstrated that Treg, while expressing low levels of CD127 at the steady state, significantly upregulate this molecule upon in vitro and in vivo activation.

Figure 3.

Treg upregulate CD127 upon in vivo activation. CD45.1+ naïve CD4+ T cells and CD45.2+ Treg or “naïve-like” Treg were sorted by FACS and co-injected into RAG-deficient mice. Phenotypic analysis was performed on mesenteric LN and spleen 6 wk after transfer. Cells were identified by gating of lymphocytes on FSC/SSC plot, followed by gating on CD4+ cells. CD45.1+FOXP3CD4+ cells and CD45.1FOXP3+ Treg were further identified. (A, C) CD127 expression on conventional FOXP3 (upper panel) and regulatory FOXP3+ (lower panel) CD4+ cells during in vivo activation when (A) total or (C) “naïve-like” Treg were co-injected. Isotype controls of activated and control fresh cells are represented by dotted and dashed lines. (B, D) CD127 MFI on CD4+ cells isolated from control WT mice (open profile) or transferred cells (gray shaded profile) when total (B) or “naïve-like” Treg (D) were co-injected. Data show mean±SEM of three independent experiments using two to three mice per group in each experiment. *p<0.05, **p<0.01, two-tailed Mann–Whitney U-test.

Activated Treg expressed CD127 in a model of acute skin inflammation

Although adoptive transfer is a useful model to study Treg activation and suppressive capacity in vivo, it requires transfer into empty hosts, which may introduce some bias concerning CD127 expression analysis. It is well known that IL-7 availability directly interferes with CD127 expression at T-cell surfaces 16, 17 and that IL-7 cytokine levels are increased in lymphopenic states 26, 27. We wanted to confirm our results in a non-lymphopenic model of Treg activation: a model of acute skin inflammation. Ear skin exposure to 2,4-dinitrofluorobenzene (DNFB) induces a strong inflammatory response accompanied by ear swelling and activation of resident and recruited T cells 28. Conventional and Treg activation was studied using ICOS expression. Indeed, ICOS is the most consistent marker for Treg activation, whereas CD103 has been described to reflect tissue-specific Treg homing. Moreover, being a shared activation marker for conventional and regulatory CD4+ T cells, it also allows distinguishing activated conventional T cells. Analyses were performed 6 days after DNFB treatment. Inflammation was confirmed by significant increase in ear thickness (Fig. 4A) and increased CD4+ T-cell numbers in cervical LN draining the ear skin (Fig. 4B). In draining LN, a fraction of conventional CD4+ FOXP3 cells underwent activation as revealed by ICOS upregulation. Simultaneously, an important fraction of CD4+ FOXP3+ Treg was activated and expressed ICOS (46±6%). We next determined CD127 expression on activated regulatory and conventional CD4+ T cells as identified by ICOS expression. As expected, ICOS-expressing conventional CD4+ T cells downregulated CD127 at their surface. However, ICOS+ Treg expressed higher levels of CD127 than their negative counterpart (Fig. 4C). To assess CD127 expression on CD4+ T-cell subsets at the site of inflammation, we analyzed lymphocytes infiltrating ear skin in both control and inflamed ear skin. In control mice, an important proportion of conventional and regulatory CD4+ T cells infiltrating ear skin expressed ICOS (Fig. 4D) suggesting that resident CD4+ T cells were in an activated state. During acute inflammation, a more diverse phenotype was observed both on conventional T cells and Treg, revealing notably an increasing proportion of ICOS conventional and Treg. The rationale for the appearance of phenotypically less activated T cells (as assessed by low ICOS expression) is unclear but may reflect recruitment from peripheral blood (PB) or downregulation of ICOS on pre-existing resident T cells during the acute phase of inflammation. Nevertheless, we assessed CD127 expression on skin-infiltrating CD4+ T-cell subsets expressing ICOS or not, similar to what we performed in draining LN. As shown in Fig. 4D, a fraction of ICOS+ conventional CD4+ T cells downregulated CD127 when compared to the ICOS population, whereas a fraction of activated ICOS+ Treg expressed higher levels of CD127 than their negative counterpart. The appearance of CD127lo Treg among ICOS+ cells in acute inflamed skin may also rely on recruitment or migration of PB-derived T cells into inflamed sites but may also reflect different kinetics of expression of ICOS and CD127 on activated Treg during the acute phase of inflammation. However, in both cases, high CD127 expression was predominantly observed on activated Treg in ear skin. Collectively, these in vivo results demonstrated that upon activation, conventional CD4+ T cells downregulate CD127, whereas Treg, which express low levels of CD127 at the steady state, strongly upregulated this molecule. These results also confirmed that CD127 expression on activated Treg can be equivalent to conventional T cells.

Figure 4.

High CD127 expression on activated Treg in a model of acute skin inflammation. (A) Ear thickness and (B) CD4+ T-cell numbers in cervical LN draining ear skin 6 days after DNFB treatment. Data show mean±SD of five mice from one representative of at least three independent experiments. **p<0.01, two-tailed Mann–Whitney U-test. (C, D) ICOS and CD127 expression on CD4+FOXP3 and CD4+FOXP3+ T cells from (C) draining LN and (D) ear skin. Histogram profiles show CD127 expression on ICOS (thick black line) or ICOS+ (gray shaded profile) cells from DNFB-treated mice. Isotype controls of ICOS and ICOS+ subsets are represented by the dotted and dashed lines, respectively. Gating strategy is shown in Supporting Information Fig. 1B. This figure is representative of at least three independent experiments using pool of 5–15 mice per experiment.

Different CD127 expression on Treg depending on their tissue localization

Considering the high CD127 expression profile from skin resident Treg in non-immunized mice, we next analyzed CD127 expression on Treg among various sites: LN, spleen (SP), PB, thymus, bone marrow (BM), peritoneal cavity, intestine, lung, liver and skin. CD127 expression was determined on tissue resident FOXP3+ and conventional FOXP3CD4+TCR+ cells and isotype staining was performed as control (Fig. 5A). CD127 expression profile on Treg differed depending on the organs studied (Fig. 5B and C). Various proportions of Treg expressing high levels of CD127 were recovered, ranging from 10±3% in the intestine to in the skin. PB (15±8%), LN (17±4%), thymus (19±1%) and SP (28±6%) exhibited intermediate proportion of CD127-expressing Treg. Strikingly, BM exhibited high percentage of CD127-expressing Treg (64±3%) (Fig. 5B). Similar results were obtained analyzing MFI (Fig. 5C). Highest levels of CD127 expression on Treg were observed in BM and skin (CD127 MFI 312±57 in BM and 383±73 in skin compared to 85±15 in LN). In these two organs, no concomitant upregulation was observed on conventional T cells when compared to LN counterpart, suggesting a Treg-restricted phenomenon. Our results demonstrated differential CD127 expression on Treg depending on their localization and they confirmed that low CD127 expression was not an intrinsic characteristic of Treg.

Figure 5.

Different CD127 expression on Treg depending on the organs studied. (A) Ex vivo CD127 and FOXP3 expression on CD4+TCRβ+ cells isolated from LN, SP, peripheral blood (PB), thymus, BM, peritoneal cavity (PC), intestine, lung, liver and skin of C57BL/6 mice. Isotype controls are presented for all organs. FSC/SSC plot was used to gate lymphocytes in all organs except skin and intestine. In these organs, CD4 gating was directly performed in a CD4/FSC plot and followed by back-gating to the FSC/SSC plot to select more accurately lymphocytes and consequently TCRβ+CD4+ cells. Quadrants were defined based on isotype controls. Percentages of cells in each quadrant are indicated. (B) Percentages of CD127-expressing cells among FOXP3+CD4+ T cells in different organs are shown. (C) CD127 MFI on FOXP3 (white bars) and FOXP3+ (grey bars) CD4+ T cells. Error bars represent SEM of three independent experiments. Data are representative (A) or show mean±SEM (B, C) of at least three independent experiments, with 5–15 mice pooled in each experiment depending on the organs studied.

Higher levels of CD127 expression on Treg correlated with increased IL-7 signal transduction

To assess whether CD127 expression on Treg was associated with efficient IL-7 signal transduction, we examined STAT5 phosphorylation following IL-7 incubation in vitro. We used Foxp3-GFP reporter mice to identify Treg; similar results were obtained using CD25 as a marker for Treg identification. IL-2 treatment induced robust STAT5 phosphorylation in CD4+FOXP3+ but not in conventional CD4+FOXP3 T cells (Fig. 6A). IL-7 treatment induced significant increase in phospho-STAT5 in Treg but to a lesser extent than observed in conventional CD4+FOXP3 cells, as previously described 13, 29, 30. Importantly, a dose-dependent effect was detectable among Treg but not among conventional T cells (Fig. 6A). We next studied STAT5 phosphorylation depending on CD127 expression on Treg. As shown in Fig. 1C, Treg exhibited heterogeneous CD127 expression. We thus determined pSTAT5 signaling on CD127hi Treg (top 10% of Treg gate) and Treg exhibiting low but significant CD127 expression (lowest 50% of Treg gate). Following treatment with low doses of IL-7 (1 ng/mL), an approximately eightfold increase in phospho-STAT5 MFI was detectable among CD127hi Treg compared to a twofold increase in CD127lo cells (Fig. 6B). A similar pattern of STAT5 phosphorylation in the two Treg sub-populations was observed after treatment with lower or higher doses of IL-7. These results indicate that CD127hi Treg transduced IL-7 signals more efficiently than CD127lo Treg.

Figure 6.

Different IL-7 signal transduction in Treg depending on CD127 expression. Cells from Foxp3-GFP mice were incubated with either medium alone, IL-2 (50 U/mL), or IL-7 (0.1, 1, 10 ng/mL) for 30 min. STAT5 phosphorylation was determined on FOXP3+ or FOXP3 cells isolated from (A) pooled LN and splenic cells or (C) BM cells. (B) p-STAT5 staining depending on CD127 expression among FOXP3+ LN and spleen cells; CD127hi Treg were identified by gating on top 10% CD127-expressing CD4+GFP+ cells (grey shaded profile); CD127lo Treg were defined as 50% of CD4+GFP+ cells expressing the lowest CD127 levels (open profile). (D) Comparison of p-STAT5 staining among LN and spleen or BM isolated Treg. Gating strategy is shown in Supporting Information Fig. 1C. Results are representative of three independent experiments, using one Fopx3-GFP mouse for each experiment.

Because Treg localized in the bone marrow (BM Treg) expressed high levels of CD127 at the steady state, we wanted to assess whether BM Treg could be more reactive to IL-7 than Treg from LN and SP. Interestingly, in vitro stimulation with IL-7 induced comparable amounts of phospho-STAT5 in conventional and Treg from BM at all doses studied. Maximal levels of STAT5 phosphorylation were reached in BM Treg after exposure to very low doses (0.1 ng/mL) of IL-7 (Fig. 6C). Accordingly, we found that BM Treg expressed significantly higher levels of phospho-STAT5 than secondary lymphoid organs-derived Treg upon IL-7 stimulation at all IL-7 doses studied (Fig. 6D).

CD127 is also part of the receptor for thymic stromal-derived lymphopoietin (TSLP). We thus assessed whether differential levels of CD127 expression could influence the capacity of Treg to react to TSLP as well. TSLP induced a dose-dependent increase in STAT5 phosphorylation in both conventional and, to a lesser extent, Treg (Supporting Information Fig. 2A). No difference in STAT5 phosphorylation was detected in CD127hi and CD127lo Treg upon TSLP treatment at the doses of 1 and 10 ng/mL and only a slight increase in pSTAT5 was detected in CD127hi Treg compared to CD127lo Treg upon exposure to high doses (100 ng/mL) of TSLP (Supporting Information Fig. 2B). Similar to what we performed for IL-7, we assessed whether BM Treg were more reactive to TSLP than Treg from LN and SP. Only a slight increase in pSTAT5 was detected in both conventional T cells and Treg from the BM upon high doses of TSLP treatment (Supporting Information Fig. 2C). Moreover, no significant difference between STAT5 phosphorylation in BM Treg and Treg from secondary lymphoid organs was observed (Supporting Information Fig. 2C). Collectively, our results demonstrate that Treg that express more cell surface CD127 are more efficient at perceiving and transducing signals specifically mediated by IL-7.

IL-7 preferentially increases the in vitro survival of Treg expressing high levels of CD127

As we demonstrated that CD127hi Treg transduced IL-7 signals more efficiently, we assessed whether IL-7 could exert functional effect preferentially on these cells. Conflicting data have been reported so far concerning the role of IL-7 on Treg survival. We thus re-address this question considering high and low CD127 expression on Treg. To this aim, we cultured FACS-sorted total, CD127lo or CD127hi CD4+ CD25+ cells (Fig. 7A) in the presence or absence of IL-7. Upon overnight incubation, IL-7 significantly increased in vitro survival of total CD4+CD25+ cells (p=0.003) (Fig. 7B). When considering CD127 expression, a minor but statistically significant increase in cell survival was observed on CD127lo Treg upon IL-7 treatment (p=0.024). Interestingly, a highly significant increase in survival was detected on IL-7 treated CD127hi Treg compared both to non-treated cells (p=0.0002) and to IL-7 treated CD127lo cells (p=0.0003). In conclusion, our results indicate that IL-7 preferentially promotes in vitro survival of Treg expressing higher levels of CD127.

Figure 7.

IL-7 preferentially promotes in vitro survival of Treg expressing higher levels of CD127. FACS-sorted total, CD127lo or CD127hi CD4+CD25+cells were cultured overnight in medium alone or in the presence of IL-7 (10 ng/mL). (A) Gating strategy to isolate the top 10% CD127-expressing CD25+ T cells (H gate) and CD127lo Treg (L gate) by FACS. (B) Percentage of live cells after culture of indicated Treg populations in medium alone (open bars) or in the presence of IL-7 (filled bars). Live cells were defined as LIVE/DEAD negative cells among CD4 expressing lymphocytes. FACS sorting was performed on pooled mesenteric LN isolated from fifteen mice for each experiment. Data show mean±SD (n=3) and are representative of three independent experiments. *p<0.05, **p<0.01, ***p<0.001, unpaired two-tailed Student's t-test.

Discussion

IL-7 and Treg are the two main regulators of homeostasis and immune responses. In contrast to their constitutive expression of IL-2Rα (CD25), FOXP3+CD25+CD4+ Treg have been shown to exhibit low expression of CD127 compared to conventional CD4+ and CD8+ peripheral T cells 10–12. Importantly, regulation of CD127 expression has proven crucial during thymocyte maturation 31 and it has been suggested to be a crucial step for effector or memory differentiation 32–34. We thus further investigated CD127 expression on Treg. When dissecting Treg heterogeneity in lymphoid organs, CD127 expression on Treg was associated with the expression of CD103 and ICOS. These two markers have been shown to identify activated Treg 23, 35. Using various context of activation (in vitro, model of adoptive transfer, skin inflammation model), we demonstrated that high CD127 expression on Treg was predominant during ongoing immune responses. These data differ from what has been described for conventional T cells which downregulate transiently IL-7Rα expression during activation and suggest different regulation of CD127 expression on conventional and Treg. Whether TCR ligation and IL-2 signaling differently regulate CD127 expression or whether TCR ligation differs between conventional and Treg remain to be further investigated. Intrinsic difference in Treg intracellular machinery may also be responsible for such differential regulation. Interestingly, the opposite regulation of CD127 on Treg and effector cells during activation led to equivalent CD127 expression on these two subsets when activated. These data provide an additional mechanism for the functional impact of IL-7 on Treg suppression, which has been shown to rescue effector cells from Treg-mediated apoptosis 4, 36. Our data support the hypothesis that Treg and conventional T cells may compete for IL-7 during activation. They also support our previous results demonstrating that IL-7R expression on Treg was essential for Treg-mediated suppression of conventional T cells during lymphopenia-induced proliferation 37.

We also extended our analysis to peripheral tissues that are important localization for activated Treg. Depending on the organs studied, we observed strikingly different CD127 expression profile on Treg compared to Treg isolated from LN. PB and secondary lymphoid organs analysis revealed low CD127 expression on Treg ex vivo. Contrasting profiles were observed in mucosal sites: high CD127 expression on Treg was predominant in the skin but barely detectable in the intestine. Such heterogeneity among various mucosal sites suggests high CD127 expression is not a ubiquitous marker of activated Treg. Difference in CD127 expression depending on the tissue considered may rely on various mechanisms: differential antigenic load, differential kinetics of activation, differential activation/differentiation pathways and/or specific signals provided in specific sites, differential migration capacities of high CD127-expressing Treg, differential persistence depending on IL-7 (or TSLP) production on sites 38. These latter hypotheses may also stand for the high percentage of CD127-expressing Treg in the BM. The physiological relevance of such a specific profile in the BM is currently under investigation and may reflect local IL-7 production. Our data questioned Treg analyses excluding CD127-expressing cells 39, 40. Such protocol, commonly used in clinical analyses, allows isolating a highly enriched Treg population but it may also exclude an activated population that could be especially relevant in chronic immune responses and pathological contexts. Although CD127hi Treg are minor in PBMC, such a protocol may prove debatable depending on the tissue and the context considered.

Demonstrating differential CD127 expression among Treg subsets did not formally demonstrate any preferential reactivity to IL-7. To ascertain whether high CD127 expression enhanced IL-7-mediated signaling among Treg, we determined STAT5 phosphorylation upon in vitro IL-7 or TSLP incubation. Two strategies were considered: comparing high and low CD127-expressing Treg isolated from LN and SP or comparing total Treg isolated from secondary lymphoid organs and BM, which exhibit significant difference in CD127 expression. In both analyses, CD127hi Treg exhibited higher phosphorylation of STAT5. Interestingly, IL-7-induced STAT5 phosphorylation in CD127hi Treg was comparable with that detected in conventional FOXP3CD4+ T cells. These results confirmed the functional relevance of differential CD127 expression on Treg in terms of reactivity to IL-7. A similar approach was performed to study the reactivity to TSLP, an IL-7-like cytokine expressed by epithelial cells, including keratinocytes, and important in allergic inflammation 41, 42. We showed that TSLP induced STAT5 phosphorylation in Treg, but we failed to reveal any association between CD127 expression on Treg and TSLP signal transduction. Nevertheless, one may hypothesize that activated Treg could compete with effector T cells for TSLP and this may be one of the mechanisms by which Treg dampen inflammation at tissue sites. Collectively, these data demonstrated that activated Treg are an important target of IL-7/IL-7Rα signaling. One obvious hypothesis is to consider IL-7 as an additional survival factor for activated Treg. So far, minor influence of IL-7/IL-7Rα signaling on peripheral Treg has been reported in accordance with the low level of CD127 expression described on Treg 13–15. However, Pandiyan et al. demonstrated that IL-7 addition enhances Treg survival during in vitro activation 43. High CD127 expression on activated Treg reconciles these conflicting data since experimental protocols studying Treg sensitivity to IL-7 were performed either on resting 13 or activated Treg 43. Indeed, we demonstrated enhanced in vitro survival of CD127hi Treg compared to CD127lo Treg upon 20 h IL-7 incubation. Finally, demonstrating high CD127 expression on activated Treg may be especially important during lymphopenic episodes 44 and in regard to the development of IL-7 based therapies in cancer and HIV infected patients notably 45–50.

In conclusion, we demonstrate that Treg exhibit high CD127 expression when activated in contrast to their low expression in non-immunized settings. Associated with the recent observation that DC express CD127 27, it is remarkable to identify the three main partners involved in immune T-cell responses, i.e. conventional T cells, DC and Treg, as potential targets of IL-7. Altogether, these data provide additional insights into the ubiquitous impact of IL-7 on various peripheral T-cell subsets. Importantly, our data substantiate a direct link between two main components of immune regulation that are Treg and IL-7 and identify activated Treg as a potential target of endogenous or therapeutic IL-7.

Materials and methods

Mice

Six- to eight-week-old C57BL/6 mice expressing CD45.1 or CD45.2 (Charles River, Janvier) and B6 RAG-2-deficient mice were used as donors and hosts, respectively. Foxp3-GFP mice were purchased from the Jackson Laboratory. All mice were kept under specific pathogen-free conditions and all experiments were performed according to institutional guidelines of the European Community.

Cell suspension

Single cell suspensions were prepared from LN, SP, liver, thymus and BM in HBSS containing 2% FBS (both from PAA Laboratories GmbH). For preparation of intestinal cell suspensions, colon pieces were incubated in PBS containing 5 mM EDTA for 15 min on a shaking incubator at 37°C prior to digestion. Supernatant was collected and remaining tissue was further digested. To digest tissues, intestine, lung and ear skin were finely minced and stirred in HBSS with 400 μg/mL Liberase (Roche) and 50 μg/mL Collagenase IV (Sigma) for 1 h at 37°C. Cells isolated from intestine, liver and lung were pelleted, resuspended in 37% Percoll (GE Healthcare) and spun at 2500 rpm for 25 min. Isolated lymphocytes were washed twice and used for subsequent flow cytometry analyses.

Flow cytometry

Extracellular staining was preceded by incubation with purified anti-CD16/32 Ab (FcgRII/III block, 2.4G2) (eBioscience) to block non-specific staining. Cells were stained with FITC-, PE-, PECy5-, PECy7-, APC- and APCAlexa750-labeled or biotinylated appropriate Ab including CD4 (GK1.5); TCR (H57-597); CD62L (MEL-14); CD69 (H1.2F3); CD25 (PC61.5); CD127 (A7R34); ICOS (7E.17G9); CD103 (2E7) or appropriate isotype Ab. Streptavidin-FITC, PECy5 or PECy7 were used to develop biotinylated Ab. All Ab were purchased from eBioscience. Intranuclear FOXP3 staining was performed using eBioscience APC-conjugated FOXP3 staining buffer set (FJK-16s). Six-color flow cytometry was performed with a FACSCanto cytometer (BD Biosciences) and data files were analyzed using FlowJo software (Tree Star Inc.).

Cell purification

Cell sorting of naïve CD4+ T cells (CD4+CD62L+CD25) and Treg (CD4+CD25+) was performed on a FACSVintage cell sorter (BD Biosciences). For suppressive assays and adoptive transfer experiments, Treg subsets were sorted based on CD62L, CD69, ICOS and CD103 expression following positive selection using anti-CD25 PE or biotinylated Ab incubation, and purification with anti-PE or anti-biotin beads, respectively (Miltenyi). For in vitro survival experiments, cells were isolated from mesenteric lymph nodes. CD127hi Treg (top 10% of Treg gate) and Treg exhibiting low CD127 expression (approximately lowest 50% of Treg gate) were next FACS sorted among CD4+CD25+ cells.

In vitro cell cultures

Cells were cultured in RPMI 1640 medium containing penicillin, streptomycin, L-glutamine, HEPES buffer, non-essential amino acids, sodium pyruvate and β-mercaptoethanol and 10% heat inactivated FBS (all from PAA Laboratories GmbH). For activation experiments, total CD4+ T cells (1×105) were cultured in 96-well plates with splenocytes from RAG−/− mice (1×105) and 0.5 μg/mL anti-CD3.

Suppression assays

Suppression assays were performed using allotypic marker CD45.1/CD45.2 to distinguish conventional from CD25+ sorted population. CFSE (Sigma) labeled CD45.1+CD62L+CD25CD4+ cells (3×104) were cultured for 72–96 h in 96-well plates with splenocytes from RAG−/− mice (3×104), 0.5 μg/mL anti-CD3 (eBiosciences), in the presence or absence of the indicated subset of CD4+CD25+CD45.2+ cells. CFSE labeling was performed using standard methods.

Adoptive transfer experiments

FACS sorted CD45.1 naïve CD4+ T cells (1×105) and CD45.2 Treg (0.5×105) were co-injected into RAG2-deficient mice. Six weeks after transfer, SP and mesenteric LN were harvested.

Contact dermatitis model

Acute skin inflammation experiments were conducted as described 28. Briefly, 0.5% DNFB (Sigma) diluted in acetone/olive oil vehicle (4:1 v/v) was applied onto ears of C57BL/6 mice. Control mice were treated with vehicle alone. Six days later, ear thickness was measured with a digital caliper, cervical draining LN and ears were isolated.

STAT5 phosphorylation experiments

Cells isolated from Foxp3-GFP mice were incubated for 30 min at 37°C with or without IL-2, IL-7 (Immuno Tools) or TSLP (R&D) and immediately fixed in 2% paraformaldehyde. Cells were made permeable by incubation in 90% methanol and then were stained with primary rabbit Ab to phosphorylated STAT5 or isotype control (Cell Signaling) revealed with an anti-rabbit Alexa-647 conjugated secondary Ab.

In vitro survival experiments

5×104 FACS-sorted total CD4+CD25+, CD4+CD25+CD127lo or CD4+CD25+CD127hi were cultured overnight (18–20 h) in medium alone or in the presence of IL-7 (10 ng/mL). Cell survival was determined using the fluorescence-based LIVE/DEAD assay (Invitrogen).

Acknowledgements

This work was supported by the ANRS (Agence Nationale de la recherche contre le SIDA et les hépatites virales) and Fondation de France for Federico Simonetta, Alejandra Urrutia, Isabelle Girault and Dr. Christine Bourgeois. Dr. Corinne Tanchot is supported by ANR (Agence Nationale de la Recherche). The authors thank Dr. Alain Venet for critical reading and helpful discussion, Pr. Jean-François Delfraissy and Pr. Marc Tardieu for their support.

Conflict of interest: The authors declare no financial or commercial conflict of interest.

Ancillary