A unique subpopulation of CD4+ regulatory T cells controls wasting disease, IL-10 secretion and T cell homeostasis

Authors


Abstract

CD25+CD4+ regulatory T cells have major roles in controlling immune responses, and use heterogeneous regulatory mechanisms. It is possible that these different activities aremediated by different subsets. Here we show that CD103+CD25+CD4+ T cells (that control inflammatory bowel disease) are highly enriched in gut-associated lymphoid tissue and have unique functional properties. In vivo, only this subpopulation is able to control wasting disease and peripheral T cell homeostasis. In vitro, only this subpopulation is able to regulate IL-10 secretion, and it might also mediate infectious suppression. These results demonstrate that regulatory T cells can be divided into discrete subpopulations with defined functional properties and regulatory mechanisms.

Abbreviations:
B6:

C57Bl/6

Ba:

C57Bl/Ba

Ct:

Threshold cycle

IBD:

Inflammatory bowel disease

KO:

Knockout

LN:

Lymph nodes

PP:

Peyer's patches

1 Introduction

CD25+CD4+ T cells with regulatory functions are major actors in the control of immune responses. They participate in immune tolerance 1, 2and in the regulation of inflammatory responses 36. They also participate in T cell homeostasis 79 and contribute to the induction and maintenance of graft tolerance 10, 11. The regulatory mechanisms used by these cells are heterogeneous. In vivo studies strongly implicate cytokines, but the cytokines involved vary according to the experimental setting. Protection from inflammatory bowel disease (IBD) is dependent on both TGF-β and IL-10 12, 13. Control of cytokine over-production in IL-2-Rα-knockout (KO) mice is dependent on IL-10 5, but inhibition of thyroiditis in the rat is mostly dependent on TGF-β secretion 14. In contrast, inhibition of CD25CD4+ T cell proliferation by CD25+CD4+ T cells in vitro appears not to be cytokine-mediated but rather to be cell-contact-dependent 1517. Moreover, IL-10 is not essential in the control of inflammatory cytokine production by CD25CD4+ T cells in vitro5.

As a whole, these results point to the existence of multiple regulatory mechanisms. The phenotype of CD25+CD4+ T cells is also heterogeneous 18. It is therefore possible that some activities of CD25+CD4+ T cells are mediated by unique subsets. Identification of these subsets and the mechanisms by which they exert their function is amajor aim in the understanding of regulation, and in the possible therapeutic use of regulatory cells.

It has been recently reported that a subpopulation of regulatory T cells express the integrin αEβ7 (CD103) 1921. CD103 is expressed by T lymphocytes adjacent to mucosal epithelia, including 90% of CD8+ intestinal T cells and 40–50% of CD4+ T cells located within the intestinal lamina propria 22. This integrin is involved in T cell function, as anti-human-αE mAb enhance T cell proliferation in response to suboptimal anti-CD3 stimulation 23, 24. CD103 also mediates adhesion to epithelial cells through its binding to E-cadherin, which is expressed selectively on epithelial cells 2527. CD103 is thus important for localization of T cells in the skin and intestine 28. Regulatory cells expressing CD103 were recently shown to have a preferential capacity to prevent IBD 29. Expression of this integrin may also be involved in immune regulation within the skin, as αE-deficient mice develop autoimmune-like skin lesions when back-crossedinto susceptible backgrounds 30. Here we studied the preferential location of CD103+CD25+CD4+ regulatory cells and their unique functional properties.

2 Results

2.1 αEβ7 integrin expression defines a subpopulation of CD25+CD4+ regulatory T cells with a particular location and phenotype

As regulatory CD103+CD25+CD4+ T cells have a predominant role in the control of IBD 29, we determined whether these cells were preferentially located in gut-associated lymphoid tissue. CD25+CD4+ T cells were present in all the lymphoid organs examined (Fig. 1A). Their frequency was lowest in the thymus (5%) and highest in lymph nodes (LN) (13%). The frequency of CD103+ cells among CD25+ cells differed in the different lymphoid organs (Fig. 1B), being lowest in the thymus and highest in lymphoid organs draining the gut. CD103+ cells represented 26% of CD25+ cells from mesenteric LN, and 39% of CD25+ cells from Peyer's patches (PP). The vast majority of CD103+ cells co-expressed CD25. Indeed, CD25CD4+ T cells contained only 1.5% CD103+ cells (data not shown).

We had previously shown that CD25+CD4+ T cells are CD45RBlow and express far higher levels of activation markers than CD25CD4+ T cells 18. Both CD103+ and CD103CD25+CD4+ T cell subsets were CD45RBlow. However, the expression level of CD54, CD44, CD69 and CTLA-4 was much higher on CD103+CD25+CD4+ T cells than on their CD103 counterparts (Fig. 1C). Thus, CD103+CD25+CD4+ T cells showed signs of increased activation compared with CD103CD25+CD4+ T cells. Expression of these different antigens was identical whether CD103+CD25+CD4+ T cells were recovered from peripheral LN, mesenteric LN or spleen (not shown). These results indicate that CD103+CD25+CD4+ T cells have the same activated phenotype, independently of their location.

Figure 1.

 αEβ7 integrin expression defines a subpopulation of CD25+CD4+ regulatory T cells with a particular location and phenotype. CD4+ T cells from peripheral and mesenteric LN, spleen and PP were stained with antibodies known to discriminate CD25+CD4+ T cells from CD25CD4+ T cells. The percentage of CD25+ cells among CD4+ T cells (A), the percentage of CD103+ cells among CD25+CD4+ T cells (B) and the expression of surface markers on CD103+CD25+CD4+ (bold line), CD103CD25+CD4+ (thin line) and CD25CD4+ T cells (dashed line) (C) are shown as histograms. The dotted line represents control staining. Representative data from one of five independent experiments (three mice per experiments) are shown.

2.2 Cytokine and cytokine-receptor mRNA expression in subpopulations of regulatory T cells

We used real-time RT-PCR to determine relative cytokine and cytokine-receptor mRNA expression by these different populations of regulatory T cells before and after in vitro activation with anti-CD3 (Fig. 2). Primers were designed so that all RT-PCR reactions had the same efficiency, thereby allowing us to compare expression among different cDNA, and to determine relative copy numbers directly by comparison with a house-keeping gene (HPRT). Such primer combinations are usually referred to as multiplexes.

We first confirmed that our RT-PCR reactions indeed had the same efficiency for each cDNA. The efficiency of real-time RT-PCR is reflected by the slope of the curve of amplicon accumulation in consecutive cycles. In multiplex PCR, the slopes corresponding to the different genes should be parallel. For this purpose we used cDNA prepared from mouse gut intraepithelial lymphocytes, which are known to contain mRNA coding for all the cytokines and cytokine-receptors of interest. As shown in Fig. 2A, amplification efficiency was the same for all the genes of interest. To calculate relative copy numbers, the threshold cycle (Ct) was determined in the linear phase of PCR and corrected for HPRT amplification studied simultaneously.

CD103+CD25+CD4+ T cells had a very different cytokine profile from their CD103 counterparts (Fig. 2B). It has been reported that only CD103+CD25+CD4+ T cells produce IL-10 after activation 29. We found that CD103+CD25+CD4+ T cells expressed IL-10 mRNA even before activation, and that the message was strongly up-regulated on activation by anti-CD3. These cells spontaneously expressed very high levels of TGF-β mRNA; indeed, TGF-β mRNA expression was 18-fold higher than that of IL-10 in activated CD103+CD25+CD4+ T cells. The CD103+CD25+CD4+ T cells expressed more INF-γ mRNA than other CD4 populations, whereas IL-2 mRNA was not detected. In contrast, CD103CD25+CD4+ T cells did not express IL-10 mRNA; indeed, IL-10 mRNA expression remained virtually undetectable even after activation. Surprisingly, CD103CD25+CD4+ T cells expressed IL-2 mRNA after activation, similarly to activated CD25CD4+ T cells. The three populations expressed similar levels of TNF-α mRNA (Fig. 2B), very little IL-4 mRNA and no TNF-β mRNA (data not shown).

These results show that the cytokine profile thought to characterize all CD25+CD4+ regulatory cells (high IL-10, low IL-2) is only found in the CD103+CD25+CD4+ population. CD103CD25+CD4+ T cells have a cytokine expression profile resembling that of CD25CD4+ T cells (low IL-10, high IL-2).

Some of the suppressive effects of regulatory cells may be mediated by IL-10 or TGF-β. To determine which cell types were able to respond to these cytokines, we evaluated cytokine receptor mRNA expression in different CD4 T cell populations. We found that only CD103+CD25+CD4+ cells expressed very high levels of IL-10-Rα mRNA (Fig. 2C). None of the CD4+ populations expressed TGF-β-R1 mRNA (not shown), which is required for TGF-β signal transmission 31, or TGF-β-R3 mRNA, which contributes to the expression of the high-affinity TGF-β-receptor 32. TGF-β-R2 mRNA (which captures and presents TGF-β) was mainly expressed after T cell activation. These results suggest that CD103+CD25+CD4+ cells can respond preferentially to the IL-10 they produce, whereas none of the CD4+ populations can respond directly to TGF-β. As expected, both CD103+ and CD103 CD25+CD4+ T cells expressed IL-2-R-α mRNA constitutively. On anti-CD3 activation, IL-2-R-α mRNA expression was enhanced in CD25+CD4+ T cell subsets and induced in the CD25CD4+ T cell subset.

Figure 2.

 IL-10 and IL-10-Rα mRNA expression is restricted to the CD103+CD25+ T cell subpopulation. Real-time quantitative PCR analysis (A) of cytokine mRNA (B) and cytokine-receptor mRNA (C) was performed on purified CD103+CD25+CD4+, CD103CD25+CD4+ and CD25CD4+ LN T cells before and after activation of each subset by anti-CD3 in vitro for 4 h. To calculate relative copy numbers, the Ct was determined in the linear phase of PCR, and corrected for HPRT amplification studied simultaneously. All data are arbitrary units and are representative of two independent experiments performed in triplicate.

2.3 Functional properties of CD103+ regulatory cells in vitro

It has been reported that both CD103+ and CD103 subpopulations inhibit the proliferation of CD25CD4+ T cells 29, 33. We confirmed this report (Fig. 3), and investigated the role of IL-10 produced by CD103+CD25+CD4+ T cells in the control of proliferation in vitro. CD25CD4+ LN T cells were cultured alone or together with each regulatory subset from normal mice or IL-10 KO mice (at a 1:1 ratio) for 2 days, in the presence of anti-CD3 and PMA. As shown in Fig. 3A and B, proliferation of CD25CD4+ T cells was inhibited by CD103+CD25+CD4+, CD103CD25+CD4+ and CD25+CD4+ T cells from both normal and IL-10 KO mice. Thus, both CD103+ and CD103 CD25+CD4+ T cells were able to suppress the anti-CD3-induced proliferation of CD25CD4+ T cells by an IL-10-independent mechanism in vitro.

We have previously shown that CD25+CD4+ regulatory T cells are able to regulate IL-10 secretion. Indeed, IL-10 production in co-cultures of CD25+CD4+ regulatory T cells with CD25CD4+ T cells is strongly increased when compared with that of each population cultured alone 34, a finding later confirmed by other authors 35. In these cultures, CD25+ cells induced IL-10 production by CD25 cells. Here, we studied the role of CD103+CD25+CD4+ T cells in the regulation of IL-10 secretion. CD103+ cells strongly up-regulated IL-10 secretion in co-culture with CD25CD4+ T cells, whereas CD103 regulatory cells had no effect (Fig. 4A). Intracytoplasmic staining with anti-IL-10 (Fig. 4B) revealed that the CD25 population, when incubated with CD103+CD25+CD4+ T cells, became able to secrete IL-10. However, we found that most of the IL-10 secreted in these co-cultures originated from CD103+ cells themselves (Fig. 4B). The increase in IL-10 secretion was totally abrogated when CD103+ cells were obtained from IL-10 KO mice (Fig. 4A and B).

In conclusion, CD103+ cells have a unique regulatory role in IL-10 secretion. They enhance their own IL-10 production when activated and co-cultured with CD25CD4+ T cells. They also induce IL-10 secretion by the CD25 subset in an IL-10-dependent manner. In contrast, CD103CD25+CD4+ T cells are unable to regulate IL-10 secretion in vitro.

Figure 3.

 Both the CD103+ and CD103 subsets of CD25+CD4+ T cells suppress the proliferation of CD25CD4+ T cells, by an IL-10-independent mechanism. Sorted (1×105) CD103+CD25+CD4+, CD103CD25+CD4+ and CD25+CD4+ LN T cells from wild-type (WT) (A) and IL-10 KO mice (B) were cultured for 2 days with anti-CD3 and PMA in the presence of 1×105 CD25CD4+ T cells. Proliferation was assessed on the basis of thymidine incorporation (counts per minute [CPM]). Representative data are shown from one of two independent experiments performed in triplicate.

Figure 4.

 Only CD103+ CD4+ T cells increase IL-10 production by CD25CD4+ T cells in co-culture. CD103+CD25+CD4+ and CD103CD25+CD4+ LN T cells from B6 and IL-10 KO mice were cultured for 2 days with anti-CD3 and PMA in the presence of CD25CD4+ T cells from Ba mice at a 1:1 ratio. IL-10 was assayed with ELISA kits (A) or by intracellular staining (B). Results are expressed as picograms per milliliter of supernatant. Mean values of two independent experiments and representative data from one of two independent experiments are shown.

2.4 CD103+CD25+CD4+ regulatory T cells have a unique role in preventing wasting disease

It has been reported that CD103+ regulatory T cells have a preferential role in protecting T-cell-deficient mice against the aggressive effect of CD25CD4+ T cells, limiting the development of IBD 29. This was shown by scoring IBD activity in mice or by evaluating the associated wasting disease and weight loss. Here, we examined the respective capacity of CD103+ and CD103 CD25+CD4+ T cells to control wasting disease, by injecting CD25CD4+ T cells (5×105) from C57Bl/Ba (Ba) mice, alone or together with CD103+ or CD103 CD25+CD4+ T cells (1×105) from C57Bl/6 (B6) mice, into CD3ϵ KO mice. As shown in Fig. 5, mice injected with only CD25CD4+ effector T cells lost at least 23% of their weight 11 weeks after transfer, whereas mice injected with CD25 plus CD103+CD25+ cells gained weight (p<0.05). In contrast, mice injected with CD25 plus CD103CD25+ cells lost weight similarly to those injected with CD25CD4+ T cells only. These results confirm that the CD103+ regulatory T cell subpopulation preferentially controls the wasting disease associated with IBD.

Figure 5.

 CD103+CD25+CD4+ T cells control wasting disease induced by CD25CD4+ T cells. CD25CD4+ LN T cells (5×105) from Ba mice were injected alone (black squares), or together with 1×105 CD103CD25+CD4+ (gray diamonds) or CD103+CD25+CD4+ (white circles) T cells from B6 mice congenic for Thy-1 expression, into CD3ϵ KO mice. The progression of wasting disease was monitored by measuring body weight loss. The percentage of initial body weight 11 weeks post-injection is shown.

2.5 Only CD103+CD25+CD4+ T cells are able to control homeostatic expansion of CD25CD4+ T cells in vivo

We and others have shown that homeostatic proliferation and accumulation of CD25CD4+ T cells are controlled by CD25+CD4+ T cells 8, 9. The ability of CD103+CD25+CD4+ T cells to control IBD could involve a differential capacity to control T cell homeostasis. To test this hypothesis, T-cell-deficient mice were injected with Thy1.1 CD25CD4+ T cells, alone or together with Thy1.2 regulatory cells, and were studied 11 weeks later. The absolute number of cells of CD25 or CD25+ origin was scored in the spleen, peripheral LN, mesenteric LN and PP (Fig. 6). When injected alone, CD25CD4+ T cells expanded 19-fold. This expansion was not significantly affected when CD25CD4+ T cells were co-injected with CD103CD25+CD4+ regulatory cells. In contrast, co-injection of CD103+CD25+CD4+ T cells considerably reduced the expansion of CD25CD4+ T cells. These results demonstrate that only the CD103+ fraction of CD25+ regulatory T cells is capable of efficiently controlling the homeostatic expansion of CD25CD4+ effector T cells. This property is not linked to a different capacity of CD103+ and CD103 CD25+CD4+ T cells to expand after transfer, as CD103+ and CD103 cells were found in similar numbers in the adoptive hosts (Fig. 6).

Figure 6.

 CD103+CD25+CD4+ T cells control homeostatic expansion of CD25CD4+ T cells. CD25CD4+ LN T cells (5×105) from Ba mice were injected alone or together with 1×105 CD103+CD25+CD4+ or CD103CD25+CD4+ LN T cells from B6 mice, into CD3ϵ KO mice. The reconstitution of CD3ϵ KO mice 11 weeks post-injection is shown in histogram form. The absolute number of total CD25CD4+ T cells (black columns) and CD103CD25+CD4+ T cells (gray columns) or CD103+CD25+CD4+ T cells (white columns) was determined in the spleen, LN and PP. Results are given as the mean absolute number ± SEM of CD4+ T cells in the spleen plus LN plus PP, as the ratio of the different cells was identical in the different organs of each mouse.

3 Discussion

Regulatory CD4+ T cells have multiple roles in the control of immune responses. Moreover, the mechanisms they use to control these responses are heterogeneous. This wide range of activities may be mediated by individual subpopulations of regulatory T cells. Identification and isolation of such subpopulations are required to unravel the regulatory mechanisms, and to determine the therapeutic potential of regulatory cells in a variety of clinical settings.

Here, we confirmed that CD103 expression characterizes a unique population of regulatory T cells able to secrete IL-10 and to preferentially control IBD 29. We then investigated how this population exerts its unique role. We found that this regulatory cell type was highly enriched in gut-associated lymphoid tissue, where it constituted up to 40% of regulatory T cells. By their location and CD103 expression, these regulatory T cells may be particularly efficient in controlling IBD.

We also found that CD103+ regulatory T cells possess unique mechanisms regulating IL-10 secretion. The IL-10 secretion was 6-times higher when CD103+ cells were co-cultured with CD25 cells than when CD103+ cells were cultured alone. Interestingly, this IL-10 up-regulation occurred on both CD103+ cells and CD25 cells. Human CD25+CD4+ T cells induce IL-10 secretion by CD25CD4+ T cells; this phenomenon is known as infectious suppression 35, as CD25CD4+ T cells then become able to suppress naive cell responses. Our results suggest that the capacity to induce infectious suppression is a unique property of CD103+CD25+ regulatory T cells: only this subpopulation was capable of regulating IL-10 secretion and of inducing IL-10 secretion by CD25 cells. Furthermore, by regulating the IL-10 secretion, CD103+ cells would be particularly efficient in controlling IBD.

Finally, we found that CD103+ regulatory T cells can use the IL-10 they produce in autocrine manner. Their high IL-10-Rα mRNA expression together with the fact that IL-10-Rβ is constitutively expressed by most T cells 36 implies that the CD103+CD25+CD4+ T subset may need and use IL-10 to ensure its own differentiation and/or maintenance in vivo. This IL-10 requirement has previously been suggested by Groux and collaborators, who showed that CD4+ T cells with immunoregulatory properties could be generated in vitro by stimulation in the presence of IL-10 37.

About 1% of CD25CD4+ T cells express CD103 and display regulatory functions 29, and may be responsible for IL-10 production by CD25CD4+T cells in co-culture. Indeed, we found that approximately 38% of CD25CD4+ T cells that produced IL-10 in co-culture with the CD103+CD25+CD4+ T cells expressed CD103 (not shown). The origin of this CD103+CD25CD4+ T subset is unclear. We found that 20% of the CD103+CD25CD4+ T cells expressed CD25 in their cytoplasm (not shown), suggesting that CD103+CD25CD4+ and CD103+CD25+CD4+ T cells might represent different maturation stages of the same lineage. Enhanced IL-10 production by the CD103+CD25+CD4+ T cell subset may induce maturation of the CD103+CD25CD4+ T cell subset and increased its capacity to produce IL-10. However, the relationship between both subsets has to be clarified.

Despite the striking difference between CD103+ and CD103 CD25+CD4+ T cells as regards cytokine production, both populations inhibited the proliferation of CD25CD4+ T cells induced in vitro by TCR engagement 29, 33. Although inhibition of proliferation induced by TCR signaling in vitro was IL-10-independent, strong IL-10 production during interaction of regulatory T cells with effector T cells may be necessary for other functions. In particular, IL-10 is involved in the immunoregulatory function of CD4+ regulatory T cells in vivo13, 37. In addition, we have previously shown that IL-10 produced by CD25+CD4+ T cells is necessary to control the superantigen-induced burst of inflammatory cytokine production in IL-2-Rα KO mice 5. Here, we found that the IL-10-producing CD103+CD25+CD4+ T cell subset most efficiently prevented wasting disease induced by CD25CD4+ effector T cells injected into T-cell-deficient mice.

One of the most striking effects of CD25+CD4+ T cells in vivo is to regulate CD4+ T cell homeostasis 8, 9. Regulatory T cells also play a major part in the control of lymphocyte homeostasis. Lymphoid hyperplasia in IL-2-Rα KO and IL-2 KO mice is associated with a lack of CD25+CD4+ T cells, whichneed IL-2 to expand 7. CD25+CD4+ T cells from normal mice can inhibit peripheral expansion of CD4+ T cells from IL-2-Rα KO mice and restore normal homeostasis 9. We found here that control of peripheral lymphocyte numbers — a property previously attributed to all regulatory T cells — is the sole property of the CD103+ subset. It is conceivable that CD103+CD25+ T cells may protect from IBD by limiting the number of disease-causing CD25CD4+ effector T cells 38.

How then do regulatory T cells control homeostatic proliferation? It is controversial whether this homeostatic control is strictly dependent on the capacity of CD103+ T cells to produce IL-10. Indeed, in contrast to a previous report 8 suggesting that IL-10 is required for the control of both wasting disease and T cell homeostasis, it was recently shown that control of IBD was IL-10-dependent whereas homeostatic control could occur in the absence of IL-10 9. The absence of homeostasis deregulation in IL-10 KO mice (with the exception of late splenomegaly) also suggests that IL-10 is not mandatory for the control of homeostasis 39. Taken together, these results suggest that the CD103+ population, although it produces IL-10, does not control homeostasis through an IL-10-dependent mechanism. It should be noted that not all CD103+CD25+CD4+ T cells are able to secrete IL-10,implying that this population may contain subpopulations using other regulatory mechanisms. It is unlikely that this regulation is mediated by TGF-β, as we found that CD4+ populations did not express the TGF-β-receptor combinations required to respond to this cytokine. Recent results also show that CD25+CD4+ T cells can mediate their suppressor function in vitro in the absence of TGF-β production and responsiveness 15.

These results support the notion that CD25+CD4+ regulatory T cells are not a homogeneous population but that they contain subtypes with different functional characteristics. We found that although the CD103+ and CD103 CD25+CD4+ T subpopulations both inhibited CD25 effector T cell proliferation in vitro independently of IL-10, their functional capabilities and mechanisms of action were different. Only CD103+CD25+CD4+ T cells produced IL-10, expressed IL-10-Rα mRNA, strongly up-regulated their IL-10 production in co-culture with CD25CD4+ T cells, and induced IL-10 secretion by a subset of effector T cells through an IL-10-dependent mechanism. These properties may be involved in the anti-inflammatory function of these cells in vivo. Indeed, CD103+CD25+CD4+ T cells are preferentially located in gut-associated lymph nodes and concentrated the capacity to control wasting disease and effector T cell expansion. These results underline the heterogeneity of regulatory T cells and of their regulatory mechanisms. The relationship between the CD103+CD25+CD4+ and CD103CD25+CD4+ T cell subsets also requires further investigation. These two subsets might be developmentally distinct and derive from different precursors. Alternatively, they may have a precursor-progeny relationship, and be cells at different maturation stages of the same lineage.

4 Materials and methods

4.1 Mice

B6 mice (H-2b, Thy 1.2) mice were from CERJ (Le Genest St Isles, France), IL-10 KO 39 mice were from Transgenic Alliance (L'arbresle, France), CD3ϵ KO mice40 were from CDTA-CNRS (Orléans, France) and Ba mice (H-2b, Thy 1.1) were from our own facilities.

4.2 mAb and flow cytometry

The following Ab were used for flow cytometry and cell culture. Purified anti-CD8 and anti-CD3, biotinylated anti-CD25 and anti-CD45RB, FITC–anti-CD25, FITC–anti-Thy-1.1 and FITC–anti-CD8 wereprepared and coupled in our laboratory. Biotinylated anti-CD103, PerCP–anti-CD4, FITC–anti-CD103, PE–anti-CD44, PE–anti-IL-10, PE–rat-IgG2b, PE–anti-CTLA-4, PE–anti-CD4, PE–anti-Thy-1.2 and PE–anti-CD25 were from Pharmingen (San Diego, CA, USA). Biotinylated antibodies were revealed with streptavidin–APC (Pharmingen). Flow cytometry was performed using a FACS Calibur device (BD Biosciences, Mountain View, CA, USA).

4.3 T cell purification

Lymphocyte suspensions were prepared from pooled peripheral LN (inguinal, axillary, cervical and popliteal). CD4+ T cells were prepared by negative selection, and CD25+CD4+ T cells were then positively selected with a Macs Separator (Miltenyl Biotech, Paris, France) as previously described 34. The CD25+CD4+ and CD25CD4+ populations were 90% and 95% pure, respectively. Purified CD25+CD4+ and CD25CD4+ T cells were labeled with FITC–anti-CD103, PE–anti-CD4 and Cychrome–streptavidin (BD Biosciences); CD103+CD25+CD4+ and CD103CD25+CD4+ T cells were then sorted using a FACS Vantage (BD Biosciences). Cells were97–99% pure. For mRNA analysis, CD25CD4+ T cells were sorted by FACS after negative selection. In some experiments, CD103+ and CD103CD25+CD4+T cells were positively selected using a Macs Separator (Miltenyl Biotech).

4.4 Real-time quantitative RT-PCR

Expression of mRNA for IL-2, IL-4, IL-10, TGF-β, IFN-γ, TNF-α, TNF-β, IL-2-Rα, IL-10-Rα, TGF-β-R1, TGF-β-R2, TGF-β-R3 and HPRT was tested on purified CD103+CD25+CD4+, CD103CD25+CD4+ and CD25CD4+ lymph node T cells ex vivo, before and after 4 h of anti-CD3 stimulation. RNA was extracted from 106 cells using the Rneasy Mini Kit (Qiagen, Courtabouef, France) according to the manufacturer's instructions. RNA was reversed-transcribed into cDNA by using random oligonucleotides from Gene Amp RNA PCR Kit Components (Applied Biosystems, Foster City, CA, USA). The cDNA was quantified by SYBR-Green incorporation (PCR Master Mix; Applied Biosystems): SYBR-Green dye binding to double-stranded DNA permits direct detection of PCR products after each amplification cycle (ABI Prism 7700 Sequence Detection System). The specificity of the amplification was checked by electrophoresis. PCR reactions were run according to the manufacturer's instructions using 0.25 pmol of primers in MicroAmp Optical 96-well reaction plates (Applied Biosystems). The cDNA were subjected to 60 cycles at 95°C for 30 sec, 60°C for 30 sec and 72°C for 30 sec. Primers for each gene of interest were designed for use in real-time PCR conditions. The efficiency of PCR amplification was the same for all the genes studied. A standard curve of cDNA from increasing numbers of cells was used to calibrate differences in cycle numbers. For each gene and all three cell populations, Ct was determined in the linear phase of PCR. To calculate relative copy numbers, Ct was corrected for HPRT cDNA amplification, which was studied simultaneously. The difference in Ct between two populations was designated n, and the difference in mRNA quantity was calculated as 2n.

4.5 In vitro activation, proliferation and cytokine assays

Purified CD103+CD25+CD4+, CD103CD25+CD4+ and CD25CD4+ LN T cells (1×105) were cultured for 4 h at 37°C in 5% CO2-in-air, in anti-CD3-coated 96-well culture plates, to study cytokine mRNA production. To examine regulation of cell proliferation and IL-10 production, CD25CD4+ T cells (1×105) were cultured for 48 h in anti-CD3-coated wells alone or with CD103+CD25+CD4+, CD103CD25+CD4+ or CD25+CD4+ T cells (1×105) or with added CD25CD4+ T cells. All cells were grown in 100 μl of RPMI 1640 medium supplemented with 10% FCS (Life Technologies, Grand Island, NY, USA), 100 U/ml penicillin, 100 μg/ml streptomycin, 1% sodium pyruvate (Biomedia, Foster City, CA, USA), 5 mM Hepes (Life Technologies), 10–5 M β2-mercaptoethanol and 50 ng/ml PMA. Proliferation was assessed by measuring incorporation of [3H]thymidine (specific activity 25 Ci/mmol; Amersham France SA, Les Ulis, France) during the last 4 h of culture. Results are expressed as counts per minute. IL-10 was assayed in the supernatants with ELISA kits (R&D Systems, Abingdon, GB) and by intracellular IL-10 staining. To analyze IL-10 expression in the different subpopulations, cells were restimulated for 4 h with ionomycin (500 ng/ml) in the presence of GolgiPlug (BD Biosciences). Before fixation and intracellular staining, cells were stained with anti-CD4, -Thy-1.1 and -CD103 to distinguish CD103+ and CD103CD25+CD4+ T cells (B6, Thy-1.2) from CD25CD4+ T cells (Ba, Thy-1.1).

4.6 Cell transfer

CD3ϵ KO mice were injected i.v. with 5×105 CD25CD4+ LN T cells from Ba mice, alone or with 1×105 sorted CD103+ or CD103 CD25+CD4+ T cells from B6 mice; the latter were congenic for Thy-1 expression, which was used to follow donor cell reconstitution of CD3ϵ KO mice. The progression of wasting disease was monitored by measuring body weight loss. Eleven weeks post-injection, LN, spleen and PP were removed and cell suspensions were prepared. Absolute numbers of total CD25CD4+T cells, CD103+CD25+CD4+ and CD103CD25+CD4+ T cells were determined by cell counting and by anti-CD4 and anti-Thy-1 labeling.

4.7 Statistical analysis

Data are expressed as means±SEM. Comparisons were made using Student's t-test; differences were considered significant at a 95% confidence level.

Acknowledgements

We thank C. Tanchot and C. Bourgeois for help with intracellular cytokine staining. This work was supported by Institut National de la Santé et de la Recherche Médicale (INSERM) and Université René Descartes Paris V. A. Banz is supported by a grant from the Association de la Recherche contre le Cancer (ARC).

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

  6. 6

    WILEY-VCH

Ancillary