Met 1 Dad

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES

Regulatory T cells expressing interleukin 10 develop from Foxp3+ and Foxp3– precursor cells in the absence of interleukin 10 Craig L Maynard1,2, Laurie E Harrington1, Karen M Janowski1, James R Oliver1, Carlene L Zindl2, Alexander Y Rudensky3 & Casey T Weaver1,2 CD4+ regulatory T cells (Treg cells) that produce interleukin 10 (IL-10) are important contributors to immune homeostasis. We generated mice with a ‘dual-reporter’ system of the genes encoding IL-10 and the transcription factor Foxp3 to track Treg subsets based on coordinate or differential expression of these genes. Secondary lymphoid tissues, lung and liver had enrichment of Foxp3+IL-10– Treg cells, whereas the large and small intestine had enrichment of Foxp3+IL-10+ and Foxp3–IL-10+ Treg cells, respectively. Although negative for Il10 expression, both Foxp3+ and Foxp3– CD4+ thymic precursor cells gave rise to peripheral IL-10+ Treg cells, with only Foxp3– precursor cells giving rise to all Treg subsets. Each Treg subset developed in IL-10-deficient mice, but this was blocked by treatment with antibody to transforming growth factor-b. Thus, Foxp3+ and Foxp3– precursor cells give rise to peripheral IL-10-expressing Treg cells by a mechanism dependent on transforming growth factor-b and independent of IL-10.

Since its discovery as a product of T helper type 2 (TH2) cells that inhibit effector cytokine production by TH1 cells1, interleukin 10 (IL-10; originally called CSIF, for ‘cytokine synthesis inhibitory factor’) has emerged as a key immunomodulatory factor that inhibits the release of proinflammatory cytokines by innate immune cells2. It is now recognized that IL-10 can be produced by a range of adaptive and innate immune cells3, although its production by CD4+ T cells seems particularly important for immune homeostasis4. Mice deficient in IL-10 develop spontaneous, lethal inflammation of the lower intestine driven by unimpeded reactivity of effector CD4+ T cells to antigens of the intestinal microbiota5,6. Adoptive transfer of IL-10-deficient CD4+ T cells into mice with severe combined immunodeficiency or mice deficient in recombination-activating gene products induces severe colitis despite the ability of the recipient’s innate immune system to produce IL-107,8, and mice deficient in IL-10 limited to CD4+ T cells similarly develop lethal, spontaneous colitis9. Such results establish a nonredundant function for IL-10-producing CD4+ T cells in controlling pathologic autoreactivity to the intestinal flora10. The production of IL-10 by CD4+ T cells has been associated with at least two general subsets of regulatory T cells (Treg cells), natural and adaptive (or induced), that seem to differ in their development, antigenic specificities and mechanisms of action. Expression of the transcription factor Foxp3 is a defining feature of natural Treg cells11, which arise during thymic development through high-affinity recognition of self antigens12,13. Foxp3 initiates and supports a program of

gene expression that specifies the development and maintenance of this lineage14–16. The phenotypic characteristics of natural Treg cells include constitutive expression of IL-2 receptor-a (CD25), the T cell activation marker CTLA-4 and the cell survival factor GITR, which seem to be necessary for the maintenance and/or suppressive function of these cells17–19. The suppressive function of natural Treg cells seems to require only cell-cell contact or proximity in vitro, whereas aspects of the in vivo function of natural Treg cells depend on the secretion of IL-10 and transforming growth factor-b (TGF-b)10. Adaptive Treg cells include Foxp3+ cells that develop extrathymically and share most phenotypic and functional features of natural Treg cells, as well as Foxp3– cells that seem to exert their regulatory activity mainly by means of secreted cytokines such as IL-10 or TGF-b20. The characterization of Foxp3– Treg cells has been limited because of the absence of specific, defining markers for this population; hence, the diversity, developmental origins and functions of these cells are poorly understood. Perhaps best characterized are T regulatory type 1 cells (Tr1 cells), originally described as a product of naive CD4+ T cells activated ex vivo in the presence of IL-10 or by IL-10-conditioned dendritic cells21,22. Tr1 cells are characterized by abundant production of IL-10 and produce suppression by a cell contact–independent, cytokine-dependent mechanism that involves both IL-10 and TGF-b23. On the basis of the absence of Foxp3 expression by Tr1 cells and their derivation from naive CD4+CD25– T cell precursor cells21,24, it has been suggested that, at least in vitro, Tr1 cells develop

1Department of Pathology and 2Department of Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294, USA. 3Howard Hughes Medical Institute, University of Washington School of Medicine, Seattle, Washington 98195, USA. Correspondence should be addressed to C.T.W. ([email protected]).

Received 30 April; accepted 20 July; published online 12 August 2007; doi:10.1038/ni1504

NATURE IMMUNOLOGY

VOLUME 8

NUMBER 9

SEPTEMBER 2007

931

independently of natural Treg cells25. However, lineage relationships between Tr1 cells and natural Treg cells or IL-10-producing Foxp3+ adaptive Treg cells have not been defined. To better define the cell populations that produce IL-10 in vivo and to facilitate studies of their development and function, we generated mice transgenic for a bacterial artificial chromosome (BAC) containing an Il10 gene in which expression of IL-10 was replaced with expression of a Thy-1.1 reporter, thereby enabling the identification of cells ‘programmed’ for expression of IL-10 (called ‘IL-10 competent’ cells here) that arise spontaneously in vivo. We crossed those mice with reporter mice with the sequence encoding green fluorescent protein (GFP) ‘knocked-into’ the gene encoding Foxp3 (Foxp3gfp mice)26 to produce ‘dual-reporter’ mice that permit simultaneous detection of Il10 and Foxp3 expression by individual cells. Characterization of the dual-reporter mice demonstrated that in the steady state in vivo, coordinate or differential expression of Il10 and Foxp3 by CD4+ T cells defined three distinct subsets of Treg cells whose prevalence differed in distinct tissue compartments. Foxp3+IL-10– Treg cells were more frequent in lymph nodes and spleen and developed from Foxp3+ or Foxp3– thymic precursor cells. Foxp3+IL-10+ Treg cells, in contrast, were more frequent in the lymphoid tissues of the large intestine and similarly arose from either Foxp3+ or Foxp3– precursor cells. In the small intestine and Peyer’s patches, Foxp3–IL-10+ Treg cells were the most prevalent and had a cytokine profile, proliferative response and suppressive function typical of Tr1 cells. This last population

a

Spleen <0.1

PLN <0.1

CLN <0.1

Liver <0.1

developed exclusively from Foxp3– precursor cells in all tissues, and, as with the IL-10-competent Foxp3+ subsets, this development was dependent on TGF-b but not IL-10. Our data establish that Treg cells able to express IL-10 develop in vivo by Foxp3-dependent and Foxp3-independent pathways, that gut-associated lymphoid tissue (GALT) is enriched for both Foxp3+IL-10+ and Foxp3–IL-10+ subsets, and that both are dependent on TGF-b for their induction and/or maintenance. RESULTS IL-10 reporter mice To generate an IL-10 reporter mouse that permits identification of Il10 expression without perturbing endogenous loci, we cloned a Thy-1.1 (CD90.1) expression cassette in-frame into the first coding exon of Il10 in a BAC including about 113 kb and 83 kb of contiguous 5¢ and 3¢ genomic DNA, respectively (Supplementary Fig. 1 online). We designed the reporter cassette to replace the endogenous coding segment of exon 1 and to introduce a premature stop codon, thus abrogating downstream Il10 expression but leaving the remainder of the locus intact to ensure normal gene regulation (Supplementary Fig. 1a). We injected this ‘Il10 BAC-in transgene’ (10BiT) DNA into single-cell embryos from C57BL/6 mice to generate founder lines on a defined strain background. We analyzed in detail one founder line containing approximately 12 integrated copies of 10BiT DNA and used this line for all subsequent studies (Supplementary Figs. 1 and 2

b

Lung <0.1

c 103

CD8αα+CD4+ CD8αβ+CD4+

B6

2

3.6 10BiT

MLN <0.1

PP <0.1

SI IEL <0.1

SI LPL <0.1

22.7

90

<0.1

13.2

28.9 10BiT

85

P C SI P ol LP on L LP L

N C LN M LN

PL

28.8

27

TCRβ

102

3.2

6

67

CD8β

B6

2.1

89

CD8αα+CD4– CD8αβ+CD4–

CD8α

Colon LPL

3 2 71 24

TCRγδ

1.2

1

CD8α+Thy-1.1+

CD4

1.7

Thy-1.1

1.8

CD8α+

Thy-1.1 (MFI)

1.9

Thy-1.1

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES

d

WT

10BiT

CD4

CD4

Thy-1.1

Small Figure 1 Intestines and associated lymphoid tissues are enriched for intestine IL-10-competent Thy-1.1+CD4+ T cells. (a) Flow cytometry of CD4+ Hoechst T cells from 6-week-old 10BiT mice and wild-type littermates stained Villin with monoclonal antibodies specific for CD4 and Thy-1.1. Plots are gated on CD4+ T cells; numbers beside outlined areas indicate the frequency of Thy-1.1+ cells. PLN, peripheral lymph node; CLN, cervical lymph node; MLN, mesenteric lymph node; PP, Peyer’s patch; SI, small intestine; B6, C57BL/6. (b) Mean fluorescence intensity CD4 (MFI) of Thy-1.1 expression by CD4+ T cells in non-GALT (white bars), Thy-1.1 GALT (gray bars) and intestinal lamina propria (black bars) for data Colon acquired as described in a. (c) Flow cytometry of IELs from the small Hoechst intestines of 10BiT mice, labeled with fluorescent antibodies specific for CD4, CD8a, CD8b, TCRb, TCRgd and Thy-1.1. Far left, gating on total CD8a+ IELs; other plots gated as indicated by colors. Numbers in plots indicate percent cells. (d) Immunofluorescence staining of intestinal tissue for CD4, Thy-1.1 and villin and/or cell nuclei (Hoechst). Solid white arrows indicate cells positive for Thy-1.1 and CD4 staining in the lamina propria of the small intestine, lymphoid aggregates of the colon or lamina propria of the colon (red-bordered arrow); open arrow identifies a CD4–Thy-1.1+ cell in the epithelium of the small intestine. WT, wild-type. Original magnification, 40. Data are representative of four (a,c), or three (d) experiments or are pooled from two experiments with six mice (b).

932

VOLUME 8

NUMBER 9

SEPTEMBER 2007

NATURE IMMUNOLOGY

ARTICLES uniformly higher expression of the Thy-1.1 reporter (Fig. 1a,b). This pattern of Thy-1.1 750 75 375 expression seemed restricted to the intestines, 500 50 250 as there were not similar frequencies of 250 25 125 Thy-1.1+CD4+ T cells in other nonlymphoid tissues such as liver and lung (Fig. 1a). Thus, 0 0 0 U S U S U S U S U S U S Thy-1.1– IL-10-competent Thy-1.1+CD4+ T cells were – + – + – + 1,000 500 present in all immune compartments exam– Thy-1.1 (unstim) ined, but the intestines were particularly 750 375 – Thy-1.1 (stim) enriched for these cells. 500 250 Thy-1.1+ (unstim) Thy-1.1 expression in the IEL compartment + 250 125 Thy-1.1 (stim) of the small intestine was not restricted to 0 0 CD4 T cell receptor-ab (TCRab) CD4+ T cells U S U S U S U S – + – + (Fig. 1c). Over one fourth of the CD8a+ + – + T cells in the small intestinal epithelium Figure 2 Cytokine phenotypes of Thy-1.1 and Thy-1.1 CD4 T cell subpopulations. Cytometric bead + – (IEL compartment) were also Thy-1.1+, array analysis of cytokines produced by 10BiT CD4 T cells sorted by flow cytometry into Thy-1.1 (–) + 5 and Thy-1.1 (+) fractions (1  10 cells each) and then left unstimulated (unstim; U) or stimulated with representation of Thy-1.1 expression for 18 h with immobilized anti-CD3 and soluble anti-CD28 (stim; S). Data (mean + s.e.m. of among all CD8+ IEL subsets: a fraction of triplicates) are from one of three experiments with similar results. CD8aa+CD4–, CD8ab+CD4–, CD8ab+CD4+ and CD8aa+CD4+ T cells expressed Thy-1.1. + + online). Mice from this line (10BiT mice) developed normally, were Most CD8 Thy-1.1 IELs of the small intestines were CD8aa+CD4– fertile and had normal survival and no immunological or histopatho- (71%), of which approximately two thirds expressed gd TCRs. Thus, logical abnormalities as late as 1.5 years of life (data not shown). expression of IL-10 by IELs of the small intestine was not restricted to To ensure the fidelity of expression of the 10BiT transgenes, we ab T cells (Fig. 1c). Detection of Thy-1.1+ T cells by immunofluorproduced TH2-polarized populations ex vivo from naive CD4+ T cells escence staining of frozen sections of small intestine from 10BiT mice enriched from 10BiT and wild-type mice and then analyzed the confirmed the high frequency of Thy-1.1+ (IL-10-competent) IELs of correlation of expression of Thy-1.1 and IL-10 by flow cytometry the small intestine in situ (Fig. 1d). (Supplementary Fig. 1b). The frequency of TH2 cells expressing both To assess the cytokine phenotype of 10BiT T cells identified in vivo, Thy-1.1 and IL-10 was high; all IL-10+ TH2 cells were Thy-1.1+, and we pooled CD4+ T cells isolated from GALT and other lymphoid the high sensitivity of detection of Thy-1.1 relative to that of tissues and fractionated them into Thy-1.1+ and Thy-1.1– subsets by intracellular IL-10 suggested that there was probably a fraction of flow cytometry, then polyclonally stimulated these cells ex vivo or, 10BiT T cells that were falsely negative for IL-10 expression. There as a control, a subset was left unstimulated (Fig. 2). Unstimulated seemed to be no ‘aberrant’ expression of Thy-1.1 in cells that were CD4+Thy-1.1+ cells but not CD4+Thy-1.1– cells had low production of IL-10–, either in vitro or in vivo (data not shown), and the Thy-1.1 IL-10 without stimulation ex vivo. IL-10 production by the transgenic reporter transgene did not seem to perturb expression of the endo- CD4+Thy-1.1+ cells increased substantially after stimulation, indicatgenous Il10 alleles. Hence, the frequencies of IL-4+ and IL-10+ cells ing that Thy-1.1 expression ‘marks’ cells that are IL-10 competent but were similar for TH2 cells derived from wild-type or 10BiT mice (IL-4, not actively secreting IL-10 as well as those that are actively producing 45% or 39%, respectively; IL-10, 78% or 80%, respectively). Expres- IL-10 (Fig. 2 and data not shown). The concentration of IL-10 sion of the 10BiT transgene therefore ‘marks’ IL-10-expressing T cells produced by activated Thy-1.1+ cells was substantially greater than with high sensitivity and fidelity. the concentration of other cytokines examined (over 15 times the concentration of interferon-g (IFN-g), and over 40 times the producIntestinal tissues are enriched for IL-10-expressing CD4+ T cells tion of IL-2 and tumor necrosis factor), consistent with production of To define IL-10-expressing cells in vivo, we examined expression of the IL-10 mainly by Treg lineages and not by effector T cell subsets. Thy-1.1 transgene in the lymphoid tissues of unmanipulated 10BiT Additionally, intracellular cytokine staining showed that the IL-10mice (Fig. 1). In contrast to littermate control mice, 10BiT mice had competent CD4+ T cells isolated from the lamina propria of both the Thy-1.1+CD4+ T cells in all secondary lymphoid tissues examined; a small and large intestines of 10BiT mice lacked expression of effector small minority (1–2%) of reporter-positive CD4+ T cells were present cytokines such as IFN-g and IL-17 (data not shown and discussed in spleen, in peripheral lymph nodes (inguinal and axillary) and in below). These data collectively establish that expression of the cervical lymph nodes. The 10BiT mice also had a modestly but 10BiT reporter gene identifies IL-10-expressing and IL-10-competent consistently higher frequency of Thy-1.1+CD4+ T cells in mesenteric (Thy-1.1+) cells in situ, and they also establish a primary link to lymph nodes and Peyer’s patches (2–4% of CD4+ T cells; Fig. 1a and noneffector lineages in the steady state. data not shown). In contrast to results obtained with secondary lymphoid tissues, Expression of IL-10 and Foxp3 defines Treg cell subsets Thy-1.1+CD4+ T cells were prominent in intestinal tissues, including a Treg cells have been subcategorized into at least two subsets on the high frequency of cells in the intraepithelial lymphocyte (IEL) com- basis of Foxp3 expression: Foxp3+ Treg cells that develop intrathymipartment of the small intestine and in the lamina propria lymphocyte cally (natural Treg cells) or extrathymically (adaptive Treg cells) and (LPL) compartments of the small and large intestines (Fig. 1). More express IL-10, which seems essential for some but not all of their than 10% of all CD4+ T cells in the LPL compartment of the small suppressive activities; and Foxp3– Treg cells with high expression of intestine and about twice that frequency in the LPL compartment of IL-10, referred to as ‘Tr1’ cells20. By examining the expression of Foxp3 the colon were Thy-1.1+ (and therefore IL-10 competent); a large and Thy-1.1 in CD4+ T cells from transgenic mice, we defined three fraction of the Thy-1.1+ cells in the colonic LPL compartment had subsets of Treg cells: Foxp3+Thy-1.1– cells, Foxp3+Thy-1.1+ cells and IL-4 (pg/ml)

100

500

NATURE IMMUNOLOGY

VOLUME 8

NUMBER 9

IL-2 (pg/ml)

TNF (pg/ml)

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

Thy-1.1

IL-10 (pg/ml)

1,000

IFN-γ (pg/ml)

Thy-1.1+

SEPTEMBER 2007

933

ARTICLES

a

PLN 0.7

CLN 1.0

0.6

1.3

9.8

Thy-1.1

4.4

1.3

26.6

SI LPL 2.0

1.7

1.0

12.6

SI IEL 0.9

8.3

b

MLN 1.0

9.8 Colon LPL

5.0

3.6

4.3

24.4

10.4

Foxp3

CD25

Foxp3– Thy-1.1–

Foxp3+ Thy-1.1–

Foxp3– Thy-1.1+

Foxp3+ Thy-1.1+

c

CD45RB

20

CD62L 83

61

CTLA-4

82

64

33

55

29

41

23

89

23

M1

23

92

10BiT.Foxp3 gfp

97

27

76

61

M1

38 M1

8

M1

M1

16

M1

16

69

ICOS

34 M1

M1

M1

M1

GITR 7

M1

M1

M1

CD103

2

M1

M1

96

66

M1

GFP– Thy-1.1–

GFP– Thy-1.1+

GFP + Thy-1.1–

GFP + Thy-1.1+

103 c.p.m.

70 Figure 3 Phenotypic and functional analysis of Treg subsets defined by 60 concordant or differential expression of Il10 and Foxp3. (a) Flow cytometry 50 of single-cell suspensions from various tissues of 10BiT mice, stained with 40 monoclonal antibodies to CD4, Thy-1.1 and Foxp3. Plots are gated on 30 + CD4 T cells; numbers in quadrants indicate percent cells in each. (b) Flow 20 cytometry of total CD4+ T cells from pooled peripheral lymphoid tissues, 10 GFP stained with monoclonal antibodies to CD4, Thy-1.1, Foxp3 and cell 0 surface markers (above plots). All plots are gated on CD4+ T cells with – α-IL-10R α-TGFβ α-IL-10R + α-TGFβ secondary gating on specific subsets (left margin). Numbers in plots Coculture indicate percent cells positive above unstained controls (shaded histograms). ICOS, inducible costimulator. (c) Flow cytometry (left) and proliferation assay (right) of pooled CD4+ T cells from the secondary lymphoid tissues of 10BiT.Foxp3gfp mice, sorted into four subsets by flow cytometry based on expression of Thy-1.1 (Il10) and GFP (Foxp3). For proliferation analysis, each of the four sorted subsets was cultured in triplicate alone (4  104 cells/well; far left) or an equal number of GFP–Thy-1.1– cells and cells of the remaining three subsets were combined (Coculture), followed by stimulation with splenic antigen-presenting cells (3  105 cells/well) and anti-CD3 (2.5 mg/ml), with [3H]thymidine added for the final 18 h of the 72-hour incubation. Blocking antibody to the IL-10 receptor (a-IL-10R) and/or TGF-b (a-TGF-b; 10 mg/ml each) was also added to some wells for the duration of the stimulation period. Data are representative of four (a) or three (b) experiments or are the mean + s.d. of one of three similar experiments (c).

Thy-1.1

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

0.5

11.3

PP 2.3

Spleen

Foxp3–Thy-1.1+ cells (Fig. 3). Although they were present in every tissue examined, their frequencies differed depending on the tissue compartment (Fig. 3). In all lymph nodes and spleen, the most common Treg cells were Foxp3+Thy-1.1–, representing approximately 10% of the total CD4+ T cell population (Fig. 3a). Of the Foxp3+CD4+ T cells in these sites, only a minor fraction (about 10%) were Thy-1.1+ and thus IL-10 competent. We also found Foxp3–Thy-1.1+ Tr1-like cells in the lymph nodes and the spleen, but these cells represented 1% or less of the total CD4+ T cells and were, therefore, about one tenth as frequent as Foxp3+ cells in these tissues. Thus, IL-10-competent Treg cells are relatively rare in peripheral lymphoid tissues, and most are in the Foxp3+ Treg cell fraction. There was a substantial shift in subset distribution among Treg cells in GALT. Peyer’s patches had substantial enrichment for Tr1-like cells, which represented approximately one third of all CD4+ T cells with a Treg phenotype there. Tr1-like cells were even more prevalent in the LPL compartment of the small intestine, where approximately one half of the Treg cells had a Tr1 phenotype (Fig. 3a, bottom row). GALT of the small intestine also had enrichment for IL-10-competent Thy1.1+Foxp3+ Treg cells, which represented most Foxp3+CD4+ T cells in the LPL and IEL compartments (58% and 54%, respectively). Tr1-like cells were particularly dominant in the epithelial compartment, where they composed over 85% of CD4+ T cells with a Treg phenotype. In contrast to the small intestine, colonic GALT was relatively devoid of Tr1-type cells. However, Treg cells constituted more than one third of the total CD4+ T cell population of the lamina propria of the colon (39.1%), despite the relative paucity of Tr1-like cells. Thus, nearly all Treg cells isolated from the colon were Foxp3+ cells (89%), most of which (70%) were IL-10 competent (Fig. 3a). Notably, many of the

934

Foxp3+Thy-1.1+CD4+ T cells were in colonic lymphoid aggregates (Fig. 1d), although we also identified scattered Foxp3+Thy-1.1+CD4+ T cells in the lamina propria of the colon. Although the lamina propria of the colon and small intestine differed in terms of the prominence of Tr1-like population, in both sites most of the Foxp3+ Treg cells were IL-10 competent. These data support the views that in both the small and large intestine, the function of CD4+Foxp3+ cells in immune regulation is mainly IL-10 dependent and that in the absence of IL-10 expression, Foxp3 is inadequate for the maintenance of intestinal immune homeostasis. To test those ideas directly, we examined the requirement for IL-10competent Treg cells in suppressing the development of colitis in a ‘CD45RB transfer’ model of colitis27. Because the Thy-1.1 reporter accurately identifies IL-10-competent cells and those cells can be specifically depleted with a monoclonal antibody to Thy-1.1, we were able to determine whether specific elimination of IL-10-competent CD4+ T cells in vivo compromised the ability of Treg cells to prevent the development of colitis when transferred together with non–Treg cells (Supplementary Fig. 3 online). Treatment with antibody to Thy-1.1 (anti-Thy-1.1) resulted in the depletion of more than 90% of Thy-1.1+ Treg cells as well as blockade of the suppression of colitis development comparable to that of mice treated with blocking antibody to the IL-10 receptor (anti-IL-10R), despite sparing a large fraction of Foxp3+IL-10– Treg cells. These data support the results of published studies indicating that IL-10 production by Foxp3+ Treg cells is essential for their ability to suppress inflammatory responses to the intestinal microbiota28 and further demonstrate that expression of Thy-1.1 identifies IL-10-competent cells in vivo. To further characterize the Treg subsets identified, we phenotyped CD4+ T cells by surface expression (Fig. 3b). Most Foxp3+ and

VOLUME 8

NUMBER 9

SEPTEMBER 2007

NATURE IMMUNOLOGY

ARTICLES

a

CD8 SP

DP

Thy-1.1

DN

0.3

0.9

CD4 SP

0.5

5.2

Foxp3 CD8α depletion

GFP +

Analysis

97.6 10BiT.Foxp3 gfp (thymus)

IV

B6(CD45.1)

Group A

B6(CD45.1)

Group B

GFP – GFP

IV 99.8 CD4

Spleen Group A

Group B

MLN 17

1.2

Thy-1.1

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

b

0.9

4.6

PP 27

0.7

0.5

3.1

SI IEL 43

4.7

1.0

3.6

10.9

0.3

66

4.2

0.3

GFP

IL-10-competent CD4+ T cells displayed a surface phenotype characteristic of activated or memory T cells, regardless of Treg subset. Thus, in contrast to Foxp3–Thy-1.1– non–Treg cells, most Treg cells were CD45RBlo (Fig. 3b). Compared with Foxp3– Thy-1.1– cells, each of the Treg subsets had higher expression of CTLA-4, which was even more highly expressed by Thy-1.1+ cells; similarly, each subset also had higher expression of GITR and IL-2 receptor-a (CD25), although these markers were more highly expressed on Foxp3+ cells whether they were IL-10 competent or not. Expression of L-selectin (CD62L) and aE integrin (CD103) by Treg subsets was bimodal, with the relative frequencies of low versus high expression of each reflecting the relative partitioning of Treg subsets in intestinal versus nonintestinal lymphoid tissues (Fig. 3b and data not shown). Both subsets of IL-10-competent cells (Foxp3–Thy-1.1+ and Foxp3+Thy-1.1+ cells) had higher expression of inducible costimulator (ICOS) than did Foxp3–Thy-1.1– cells, consistent with the reported relationship between high expression of ICOS and predisposition toward high expression of IL-10 in the steady state29. The surface phenotype of each of the subsets defined by Foxp3 and/or Thy-1.1 (IL-10) expression was consistent with the reported characteristics of Treg subsets, thus establishing that those markers are common to all Treg cells that develop spontaneously in vivo. A unique feature of 10BiT mice was that they enabled the identification of IL-10-competent cells that developed spontaneously in vivo, permitting the isolation and functional assessment of such cells without a requirement for in vivo or ex vivo activation. To take further advantage of this property, we crossed 10BiT mice with Foxp3gfp mice26. We evaluated the expression of both reporter transgenes in the resultant dual-reporter ‘10BiT.Foxp3gfp’ mice to confirm the identification of the Treg subsets characterized above (Fig. 3c, left, and data not shown). We next evaluated GFP+Thy-1.1–, GFP+Thy-1.1+ and GFP–Thy-1.1+ cells from the 10BiT.Foxp3gfp mice

NATURE IMMUNOLOGY

VOLUME 8

NUMBER 9

Colon LPL

SI LPL 73

SEPTEMBER 2007

65

Figure 4 IL-10-competent Treg cells develop extrathymically from both Foxp3+ and Foxp3– thymic precursor cells. (a) Flow cytometry of 10BiT thymocytes stained with monoclonal antibodies specific for CD4, CD8, Thy-1.1 and Foxp3. Cells are gated on thymocyte subsets (above plots); numbers indicate percent Foxp3+ cells. DN, double negative; DP, double positive; SP, single positive. Data are representative of five experiments. (b) Flow cytometry of 10BiT.Foxp3 gfp thymocyte samples depleted of CD8+ cells before being sorted by flow cytometry into CD45.2+CD4+GFP– and CD45.2+CD4+GFP+ fractions (top left). Top right, intravenous (IV) transfer of CD4+GFP– cells (3  106 to 5  106) or CD4+GFP+ cells (2  105 to 5  105) into sublethally irradiated age- and sex-matched CD45.1 recipients. Below, expression of GFP (Foxp3) and Thy-1.1 (Il10) by donor CD45.1+ cells at 4 weeks after transfer; plots are gated on CD4+CD45.2+ cells. Numbers in quadrants indicate percent cells in each. Data are from one of three similar experiments with three to five recipients per group.

to determine the relative proliferative capacity and suppressor activity of each Treg subset ex vivo (Fig. 3c, right). We isolated and pooled CD4+ T cells from GALT and extra2.4 2.6 intestinal lymphoid tissue, sorted the cells by flow cytometry into four fractions defined by expression of GFP (Foxp3 marker) and Thy-1.1 (IL-10 marker), and then tested the cells for proliferation and suppressive function. Compared with control non–Treg cells (GFP–Thy-1.1–CD4+ T cells), all the Treg subsets demonstrated profound anergy (o10% proliferation of controls). GFP+Thy-1.1+ and GFP+Thy-1.1– Treg cells much more potently inhibited responder cells (GFP–Thy-1.1–) than did GFP–Thy-1.1+ Tr1 cells (over 95% inhibition and about 60% inhibition, respectively), yet we found no substantial difference in the suppressor potency of the GFP+Thy-1.1– and GFP+Thy-1.1+ subsets. In accordance with published studies using Tr1 cells derived in vitro21,30, the inhibitory activity of GFP–Thy-1.1+ cells was partially reversed by the addition of anti-IL-10 or anti-TGF-b and was completely reversed by the addition of both antibodies. These data establish that an in vivo correlate of the functional phenotype previously described for Tr1 cells is enriched in intestinal tissues and can act by cytokine-dependent mechanisms to exert suppressive function31. There was no similar inhibition of suppression of either GFP+ (Foxp3-competent) subset by anti-IL-10 and anti-TGF-b, although there was modest inhibition of suppression of IL-10-competent, Foxp3-competent, Thy-1.1+GFP+ Treg cells in some experiments. These data are consistent with a contact- or proximity-dependent mode of suppression by both Foxp3+ Treg subsets in blocking the proliferation of naive T cells and indicate that acquisition of Il10 expression by Foxp3+ Treg cells does not alter their dominant mode of suppressive activity in vitro. 1.1

1.3

2.9

IL-10-competent Treg cells from Foxp3+ and Foxp3– precursors CD4+Foxp3+ Treg cells, including both those that develop in the thymus during repertoire selection and those that develop extrathymically from naive CD4+ T cell precursors, have been shown to express IL-10, although details of the frequency and origin of IL-10-competent Foxp3+ Treg cells are limited. The origins of IL-10competent Foxp3– Treg cells, such as Tr1 cells, have been difficult to

935

ARTICLES

a

c

10BiT.II10 –/–

Spleen 0.6

MLN 2.4

PP

0.8

2.1

1.9

3.8 MLN 0.0

0.0 16.8

SI IEL 17.3

1.8

16.6

SI LPL 4.8 8.2

SI LPL 0.0

0.0

Thy-1.1

0.0

0.0

0.0

0.0

WT

Colon LPL 7.3 18.0

2.4 5.5

0.2

0.4 2.6

<0.1

1.7 12.6

0.3

5.5 0.5

12.8

3.3 0.2

21.6

3.7 0.7

21.4

7.9 10BiT

Foxp3

b

Foxp3– Thy-1.1+

1.0

Foxp3+ Thy-1.1+ Foxp3+ Thy-1.1–

20

0.2

2.0

0.2 0.2

2.1

1.7 0.1

8.9

4.5 9.0

0.2

4.8 16.0

0.5

2.6 0.9

15.9

15

10BiT. II10 –/–

Thy-1.1

CD4+ T cells (%)

10 5

0.5

IFN-γ

1.0

3.9

IL-17A

9.0

8.4

IFN-γ

Il-17A

LP

L

L

IFN-γ

on

LP

L

0.4

IL-17A

C

ol

SI

PP

IE SI

LN M

le

en

0 Sp

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

0.0

0.0

13.5

1.0

0.4

0.0

0.0

Colon LPL

Figure 5 IL-10 is dispensable for the development of IL-10-competent subsets of Foxp3– and Foxp3+ CD4+ T cells. Flow cytometry of cells from lymphoid tissues of 6-week-old 10BiT.Il10–/– mice, stained with monoclonal antibodies specific for CD4, Thy-1.1, and Foxp3. (a) Plots gated on CD4+ T cells; numbers in quadrants indicate percent cells in each. Data are from one of five mice examined. (b) Average frequency of Foxp3+Thy-1.1–, Foxp3+Thy-1.1+ and Foxp3–Thy-1.1+ CD4+ T cells in tissues of healthy 10BiT.Il10–/– mice. Data are the mean + s.e.m. of the frequency values from five individual mice. (c) Single-cell suspensions prepared from mesenteric lymph nodes and lamina propria of both the small and large intestine and stimulated in vitro with phorbol 12-myristate 13-acetate and ionomycin for 5 h, then with monensin for the final 4 h, and then stained extracellularly for CD4 and Thy-1.1 and intracellularly for IFN-g and IL-17A. Plots are gated on CD4+ T cells; numbers in quadrants indicate percent cells in each. Data are representative of analyses of at least five mice in two separate experiments.

define in vivo. To address the origins of these subsets and their possible lineage relationships, we first analyzed thymocytes from 10BiT mice for expression of Thy-1.1 (Fig. 4a). In contrast to results obtained with wild-type control mice, we detected no Thy-1.1 expression by any of the thymic subsets of 10BiT mice defined by CD4 and CD8 markers, including the GFP+ fraction of CD4 single-positive thymocytes. Thus, the IL-10-competent Treg cells that we characterized in extrathymic tissues (Figs. 1,3 and Supplementary Fig. 3) do not arise from a pool of precommitted thymic precursor cells, indicating that induction of Il10 expression is initiated extrathymically regardless of Foxp3 expression. To define the thymic precursor cells from which peripheral Treg subsets develop, we adoptively transferred CD4 single-positive (CD4+CD8–) thymocytes from 10BiT.Foxp3gfp mice (Fig. 4b). For this, we isolated CD8-depleted thymocyte samples from male 10BiT.Foxp3gfp donor mice (CD45.2+) and sorted them by flow cytometry into CD4+GFP+ and CD4+GFP– fractions before transferring them into sublethally irradiated, sex-matched wild-type CD45.1+ recipient mice. We monitored the transferred thymocytes for expression of GFP (Foxp3) and Thy-1.1 (Il10) by donor-derived T cells in various tissues over a 2- to 6-week period. By 2 weeks after transfer, we identified all three Treg subsets (GFP+Thy-1.1–, GFP+Thy-1.1+ and GFP–Thy-1.1+) in the secondary lymphoid tissue and GALT of mice that received CD4+GFP– cells. Notably, the relative distribution of GFP–Thy-1.1– and GFP+Thy-1.1+ resembled that of intact 10BiT mice, albeit with different absolute values that reflected variable rates of reconstitution of the different tissues (Fig. 4b, group B, and data not shown). The small intestine was enriched for Tr1-like GFP–Thy-1.1+ (Foxp3–IL-10+) cells in both the epithelial and lamina propria compartments, and the lamina propria of the small intestine was enriched for Thy-1.1+GFP+ cells. These data support published studies

936

providing evidence of peripheral conversion of Foxp3– thymic precursor cells into Foxp3+ Treg cells and indicate that there is no intrinsic bias toward or against Il10 expression by this population. Thus, expression of both Foxp3 and Il10 by Foxp3– precursor cells was induced extrathymically, and the kinetics of the development of IL-10-competent Treg cells were similar for both GFP+ (Foxp3+) and GFP– (Foxp3–) populations. Essentially all of the transferred CD4+GFP+ thymocytes retained Foxp3 expression and were present in all tissues examined. Notably, there was prominent expression of Thy-1.1 by a substantial fraction of these cells, demonstrating rapid acquisition of IL-10 competency, with relative enrichment in GALT that paralleled the distribution in 10BiT mice. Indeed, the frequencies of IL-10-competent cells were typically higher in recipients of transferred GFP+ (Foxp3+) cells than in 10BiT mice, suggesting a possible predisposition toward Il10 expression by recent thymic emigrants (Fig. 4b, group A, and data not shown). Notably, we found no downregulation of GFP (Foxp3) by transferred cells that upregulated Thy-1.1, establishing that prior expression of Foxp3 is not a prerequisite for expression of IL-10 and that the Foxp3–IL-10+ Treg cells develop by a lineage distinct from that of Foxp3+ Treg cells. Development of IL-10-competent Treg cells does not require IL-10 An established feature of Tr1 cells derived in vitro is the induction of their development by IL-10 (refs. 21,22,32). It has also been reported that mice transgenic for higher IL-10 expression in vivo more readily develop IL-10-producing Tr1-type cells than wild-type controls do33. Although such studies indicate that IL-10 enhances the development of Tr1 cells, it is unknown whether IL-10 is strictly required for the development of Tr1 cells in vivo. Given our results reported above indicating that IL-10 competency developed in a subpopulation of

VOLUME 8

NUMBER 9

SEPTEMBER 2007

NATURE IMMUNOLOGY

ARTICLES Figure 6 TGF-b is required for the development IFN-γ IL-17 IL-4 – IL-10 TGF-β of IL-10-competent Treg cells. (a,b) Flow 1.0 0.0 1.2 0.1 3.2 2.5 cytometry of naive CD4+Foxp3–CD62Lhi T cells 3.8 1.3 0.4 from 10BiT.Foxp3gfp OT-II mice, stimulated in vitro with bone marrow–derived dendritic cells 0.2 0.2 32.0 and ovalbumin peptide (1 mg/ml) either alone (–) or in the presence of recombinant IL-10 GFP 1.1 1.3 0.2 (5 ng/ml) or TGF-b (5 ng/ml). (a) Expression of GFP (Foxp3) and Thy-1.1 (Il10) after 3 d. Plots Anti-TGF-β Control IgG 1.2 1.7 are gated on CD4+ T cells; numbers in quadrants (8.5) (8.7) indicate percent cells in each. (b) Intracellular expression of IFN-g, IL-17 and IL-4 by TGF-b4.0 0.4 0.0 treated cells from a stimulated, at 5 d after 12.8 17.8 induction, for 5 h with phorbol 12-myristate CD4 13-acetate and ionomycin in the presence of GFP– Thy-1.1+ GFP+ Thy-1.1+ GFP+ Thy-1.1– monensin. Left margin, gates for plots; numbers 3.0 12 5 above bracketed lines indicate frequency of 2.5 10 4 Control IgG 2.0 8 cytokine-positive cells. (c) Expression of GFP 3 Anti-TGF-β 6 1.5 (Foxp3) and Thy-1.1 (Il10) by donor CD45.1+ 2 1.0 4 + – 6 cells among CD4 Foxp3 thymocytes (4  10 ) 1 2 0.5 0 from male 10BiT.Foxp3gfp mice transferred into 0.0 0 congenic, age- and sex-matched CD45.1 mice (as described in Fig. 4); recipient mice were treated with neutralizing monoclonal anti-TGF-b (1D11) or isotype control antibody (immunoglobulin G (IgG)) and received 250 mg antibody (per mouse) at 24 h before cell transfer and 125 mg antibody (per mouse) every 4 d thereafter, then were killed for analysis 2 weeks after transfer. Top, flow cytometry of cells from Peyer’s patches of recipient mice in a lymphocyte gate; numbers in quadrants indicate frequency of total cells in each (numbers in parentheses indicate percent CD45.2+ donor cells in the total CD4+ T cell pool). Bottom, frequency of GFP–Thy-1.1+, GFP–Thy-1.1+ and GFP+Thy-1.1+ donor cells in tissues of recipient mice. Spl, spleen. Data are representative of two experiments.

b

GFP+ Thy-1.1–

Thy-1.1

Total + Thy-1.1

a

GFP– Thy-1.1–

CD45.2

Foxp3+ Treg cells after thymic export, we also sought to determine whether IL-10 might be involved in inducing IL-10 competency in this population. We crossed the 10BiT reporter onto an IL-10-deficient background (10BiT.Il10 –/–) and analyzed expression of the Thy-1.1 reporter and endogenous Foxp3 by CD4+ T cells. Notably, reporter-positive, IL-10competent cells were present in both Foxp3– and Foxp3+ CD4+ populations at relative frequencies approximating those of wild-type 10BiT mice in all tissue compartments examined (Fig. 5). The Thy-1.1+ fraction of Foxp3– Tr1-type 10BiT.Il10 –/– cells had slightly less Thy-1.1 expression in some experiments (Fig. 5a), but this was not a consistent finding (Fig. 5c), and there were no gross differences in the frequencies of these cells in any of the tissue compartments examined (Figs. 5a,b and 3a). Thus, induction of IL-10 competency in CD4+ T cells does not require IL-10 signaling in either the Foxp3+ or Foxp3– lineage. IL-10 can be produced by a subset of TH1 or TH2 cells in the context of chronic inflammatory conditions or infection34–38, consistent with the original association of IL-10 with an effector lineage (TH2) and reinforcing the fact that expression of IL-10 is not unique to a single lineage. Because intestine-associated lymphoid tissues normally contain a fraction of effector T cells, which are increased during colitis or intestinal infection, we sought to determine whether the IL-10competent cells that developed in the intestine-associated lymphoid tissues of IL-10-deficient and wild-type mice were linked to effector T cells (such as TH1 and IL-17-producing T helper cells) in these sites. We thus evaluated the frequencies of IFN-g+ and IL-17+ T cells in intestine-associated lymphoid tissue of wild-type and 10BiT mice and found small numbers of IFN-g+ and IL-17+ T cells that were greater in the lamina propria of the colon in association with the early development of colitis in 10BiT.Il10 –/– mice (Fig. 5c and Supplementary Fig. 3). However, we found no substantial coexpression of IL-10 in IFN-g+ or IL-17+ cells, indicating that the IL-10-competent Treg subpopulations did not overlap with those effector T helper cells. Thus, most if not all of the IL-10

NATURE IMMUNOLOGY

VOLUME 8

NUMBER 9

SEPTEMBER 2007

IE L LP L

PP

Sp l M LN

IE L LP L

PP

Sp l M LN

IE L LP L

PP

Sp l M LN

CD45.2+ (%)

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

c

expression not associated with Foxp3+ cells was more typical of the Tr1 subset. Given the IL-10-independent development of IL-10-competent Treg cells, we assessed the involvement of TGF-b in this process. We initially compared the effects of supplementation of IL-10 and TGF-b on the expression of Thy-1.1 by naive 10BiT.Foxp3gfp (Foxp3–Thy-1.1–CD62Lhi) CD4+ T cells activated ex vivo. We activated cells for 3 d in culture with immature bone marrow–derived dendritic cells in the presence or absence of exogenous IL-10 or TGF-b and then analyzed expression of GFP (Foxp3) and Thy-1.1 (IL-10; Fig. 6a). Although the addition of IL-10 had a minimal effect on the expression of either GFP or Thy-1.1, TGF-b induced substantial GFP (Foxp3) expression and less but considerable Thy-1.1 (IL-10) expression in both the GFP+ and GFP– fractions. This expression was independent of induction of substantial numbers of effector cytokine–positive cells and seemed to be a direct effect on T cells rather than an indirect effect of cytokine modulation (Fig. 6b and data not shown). Thus, TGF-b can induce the development of all Treg subsets (Foxp3+IL-10–, Foxp3+IL-10+ and Foxp3–IL-10+) in vitro. To analyze the requirement for TGF-b in vivo, we adoptively transferred CD4+CD8– GFP– thymocytes from 10BiT.Foxp3gfp mice into congenic (CD45.1) hosts and assessed the development of Treg cells in conditions of TGF-b neutralization (Fig. 6c). Mice that received neutralizing anti-TGF-b had much lower frequencies of all Treg subsets than did recipient mice treated with control antibody, particularly in intestine-associated lymphoid tissues, despite the recovery of similar total numbers of donor cells. Thus, in contrast to IL-10, TGF-b seems to be essential for the development of IL-10-competent Treg cells from CD4+ precursor cells, regardless of Foxp3 status. DISCUSSION Here we used transgenic reporter mice to gain new insight into the diversity, distribution and developmental origins of Treg cell subsets. We found that in addition to natural and adaptive Foxp3+ Treg cells,

937

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES Tr1-type adaptive Treg cells represented a substantial component of the normal Treg repertoire that developed IL-10 competency extrathymically by a Foxp3-independent program of differentiation. Acquisition of IL-10 competency by a fraction of natural and adaptive Foxp3+ Treg cells also occurred extrathymically, and in both Foxp3+ and Foxp3– lineages, induction of IL-10 was independent of a requirement for IL-10 signaling but was dependent on TGF-b, extending the diversity of effects of this pleiotropic cytokine in T cell developmental programs. Thus, TGF-b participates in the development and/or maintenance of all Treg subsets whether they originate from Foxp3+ or Foxp3– thymic precursor cells39–43, in addition to its involvement in the development of the IL-17-producing T helper cell effector lineage44–46. Our finding that the development of IL-10-competent Foxp3– Treg cells was mostly unimpaired in IL-10-deficient mice was unanticipated. Reports that have characterized the requirements for the induction of Tr1 cells have consistently indicated that IL-10 is a critical factor, acting indirectly to modulate the differentiation or activation of immature dendritic cells that prime Tr1 development. Specifically, it has been shown that in concert with granulocytemacrophage colony-stimulating factor and tumor necrosis factor, IL-10 promotes the differentiation from bone marrow precursor cells of a plasmacytoid dendritic cell population characterized by a CD11cloCD45RBhi phenotype, which induces Tr1 cells in vitro and in vivo22. Other studies have also shown that IL-10 induces the development of Tr1 cells25,47. In contrast, we found here that although Tr1 cells that developed in 10BiT.Il10 –/– mice may have had slightly lower 10BiT reporter expression, the frequencies of IL-10-competent cells were similar to those of wild-type 10BiT mice, indicating that IL-10 is dispensable for the induction of Tr1 cell development in vivo. Similarly, Foxp3+ cells acquired IL-10 competence in the absence of IL-10. A critical function for IL-10 in the development of IL-10-competent Foxp3+ Treg cells has been neither supported nor refuted by published studies, in part because of the difficulty in identifying the relatively small subpopulation of IL-10+ cells defined here in the peripheral lymphoid tissues (which have been the main source of Foxp3+ cells analyzed before). Thus, IL-10 does not directly beget IL-10 Treg cells, and although our data do not negate published studies indicating that IL-10 can induce the development of adaptive Treg cells, clearly it is not required. Although IL-10 was dispensable for the development of all Treg cells that arose from Foxp3– precursor cells after thymic export, TGF-b seemed to be essential. Thus, blockade of TGF-b suppressed the development of both Foxp3+ and Foxp3– adaptive Treg cells. Given the importance of TGF-b in the maintenance of Foxp3 expression48, and the decreased IL-10 expression in mature Treg cells by induced ablation of the Foxp3 allele16, interpretation of our results in terms of a requirement for extrathymic development of adaptive Foxp3+ Treg cells is complex and will require further study. However, these data indicate TGF-b is an essential factor in the development of Tr1 cells, and it should now be possible to delineate the mechanism for such development in greater detail. These data also provide a basis for understanding the effects of TGF-b neutralization in inhibiting the protective function of Treg cells in models of colitis10 and are consistent with a model in which TGF-b participates in the development of all IL-10-competent Treg cells that mediate intestinal immune regulation. In the steady state, IL-10 expression is limited to T cells, particularly T cells in the gut. We detected expression of the Thy-1.1 reporter in 10BiT mice outside the T cell compartment in only a small subset of B cells and in rare dendritic cells present in the intestinal lymphoid tissues (data not shown). We detected no expression of the Thy-1.1

938

reporter by nonhematopoietic cells in any tissue examined, including the gut. In contrast, the main subpopulations of Foxp3+ and Foxp3– subsets of Treg cells were IL-10 competent in the normal T cell repertoire and were concentrated in GALT, albeit with notable differences in distribution. Thus, whereas IL-10 expression is limited to a minor fraction of Foxp3+ Treg cells that circulate through peripheral lymph nodes, it is a prominent feature of Foxp3+ Treg cells that traffic to lymphoid tissues of the gut, particularly the large intestine. This finding may explain why IL-10-deficient mice and mice treated with blocking anti-IL-10R succumb to intestinal inflammation despite having substantial numbers of Foxp3+ cells in the intestines. Similarly, Foxp3– Treg cells are a minor fraction of the CD4+ T cells that circulate in extraintestinal lymphoid tissues but are particularly abundant in the small intestine, where they seem to constitute the single most abundant CD4+ T cell subset. Notably, IL-10 competency was not limited to the TCRab lineage or the CD4+ subset, as we identified many IL-10-competent gd T cells and CD8+ T cells in the intestinal epithelium, where their functions remain to be defined. In this context, TCRab, CD8aa T cells isolated from IELs of the small intestine have been reported to exert regulatory effects by an IL-10-dependent mechanism49. Given the fact that the intestines have the largest number of T cells in the body, the finding that the highest density of IL-10-competent T cells is in this tissue establishes that most IL-10-competent T cells reside in the intestines. Indeed, it is likely that in the steady state, IL-10-competent T cells represent the single largest population of cytokine-positive T cells, regardless of regulatory or effector lineage. Although the basis for the enrichment of IL-10+ Treg cells in the gut no doubt reflects the nonredundant function of IL-10 in maintaining immune homeostasis to the diverse intestinal microbiota, the basis for the apparent ‘subdivision of labor’ among Foxp3+ and Foxp3– Treg subsets is unclear. In different studies, both Foxp3+ and Tr1-type Foxp3– Treg cells have been linked to the suppression of colitis in transfer models28,50, but the relative contribution of each to normal homeostasis has been contentious. A confounding variable in each of these CD4+ T cell transfer studies has been the heterogeneity of T cell subpopulations in the isolated CD45RBlo and even CD25hi fractions used for adoptive transfer. As shown in our analysis of the surface phenotypes of the Treg subsets defined by expression of Foxp3 and IL-10, all three Treg subsets were present in both the CD45RBlo and CD25hi fractions, emphasizing the problem of assigning a lineage or functional importance to cells fractionated on the basis of these markers. The ability to discriminate and isolate distinct subsets of Treg cells from 10BiT.Foxp3gfp reporter mice should permit even more definitive studies of the specific functions of Foxp3+ and Foxp3– Treg cells in the future. Assuming that Tr1 cells normally are important in preventing inflammatory responses to the intestinal flora, it is unclear why most of these cells should be restricted to the small intestine, whereas most intestinal inflammation in mouse models is restricted to the large intestine. Although this will require further study, one possibility is that IL-10-competent Foxp3+ and Foxp3– subsets might be ‘preferentially’ induced by distinct classes of flora antigens that reside in different microenvironmental niches along the intestinal tract and elicit distinct programs of adaptive Treg development. Thus, Tr1 cells might be specifically induced by a distinct class of commensal organisms that reside ‘preferentially’ in the upper gastrointestinal tract. Analysis of the frequency of Foxp3+ and Foxp3– IL-10-competent cells along the longitudinal axis of the intestine and correlation with the regional distribution of microbial species in mice with defined flora versus germ-free mice might be instructive. Alternatively, Tr1 cells could be

VOLUME 8

NUMBER 9

SEPTEMBER 2007

NATURE IMMUNOLOGY

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES induced mainly by ingested antigens as a mechanism of inducing tolerance to food, although studies have indicated that Foxp3+ cells are induced by oral protein antigens and are dependent on TGF-b and not IL-10 for their function51. Finally, it is possible that Tr1 cells have a very different function in the repertoire, perhaps specializing in the antigendriven development of B cells to provide help for secretory immunoglobulin A responses, for which class switching may be promoted by IL-10 and TGF-b and is dominant in the small intestine52,53. This possibility could conceivably contribute to regulation of the composition, density and translocation of the commensal flora, as well as protection against invasive mucosal pathogens. In addition to its central involvement in the maintenance of immune homeostasis in the milieu of the intestinal microbiota, IL-10 is important in regulating effector responses that emerge in response to infection. T cells with IL-10-dependent regulatory activity have been described in an increasing number of models of intracellular infection, including those induced by bacterial, protozoal and viral pathogens34,37,38,54. The production of IL-10 in many of these settings seems to represent a feedback mechanism to curb tissue injury during pathogen clearance, although in circumstances in which pathogen clearance is not achieved, the emergence of IL-10-producing cells ultimately prevents host clearance of the pathogen, setting the stage for chronic infection. In this sense, the host becomes ‘tolerant’ of the pathogen by an IL-10-dependent mechanism that promotes symbiosis rather than pathogen clearance. Although classical Treg cells, both Foxp3+ and Tr1, have been suggested as cellular sources of IL-10 in many studies, several reports have demonstrated that IL-10 production by TH1 cells that also produce IFN-g is responsible for pathogen latency in leishmania and toxoplasma infections34,37,38, establishing yet another pathway to IL-10-based immune regulation. In the sense that the intestinal microbiota represent a ‘chronic infection’ (that is, chronic colonization) that does not lead to clearance, IL-10-producing Treg cells might be seen as a natural adaptation to microbial antigens that resist clearance, in this case setting the stage for the evolution of a mutually beneficial symbiosis. In our analyses of GALT T cells isolated from 10BiT.Il10–/– mice, we did not identify a substantial population of T cells that coexpressed IL-10 and effector cytokines (such as IFN-g or IL-17) or find coexpression of Foxp3 and IL-10 in CD45RBhi cells transferred from 10BiT.Foxp3gfp donors that acquired expression of IFN-g and IL-17 in recipients that developed colitis. Nevertheless, our studies emphasize the idea that unlike Foxp3, which defines a regulatory lineage, there are many pathways to IL-10 that defy simple lineage associations. An IL-10 reporter ‘knock-in’ mouse has been reported55 in which GFP-encoding sequence introduced into Il10 downstream of an internal ribosomal entry site element permits identification of cells induced to express Il10. In the steady state, few if any reporter-positive cells are identified in vivo in this system, and substantial numbers of IL-10-expressing T cells are detected only in conditions of robust polyclonal activation. Although this observation might reflect limited sensitivity of reporter detection in mice transgenic for this IL-10 reporter, it is more likely that the half-life of transcripts of this reporter reflects that of the native Il10 transcripts, which, because of the inherent instability typical of cytokine mRNA molecules, results in very transient production of IL-10 by individual cells and thus few IL-10 protein–producing cells in a T cell population at any given time. The 10BiT-transgenic model described here, in contrast, stably identifies essentially all cells in which Il10 alleles have been previously activated and those cells actively transcribing Il10, making it uniquely useful for the delineation and isolation of Treg subsets in the steady state as well as analysis of the diversity of

NATURE IMMUNOLOGY

VOLUME 8

NUMBER 9

SEPTEMBER 2007

IL-10-competent T cells and possible non–T cells that arise in the course of infection or autoimmunity. METHODS Mice. C57BL/6, C57BL/6.Ly5.1 (CD45.1+) and C57BL/6 IL-10-deficient mice were from Jackson Laboratories. The generation of 10BiT reporter mice is described in the Supplementary Methods online. Foxp3gfp reporter mice have been described26 and were backcrossed onto the C57BL/6 background before being intercrossed with 10BiT mice. All mice were bred and maintained at the University of Alabama at Birmingham in accordance with regulations of the Institutional Animal Care and Use Committee. CD4+ T cell isolation and activation. Spleens and lymph nodes were collected from the various strains of mice and single-cell suspensions were prepared by mechanical disruption in RPMI 1640 medium supplemented with 10% (vol/vol) FCS, penicillin (100 IU/ml), streptomycin (100 mg/ml), 1 nonessential amino acids, sodium pyruvate (1 mM), b-mercaptoethanol (2.5 mM) and L-glutamine (2 mM; ‘R-10 medium’). CD4+ T cells were isolated with Dynabeads mouse CD4 beads followed by DETACHaBEAD mouse CD4, according to the manufacturer’s directions (Dynal Biotech). For T cell polarization, CD4+ T cells were stimulated with anti-CD3 (2.5 mg/ml; 145-2C11; prepared ‘in-house’) and irradiated (3,000 rads) splenic feeder cells, which were cultured with CD4+ T cells at a ratio of 5:1. TH2 cells were differentiated by the addition of recombinant IL-4 (1,000 U/ml; R&D Systems), anti-IL-12 (10 mg/ml; C17.8) and anti-IFN-g (10 mg/ml; XMG). For the differentiation of Treg cells ex vivo, bone marrow–derived dendritic cells at day 5, used as antigenpresenting cells, were cultured with CD4+ T cells at a ratio of 1:5. CD4+ T cells were isolated from 10BiT.Foxp3gfp.OT-II mice as described above and the CD25–Foxp3–CD62Lhi fraction was sorted by flow cytometry and was stimulated with ovalbumin peptide (1 mg/ml) in the presence of recombinant IL-2 (50 U/ml; R&D Systems) or recombinant TGF-b (5 ng/ml; R&D Systems) or both. In all cases, CD4+ T cell cultures were split 1:2 on day 3 after activation and were further supplemented with recombinant IL-2. Antibodies and flow cytometry. The following antibodies were from BD Biosciences: fluorescein isothiocyanate–conjugated anti-CD8a (53.67), anti-CD90.1 (OX-7), anti-TCRgd (GL3), anti-CTLA-4 (UC10-4F10-11), anti-CD103 (M290), anti-IL-4 (11B11), anti-CD62L (MEL-14), peridinin chlorophyll protein–conjugated anti-CD90.1 (OX-7), biotinylated anti-CD8b (53-5.8), anti-TCRb (H57-597), anti-CD69 (H1-2F3), anti-CD25 (7D4) and anti-CD45RB (16A). The following antibodies were from eBiosciences: fluorescein isothiocyanate–conjugated anti-Foxp3 (FJK-16s), anti-IL-10 (JES516E3), phycoerythrin-conjugated anti-CD8a (53-6.7), anti-ICOS (7E.17G9), phycoerythrin–indotricarbocyanine-conjugated anti-CD4 (L3T4), allophycocyanin-conjugated anti-CD4 (L3T4) and biotin-conjugated anti-GITR (DTA-1). Acquisitions were made with a FACSCaliber or LSRII and data were analyzed with CellQuest Pro software (BD Biosciences). Intestinal lymphocyte isolation. Intestines were removed and Peyer’s patches were dissected from the small intestines. Intestines were opened longitudinally and then were cut into strips 1 cm in length. Tissues were washed in cold 1 Hank’s balanced-salt solution supplemented with 2% (vol/vol) FCS plus penicillin (100 IU/ml) and streptomycin (100 mg/ml; ‘H2 solution’). IELs were isolated as follows: gut pieces were incubated for 30 min at 37 1C with gentle magnetic stirring in H2 solution containing L-dithioerythreitol (154 mg/l). Cell suspensions were passed through a hand-held sieve and tissues were washed with H2 solution. Tissues were then incubated for 30 min at 37 1C with gentle magnetic stirring in H2 solution containing 2 mM EDTA. After being strained and washed, the remaining gut tissues were digested completely for 45 min at 37 1C with gentle magnetic stirring in R-10 medium with collagenase D (100 U/ml) and DNase (20 mg/ml; Sigma) for the isolation of LPLs. IELs and LPLs were purified on a 40%/75% Percoll gradient by centrifugation for 20 min at 25 1C and 600g. with no brake. Cytokine analysis. Specific populations of CD4+ T cells were left unstimulated or were stimulated for 18 h with immobilized anti-CD3 (10 mg/ml) and soluble anti-CD28 (2.5 mg/ml). Supernatants were assayed for cytokine

939

ARTICLES

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

expression with Cytometric Bead Array Flex Sets according to the manufacturer’s protocol (BD Biosciences). In vitro proliferation assay. CD4+ T cells (2  104) of various phenotypes were cultured alone or in the presence of an equal number of splenic CD25–Thy-1.1– T cells. T cells were stimulated in triplicate with irradiated splenic feeder cells (2  105 cells/well) and soluble anti-CD3 (2.5 mg/ml; 145-2C11). Anti–IL-10 receptor (10 mg/ml; 1.B1.3a) or anti-TGF-b (10 mg/ml; 1D11) was added to some wells. Cultures were pulsed with [3H]thymidine (1 mCi/well) for the final 18 h of the 72-hour incubation. CD4+ thymocyte isolation and adoptive transfer. Thymi were pooled from 4- to 6-week-old male 10BiT.Foxp3gfp mice and single-cell suspensions were prepared by mechanical disruption in R-10 medium. Cells were washed and resuspended in PBS supplemented with 1% (vol/vol) FCS and 2 mM EDTA. Cells were labeled with phycoerythrin-conjugated anti-CD8a or phycoerythrinconjugated anti-Thy-1.1 (and allophycocyanin-conjugated anti-CD4) followed by anti-phycoerythrin microbeads (Miltenyi Biotec). Samples were depleted of CD8a+ thymocytes on a MACS LD separation column (Miltenyi Biotec). CD4+ cells in the CD8a-depleted fraction were sorted into GFP– and GFP+ fractions with a FACSAria (BD Biosciences). Foxp3– cells (3  106 to 5  106) or Foxp3+ cells (2  105 to 5  105) were transferred retro-orbitally into sublethally irradiated (400 rads), sex-matched CD45.1+ recipients. Immunohistology and CD45RBhi T cell isolation, adoptive transfer and histopathology. These procedures are described in the Supplementary Methods. Statistical analyses. Statistical significance was calculated with an unpaired Student’s t-test, Mann-Whitney U-test or analysis of variance. All P values of 0.05 or less were considered significant, unless indicated otherwise. Note: Supplementary information is available on the Nature Immunology website.

ACKNOWLEDGMENTS We thank D. Chaplin, C. Elson, R. Lorenz, L. Timares, M. Walter and members of the Weaver laboratory for comments and suggestions; C. Song and M. Blake for technical assistance; and N. LeLievre for editorial assistance. Supported by the National Institutes of Health (C.T.W.), the Crohn’s and Colitis Foundation of America (C.T.W. and L.E.H.) and the Howard Hughes Medical Institute (A.Y.R.). AUTHOR CONTRIBUTIONS C.L.M. did all experiments and collected and analyzed all data with assistance from L.E.H. and C.T.W.; K.M.J. assisted in the development of transgenic mice; J.R.O. provided animal breeding and genotyping technical support; C.L.Z. assisted in the design and execution of immunofluorescence studies; A.Y.R. contributed mice and expertise for some reporter mouse studies; and C.L.M. and C.T.W. wrote the manuscript. COMPETING INTERESTS STATEMENT The authors declare no competing financial interests. Published online at http://www.nature.com/natureimmunology Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions

1. Fiorentino, D.F., Bond, M.W. & Mosmann, T.R. Two types of mouse T helper cell. IV. Th2 clones secrete a factor that inhibits cytokine production by Th1 clones. J. Exp. Med. 170, 2081–2095 (1989). 2. Fiorentino, D.F. et al. IL-10 acts on the antigen-presenting cell to inhibit cytokine production by Th1 cells. J. Immunol. 146, 3444–3451 (1991). 3. Moore, K.W., de Waal Malefyt, R., Coffman, R.L. & O’Garra, A. Interleukin-10 and the interleukin-10 receptor. Annu. Rev. Immunol. 19, 683–765 (2001). 4. O’Garra, A., Vieira, P.L., Vieira, P. & Goldfeld, A.E. IL-10-producing and naturally occurring CD4+ Tregs: limiting collateral damage. J. Clin. Invest. 114, 1372–1378 (2004). 5. Kuhn, R., Lohler, J., Rennick, D., Rajewsky, K. & Muller, W. Interleukin-10-deficient mice develop chronic enterocolitis. Cell 75, 263–274 (1993). 6. Sellon, R.K. et al. Resident enteric bacteria are necessary for development of spontaneous colitis and immune system activation in interleukin-10-deficient mice. Infect. Immun. 66, 5224–5231 (1998).

940

7. Davidson, N.J. et al. T helper cell 1-type CD4+ T cells, but not B cells, mediate colitis in interleukin 10-deficient mice. J. Exp. Med. 184, 241–251 (1996). 8. Asseman, C., Mauze, S., Leach, M.W., Coffman, R.L. & Powrie, F. An essential role for interleukin 10 in the function of regulatory T cells that inhibit intestinal inflammation. J. Exp. Med. 190, 995–1004 (1999). 9. Roers, A. et al. T cell-specific Inactivation of the interleukin 10 gene in mice results in enhanced T cell responses but normal innate responses to lipopolysaccharide or skin Irritation. J. Exp. Med. 200, 1289–1297 (2004). 10. Izcue, A., Coombes, J.L. & Powrie, F. Regulatory T cells suppress systemic and mucosal immune activation to control intestinal inflammation. Immunol. Rev. 212, 256–271 (2006). 11. Zheng, Y. & Rudensky, A.Y. Foxp3 in control of the regulatory T cell lineage. Nat. Immunol. 8, 457–462 (2007). 12. Jordan, M.S. et al. Thymic selection of CD4+CD25+ regulatory T cells induced by an agonist self-peptide. Nat. Immunol. 2, 301–306 (2001). 13. Apostolou, I., Sarukhan, A., Klein, L. & von Boehmer, H. Origin of regulatory T cells with known specificity for antigen. Nat. Immunol. 3, 756–763 (2002). 14. Fontenot, J.D., Gavin, M.A. & Rudensky, A.Y. Foxp3 programs the development and function of CD4+CD25+ regulatory T cells. Nat. Immunol. 4, 330–336 (2003). 15. Hori, S., Nomura, T. & Sakaguchi, S. Control of regulatory T cell development by the transcription factor Foxp3. Science 299, 1057–1061 (2003). 16. Williams, L.M. & Rudensky, A.Y. Maintenance of the Foxp3-dependent developmental program in mature regulatory T cells requires continued expression of Foxp3. Nat. Immunol. 8, 277–284 (2007). 17. Read, S., Malmstrom, V. & Powrie, F. Cytotoxic T lymphocyte-associated antigen 4 plays an essential role in the function of CD25+CD4+ regulatory cells that control intestinal inflammation. J. Exp. Med. 192, 295–302 (2000). 18. Sakaguchi, S., Sakaguchi, N., Asano, M., Itoh, M. & Toda, M. Immunologic selftolerance maintained by activated T cells expressing IL-2 receptor a-chains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases. J. Immunol. 155, 1151–1164 (1995). 19. Tone, M. et al. Mouse glucocorticoid-induced tumor necrosis factor receptor ligand is costimulatory for T cells. Proc. Natl. Acad. Sci. USA 100, 15059–15064 (2003). 20. Weaver, C.T., Harrington, L.E., Mangan, P.R., Gavrieli, M. & Murphy, K.M. Th17: an effector CD4 T cell lineage with regulatory T cell ties. Immunity 24, 677–688 (2006). 21. Groux, H. et al. A CD4+ T-cell subset inhibits antigen-specific T-cell responses and prevents colitis. Nature 389, 737–742 (1997). 22. Wakkach, A. et al. Characterization of dendritic cells that induce tolerance and T regulatory 1 cell differentiation in vivo. Immunity 18, 605–617 (2003). 23. Groux, H. Type 1 T-regulatory cells: their role in the control of immune responses. Transplantation 75, 8S–12S (2003). 24. Vieira, P.L. et al. IL-10-secreting regulatory T cells do not express Foxp3 but have comparable regulatory function to naturally occurring CD4+CD25+ regulatory T cells. J. Immunol. 172, 5986–5993 105, 1162–1169 (2004). 25. Levings, M.K. et al. Differentiation of Tr1 cells by immature dendritic cells requires IL-10 but not CD25+CD4+ Treg cells. Blood 105, 1162–1169 (2004). 26. Fontenot, J.D. et al. Regulatory T cell lineage specification by the forkhead transcription factor Foxp3. Immunity 22, 329–341 (2005). 27. Powrie, F., Leach, M.W., Mauze, S., Caddle, L.B. & Coffman, R.L. Phenotypically distinct subsets of CD4+ T cells induce or protect from chronic intestinal inflammation in C. B-17 scid mice. Int. Immunol. 5, 1461–1471 (1993). 28. Uhlig, H.H. et al. Characterization of Foxp3+CD4+CD25+ and IL-10-secreting CD4+CD25+ T cells during cure of colitis. J. Immunol. 177, 5852–5860 (2006). 29. Lohning, M. et al. Expression of ICOS in vivo defines CD4+ effector T cells with high inflammatory potential and a strong bias for secretion of interleukin 10. J. Exp. Med. 197, 181–193 (2003). 30. Levings, M.K. et al. IFN-a and IL-10 induce the differentiation of human type 1 T regulatory cells. J. Immunol. 166, 5530–5539 (2001). 31. Cong, Y., Weaver, C.T., Lazenby, A. & Elson, C.O. Bacterial-reactive T regulatory cells inhibit pathogenic immune responses to the enteric flora. J. Immunol. 169, 6112–6119 (2002). 32. Roncarolo, M.G. et al. Interleukin-10-secreting type 1 regulatory T cells in rodents and humans. Immunol. Rev. 212, 28–50 (2006). 33. Groux, H. et al. A transgenic model to analyze the immunoregulatory role of IL-10 secreted by antigen-presenting cells. J. Immunol. 162, 1723–1729 (1999). 34. Belkaid, Y. et al. The role of interleukin (IL)-10 in the persistence of Leishmania major in the skin after healing and the therapeutic potential of anti-IL-10 receptor antibody for sterile cure. J. Exp. Med. 194, 1497–1506 (2001). 35. Akbari, O. et al. Antigen-specific regulatory T cells develop via the ICOS-ICOS-ligand pathway and inhibit allergen-induced airway hyperreactivity. Nat. Med. 8, 1024–1032 (2002). 36. Stock, P. et al. Induction of T helper type 1-like regulatory cells that express Foxp3 and protect against airway hyper-reactivity. Nat. Immunol. 5, 1149–1156 (2004). 37. Anderson, C.F., Oukka, M., Kuchroo, V.J. & Sacks, D. CD4+CD25–Foxp3– Th1 cells are the source of IL-10-mediated immune suppression in chronic cutaneous leishmaniasis. J. Exp. Med. 204, 285–297 (2007). 38. Jankovic, D. et al. Conventional T-bet+Foxp3– Th1 cells are the major source of hostprotective regulatory IL-10 during intracellular protozoan infection. J. Exp. Med. 204, 273–283 (2007). 39. Weiner, H.L. Oral tolerance: immune mechanisms and the generation of Th3-type TGF-b-secreting regulatory cells. Microbes Infect. 3, 947–954 (2001).

VOLUME 8

NUMBER 9

SEPTEMBER 2007

NATURE IMMUNOLOGY

© 2007 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES 40. Chen, W. et al. Conversion of peripheral CD4+CD25- naive T cells to CD4+CD25+ regulatory T cells by TGF-b induction of transcription factor Foxp3. J. Exp. Med. 198, 1875–1886 (2003). 41. Li, M.O., Sanjabi, S. & Flavell, R.A. Transforming growth factor-b controls development, homeostasis, and tolerance of T cells by regulatory T cell-dependent and -independent mechanisms. Immunity 25, 455–471 (2006). 42. Marie, J.C., Liggitt, D. & Rudensky, A.Y. Cellular mechanisms of fatal early-onset autoimmunity in mice with the T cell-specific targeting of transforming growth factor-b receptor. Immunity 25, 441–454 (2006). 43. Kitani, A. et al. Transforming growth factor (TGF)-b1-producing regulatory T cells induce Smad-mediated interleukin 10 secretion that facilitates coordinated immunoregulatory activity and amelioration of TGF-b1-mediated fibrosis. J. Exp. Med. 198, 1179–1188 (2003). 44. Veldhoen, M., Hocking, R.J., Atkins, C.J., Locksley, R.M. & Stockinger, B. TGFb in the context of an inflammatory cytokine milieu supports de novo differentiation of IL-17producing T cells. Immunity 24, 179–189 (2006). 45. Mangan, P.R. et al. Transforming growth factor-b induces development of the TH17 lineage. Nature 441, 231–234 (2006). 46. Bettelli, E. et al. Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells. Nature 441, 235–238 (2006). 47. Van Montfrans, C. et al. Prevention of colitis by interleukin 10-transduced T lymphocytes in the SCID mice transfer model. Gastroenterology 123, 1865–1876 (2002).

NATURE IMMUNOLOGY

VOLUME 8

NUMBER 9

SEPTEMBER 2007

48. Marie, J.C., Letterio, J.J., Gavin, M. & Rudensky, A.Y. TGF-b1 maintains suppressor function and Foxp3 expression in CD4+CD25+ regulatory T cells. J. Exp. Med. 201, 1061–1067 (2005). 49. Poussier, P., Ning, T., Banerjee, D. & Julius, M. A unique subset of selfspecific intraintestinal T cells maintains gut integrity. J. Exp. Med. 195, 1491–1497 (2002). 50. Foussat, A. et al. A comparative study between T regulatory type 1 and CD4+CD25+ T cells in the control of inflammation. J. Immunol. 171, 5018–5026 (2003). 51. Mucida, D. et al. Oral tolerance in the absence of naturally occurring Tregs. J. Clin. Invest. 115, 1923–1933 (2005). 52. Defrance, T. et al. Interleukin 10 and transforming growth factor b cooperate to induce anti-CD40-activated naive human B cells to secrete immunoglobulin A. J. Exp. Med. 175, 671–682 (1992). 53. Cazac, B.B. & Roes, J. TGF-b receptor controls B cell responsiveness and induction of IgA in vivo. Immunity 13, 443–451 (2000). 54. Brooks, D.G., Teyton, L., Oldstone, M.B. & McGavern, D.B. Intrinsic functional dysregulation of CD4 T cells occurs rapidly following persistent viral infection. J. Virol. 79, 10514–10527 (2005). 55. Kamanaka, M. et al. Expression of interleukin-10 in intestinal lymphocytes detected by an interleukin-10 reporter knockin tiger mouse. Immunity 25, 941–952 (2006).

941