Human CD4+CD25+Foxp3+ Regulatory T Cells Do Not Constitutively Express IL-35 Emilie Bardel, Frédérique Larousserie, Pascaline Charlot-Rabiega, Aurore Coulomb-L'Herminé and Odile This information is current as Devergne of September 25, 2021. J Immunol 2008; 181:6898-6905; ; doi: 10.4049/jimmunol.181.10.6898 http://www.jimmunol.org/content/181/10/6898 Downloaded from

References This article cites 32 articles, 13 of which you can access for free at: http://www.jimmunol.org/content/181/10/6898.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 25, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2008 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Human CD4؉CD25؉Foxp3؉ Regulatory T Cells Do Not Constitutively Express IL-35

Emilie Bardel,1* Fre´de´rique Larousserie,1*† Pascaline Charlot-Rabiega,* Aurore Coulomb-L’Hermine´,‡ and Odile Devergne2*

EBV-induced 3 (EBI3) can associate with p28 to form the heterodimeric IL-27, or with the p35 subunit of IL-12 to form the EBI3/p35 heterodimer, recently named IL-35. In mice, IL-35 has been shown to be constitutively expressed by CD4؉CD25؉Foxp3؉ regulatory T cells (Treg cells) and suggested to contribute to their suppressive activity. In this study, we investigated whether human Treg cells express IL-35. Double-staining analysis of human thymuses showed that neither Foxp3؉ nor CD25؉ cells coexpressed EBI3. Similarly, Foxp3؉ cells present in human lymph nodes, tonsils, spleens, and intestines did not express EBI3. Consistent with these in situ observations, Treg cells purified from blood or tonsils were negative for EBI3 by ؉ ؉ immunoblotting. Other human subsets, including effector T cells, naive and memory CD4 T cells, CD8 and ␥␦ T cells Downloaded from -also did not constitutively express EBI3, which contrasts with IL-35 expression observed in murine CD8؉ and ␥␦ T cells. Fur thermore, although CD3/CD28 stimulation consistently induced low levels of EBI3 in various CD4؉ T cell subsets, no EBI3 could be detected in CD3/CD28-stimulated Treg cells. RT-PCR analysis showed that, whereas p35 transcripts were detected in both Teff and Treg cells, EBI3 transcripts were detected only in activated Teff cells, but not in resting or activated Treg cells. Thus, in contrast to their murine counterpart, human Treg cells do not express detectable amounts of IL-35. The Journal of Immunology, 2008, 181: 6898–6905. http://www.jimmunol.org/

BV-induced gene 3 (EBI3)3 is a member of the IL-12 CD4ϩCD25ϩFoxp3ϩ regulatory T (Treg) cells (4). Transcripts family that was identified in B lymphocytes based on its coding for EBI3 and p35 were observed to be constitutively co- E induction following EBV infection (1). It codes for a 34 expressed by mouse Treg cells and EBI3/p35 heterodimer was kDa-secreted glycoprotein homologous to the p40 subunit of IL- coprecipitated from the cell culture supernatant of these cells. In 12. It can associate with p28, a peptide analogous to the p35 sub- addition, in vitro and in vivo studies suggested that the expression unit of IL-12, to form the heterodimeric cytokine EBI3/p28, named of IL-35 by mouse Treg cells contributed to their suppressive func- IL-27 (2). It can also dimerize with IL-12 p35 to form another tion. In transfer experiments, Treg cells from EBI3Ϫ/Ϫ or p35Ϫ/Ϫ by guest on September 25, 2021 heterodimeric cytokine of the IL-12 family, EBI3/p35 (3), that was mice showed reduced capacity to control homeostatic expansion of recently called IL-35 (4). IL-27 signals through a heterodimeric CD4ϩCD25Ϫ effector T (Teff) cells and were less effective than receptor composed of a specific component, IL-27R (also called wild-type Treg cells to cure mice in a model of inflammatory TCCR, WSX-1 or IL-27R␣) and the common chain, gp130 (5). It bowel disease. This defective function was ascribed to a lack of ϩ ϩ acts on multiple cell types including CD4 and CD8 T cells, B IL-35 production, as the alternative partners for EBI3 and p35, cells, NK cells, macrophages, neutrophils, mast cells, and endo- p28, and p40, respectively, are not expressed by mouse Treg cells. thelial cells, and plays important roles in the immune system, spe- In addition, recombinant mouse IL-35 was shown to inhibit the cially in the regulation of Th and inflammatory responses (re- proliferation of mouse Teff cells in vitro (4). In another recent viewed in Refs. 6, 7). study, a single chain mouse IL-35-Fc fusion was demon- Although the in vivo association between EBI3 and p35 was strated to enhance the proliferation of mouse Treg cells, while originally evidenced in human placental extracts (3), it was re- inhibiting the development of Th17 cells (8). The signaling cas- cently shown that IL-35 is constitutively expressed by mouse cade induced by IL-35 and its receptor have not been described. The p35 gene is constitutively expressed at low levels in many *Centre National de la Recherche Scientifique Unite´Mixte de Recherche 8147, Uni- cell types and therefore displays an almost ubiquitous expression versite´Paris Descartes, Paris, France; †Service d’Anatomie Pathologique, Assistance (9). In contrast, EBI3 gene expression is restricted to specific cell Publique Hoˆpitaux de Paris, Universite´ Paris Descartes, Hoˆpital Cochin, Paris, France; and ‡Service d’Anatomie Pathologique, Assistance Publique Hoˆpitaux de types and highly inducible. In previous in situ and in vitro studies Paris, Hoˆpital Trousseau, Paris, France in humans, we found that EBI3 was expressed at high levels in Received for publication May 5, 2008. Accepted for publication September 6, 2008. placental trophoblast cells and activated dendritic cells, and at The costs of publication of this article were defrayed in part by the payment of page lower levels in macrophages and endothelial cells. Regarding lym- charges. This article must therefore be hereby marked advertisement in accordance phocytes, EBI3 expression was readily detected in activated nor- with 18 U.S.C. Section 1734 solely to indicate this fact. mal B cells as well as tumoral B and T cells, but was undetectable 1 E.B. and F.L. share first authorship. ϩ in normal resting CD3 T cells (10–15). However, because Treg 2 Address correspondence and reprint requests to Dr. Odile Devergne, Centre Na- tional de la Recherche Scientifique Unite´Mixte de Recherche 8147, Hoˆpital Necker, cells constitute a minor population among total T cells, EBI3 ex- Baˆtiment Se`vres, 161 rue de Se`vres, Paris, France. E-mail address: odile. pression by this specific T cell subset may have gone unnoticed. [email protected] Therefore, in this study we specifically investigated EBI3 expres- 3 Abbreviations used in this paper: EBI3, EBV-induced gene 3; Treg cell, regulatory sion by human Treg cells to determine whether they can express T cell; Teff cell, effector T cell. IL-35. Our in situ and in vitro analyses indicate that these cells do Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00 not express IL-35. www.jimmunol.org The Journal of Immunology 6899

Materials and Methods cells were treated with brefeldin A (Sigma-Aldrich, 10 ␮g/ml) for the last Human tissues 4 h of the culture. Postnatal normal thymus or tonsil specimens were obtained from children Proliferation assay who underwent corrective cardiac surgery or tonsillectomy, respectively, at ϫ 4 Necker Hospital (Paris). Fetal thymus specimens were collected after mis- To evidence the suppressive activity of Treg cells, Teff cells (5 10 per carriage (Trousseau Hospital, Paris). Non-neoplastic spleens, lymph nodes, well) were cultured in triplicate for 4 days in X-Vivo 15 medium in 96-well ␮ and intestinal tissues were previously described (12, 15). Fresh tissues, plates coated with anti-CD3 (UCHT1, R&D Systems, 2.5 to 5 g/ml) in the ϫ 4 obtained with informed consent of the patient, and fixed tissues, collected presence of 5 10 irradiated (5000 rad) PBMC and of variable numbers of Teff cells. Proliferation was measured by adding 0.5 ␮Ci per well of for histological examination and diagnostic purposes, were studied in ac- 3 cordance with French ethical guidelines. [ H]thymidine (Amersham Biosciences) for the last8hofthefourth day of incubation. Immunohistochemistry and immunocytochemistry Western blot analysis and ELISA Before staining, formalin-fixed paraffin-embedded tissues were subjected to Ag retrieval by heat pretreatment in citrate (pH 6) or Tris/EDTA (pH 9) Cells were washed in ice-cold PBS and lysed for1honiceinlysis buffer buffer. Single immunostaining was performed on serial tissue sections us- (1% Nonidet P-40, 50 mM Tris (pH 7.4), 150 mM NaCl, 3% glycerol, 1.5 ing an indirect avidin-biotin peroxidase or alkaline phosphatase kit (Bio- mM EDTA) supplemented with protease inhibitors (1 mM PMSF, 1 ␮g/ml Genex). In double immunostaining experiments, binding of the primary Ab pepstatin, 1 ␮g/ml leupeptin). Cell lysate was centrifuged for 15 min at in the first label was detected using peroxidase-conjugated anti-mouse En- 13,000 ϫ g and the supernatant was assayed for protein concentration ϩ Vision reagent (DakoCytomation) and diaminobenzidine as chromogen. using the microBCA protein assay (Pierce). Lysates (from 25 to 50 ␮g) Binding of the primary Ab in the second label was detected using an in- were subjected to SDS-PAGE and transferred to nitrocellulose for immu- direct avidin-biotin alkaline phosphatase kit (BioGenex) and Fast Red noblotting. EBI3 and Foxp3 were detected using 2G4H6 and 236A/E7 Downloaded from (DakoCytomation) as chromogen. Sections were counterstained with mouse mAbs, respectively. Actin was detected using goat polyclonal Abs Mayer hematoxylin. EBI3 was detected using 2G4H6 mAb (10) at 2–4 (I19, Santa Cruz Biotechnology). Binding of primary Abs was detected ␮g/ml, and Foxp3 was detected using 236A/E7 mAb (Abcam) at 10–20 with HRP-conjugated anti-mouse (Amersham Biosciences) or anti-goat ␮g/ml. CD25 mAb (clone 4C9, Novocastra) was used at a 1/100 dilution. (Santa Cruz Biotechnology) Abs. Peroxidase reaction was developed with For immunostaining on methanol/acetone (1/1)-fixed cytospin prepara- chemiluminescence reagents (Pierce). EBI3 ELISA was performed as pre- tion, slides were rehydrated in Tris-buffered saline, blocked with Tris-buff- viously described (10). It detects both free EBI3 and EBI3 complexed with ered saline containing 5% human Ig and 1% BSA, and then incubated with either p28 or p35 (detection limit, 1 ng/ml). In some cases, cell culture ϩ control rabbit IgG (Sigma-Aldrich) or rabbit anti-EBI3 IgG (11) (20 ␮g/ supernatants from CD4 T, Teff, or Treg cells stimulated for 2 days with http://www.jimmunol.org/ ml) for 60 min. Ab binding was detected using anti-rabbit peroxidase- beads coated with CD3 and CD28 mAbs were concentrated using Amicon ϩ conjugated EnVision reagent and diaminobenzidine as chromogen. Ultra centrifugal filter unit (Millipore) before being tested by ELISA. Photomicrographs were taken on a Leica DMRB microscope using a 3 CCD color video camera (Sony DXC-950P) using TRIBVN ICS software. RNA extraction and RT-PCR analysis Isolation of human B and T cells RNA was isolated by TRIzol extraction, followed by DNase I digestion and reverse transcription using M-MLV reverse transcriptase and oligo(dT) B and T cells were purified from adult peripheral blood (14 donors) or primer (all reagents from Invitrogen). EBI3 primers were as follows: 5Ј- pediatric tonsil (four donors) by magnetic separation using reagents from GCAGACGCCAACGTCCAC-3Ј (sense) and 5Ј-CCAGTCACTCAGT Miltenyi Biotec. Ј ␤ TCCCCGT-3 (antisense). p35 and 2-microglobulin primers were as pre-

B cells were purified from tonsils by negative selection by depletion of by guest on September 25, 2021 viously described (11, 17). The thermal cycle profile was as follows: 1 min tonsillar mononuclear cells from T/NK cells by incubation with CD2 mi- 94°C, 30 s at 55°C, and 20 to 35 s at 72°C for 30 (␤ -microglobulin), 32 crobeads as described (16). 2 ϩ ϩ (EBI3), or 34 (p35) cycles. CD4 T, CD8 T, or ␥␦ T cells were purified from PBMC isolated by Ficoll-Paque Plus (Amersham Biosciences) gradient centrifugation by us- ing CD4ϩ T cell isolation kit II, CD8ϩ T cell isolation kit II, or ␥␦ mi- Results ϩ crobead kit, respectively. In some cases, CD4 T cells were further sep- Treg cells present in the thymus do not express IL-35 arated into naive or memory T cells by incubation with CD45RO microbeads and separation by using a LS column. The negative fraction First, we investigated by immunohistochemistry whether naturally (naive CD4ϩ T cells) was collected and the positive fraction was passed ϩ occurring Treg cells developing in the thymus express EBI3. Treg again over a MS column to enrich in CD45RO cells. cells can be identified in situ based on their constitutive expression Treg and Teff cells were isolated from PBMC using the CD4ϩCD25ϩ Treg cell isolation kit. In brief, CD4ϩ T cells were first purified by negative of Foxp3 and high expression of CD25. Because the expression of selection and then incubated with CD25 microbeads followed by separa- EBI3 had not been previously analyzed in the human thymus, we tion by using a MS column. The negative fraction (CD4ϩCD25Ϫ T cells) first stained serial sections from two fetal thymuses (18 and 22 wk was collected and the positive fraction was passed over a second MS col- of gestation) and six pediatric thymuses (aged 7 days to 8 mo) with umn to enrich in Treg cells. Purified Treg cells accounted for 1–3% (mean ϩ ϩ ϩ anti-EBI3 and anti-Foxp3 mAbs to determine whether EBI3 cells 1.8%) of the total CD4 T fraction. In some cases, CD4 T cells, Teff and ϩ Treg cells were purified from tonsils. In these cases, tonsillar mononuclear could be detected in the same compartments as Foxp3 cells (Fig. ϩ ϩ cells were first depleted of B cells by incubation with CD19 microbeads 1, A and B). Numerous Foxp3 or EBI3 cells were detected in all and separation on a LS column, and then submitted to purification using the ϩ ϩ cases and their distribution did not vary with the age of the thymus. CD4 CD25 regulatory T cell isolation kit as described above. Both Foxp3ϩ or EBI3ϩ cells were predominantly located in the In each case, purity of the cell separation was verified by immunostain- medulla, although some positive cells were also scattered in the ing and FACS analysis (FACSCalibur or FACSCanto II, BD Biosciences), ϩ and was as follows: B cells: Ͼ 98% CD19ϩ, CD4ϩ T cells: 95–99%, CD8ϩ cortex (Fig. 1, A and B). However, while Foxp3 cells had a lym- T cells: 90–94%, ␥␦ T cells: 90–94%, naive CD4ϩ T cells: 92–97% phoid morphology, most EBI3ϩ cells were morphologically con- CD4ϩCD45RAϩ, memory CD4ϩ T cells: Ͼ97% CD4ϩCD45ROϩ, Teff ϩ Ϫ ϩ ϩ sistent with dendritic cells. Double-staining experiments per- cells: 95–99% CD4 CD25 and Treg cells: 95–99% CD4 CD25 . formed in the eight cases of thymus demonstrated that all Foxp3ϩ In vitro culture cells were negative for EBI3 (Fig. 1, C–E). It also showed that ϩ ϫ 6 some Foxp3 cells in the medulla were in close association with T cell subsets (2.5 10 /ml) were stimulated for 2 days with beads coated ϩ with CD3 and CD28 Abs (10 ␮g/ml each, T cell activation/expansion kit, EBI3 dendritic cells (Fig. 1E). Similarly, when thymic sections Miltenyi Biotec), or PHA (4 ␮g/ml, Roche Diagnostics), in RPMI 1640 were double-stained with anti-EBI3 and anti-CD25 mAbs, no medium supplemented with 10% FBS, L-glutamine, and antibiotics (com- CD25ϩ cell showed positivity for EBI3 (Fig. 1, F–H). As observed plete RMPI medium) or in X-Vivo 15 medium (BioWhittaker), in the ab- ϩ ϩ for Foxp3 cells, some medullar CD25 thymocytes could be ob- sence or presence of IL-2 (Roche Diagnostics, 20–500 U/ml). B cells were ϩ cultured for 2–3 days in complete RPMI 1640 medium in the presence of served in direct contact with EBI3 dendritic cells (Fig. 1H). goat polyclonal CD40 Ab (0.5 ␮g/ml, R&D Systems). In some cases, T These findings are in agreement with a previous in situ analysis 6900 LACK OF IL-35 EXPRESSION BY HUMAN Treg CELLS Downloaded from http://www.jimmunol.org/

FIGURE 1. Immunohistochemical analysis of EBI3 expression by human thymic Treg cells. A and B, Serial sections from a pediatric thymus stained by guest on September 25, 2021 with anti-Foxp3 (A) or anti-EBI3 mAbs (B). Foxp3ϩ or EBI3ϩ cells are abundant and preferentially located in the medulla. C–E, Double staining with anti-Foxp3 and anti-EBI3 Abs in sections from pediatric thymuses of various ages. Foxp3ϩ cells (indicated by arrows, brown nuclear staining) located in the cortex (C) or in the medulla (D and E), are negative for EBI3 (red cytoplasmic staining). In E, a Foxp3ϩ cell in direct contact with a EBI3ϩ cell having F–H, Double staining with anti-EBI3 and anti-CD25 .(ء) extended cytoplasmic processes and morphologically consistent with a dendritic cell is shown mAbs in sections from pediatric thymuses. CD25ϩ cells (indicated by arrows, red plasma membrane staining) present in the cortex (F) or in the medulla (G and H) are negative for EBI3 (brown cytoplasmic staining). In H, two examples of direct interaction between a CD25ϩ lymphoid cell and an EBI3ϩ dendritic cell are shown. In one case, this interaction takes place close to a Hassall’s corpuscle consistent with a previous study showing its role to instruct dendritic cells to induce Treg cells (18). On each picture, the age of the thymus is indicated at the bottom left. Co, Cortex; Me, medulla; Ha, Hassall’s corpuscle. Objective, ϫ10 in A and B, ϫ40 in C–H. showing that human thymic CD25ϩCTLA4ϩ Treg cells preferen- digestive tissues (large and small bowel, appendix), Foxp3ϩ cells tially localized within the thymic medulla in close association with were detected in the T cell zones surrounding follicles and activated dendritic cells (18). Importantly, they indicate that Treg were also scattered in the lamina propria. Again, these cells were cells developing in the human thymus do not express IL-35, inde- in all cases negative for EBI3 (Fig. 2, E–H). This lack of EBI3 pendently of the age of the thymus. detection in Foxp3ϩ cells was not due to a lack of sensitivity of

ϩ EBI3 immunohistochemistry, because many other cell types in- Lack of IL-35 expression by Foxp3 cells in peripheral tissues cluding dendritic cells, macrophages, B cell blasts, and plasma Next, we analyzed the in situ expression of EBI3 by Treg cells cells were positive for EBI3 in these tissues (Fig. 2). Thus, as present in peripheral tissues. To this end, various human tissues observed for natural thymic Treg cells, mature Treg cells present that had been previously analyzed for EBI3 expression by immu- in peripheral tissues do not express detectable amounts of EBI3. In nohistochemistry and included lymph nodes (n ϭ 5), tonsils (n ϭ these tissues, no double-staining with anti-EBI3 and anti-CD25 3), spleens (n ϭ 5), and digestive tissues (n ϭ 4) (12, 15) were mAbs was performed to investigate the expression of EBI3 by ϩ subjected to double staining with anti-Foxp3 and anti-EBI3 Abs. In Treg cells because of the presence of many CD25 cells others lymph nodes and tonsils, Foxp3ϩ cells were mainly located in the that Treg cells that could express EBI3, such as activated B cells. T cell zones, but were also present at the border of follicles and inside follicles in agreement with previous studies (19). However, Treg cells purified from adult peripheral blood do not whatever their localization, Foxp3ϩ cells were not costained with constitutively express IL-35 anti-EBI3 Ab (Fig. 2, A–D, and data not shown). In spleens, To further investigate the expression of IL-35 by human Treg cells, Foxp3ϩ cells were largely restricted to the peri-arteriolar T cell we examined whether Treg cells present in peripheral blood ex- areas and similarly were negative for EBI3 (data not shown). In press EBI3. For this purpose, CD4ϩ T cells were isolated from The Journal of Immunology 6901

FIGURE 2. Immunohistochemical analysis of EBI3 expression by Foxp3ϩ cells in peripheral human tis- sues. Double staining with anti-Foxp3 and anti-EBI3 mAbs in sections from a reactive lymph node (A–D), an ap- pendix (E), and small and large bowel (F–H). Foxp3ϩ cells (brown nuclear staining) are indicated by arrows. In lymph nodes, these cells are abundant in interfollicular areas (A and D), but are also found in follicles, crossing the mantle zone (B) and scattered in- side follicles (C). In the intestine, Foxp3ϩ cells are located around B follicles of the GALT (E–G) or scat- tered in the lamina propria (H). In all cases, Foxp3ϩ cells are negative for EBI3, although numerous EBI3ϩ cells (red cytoplasmic staining) including dendritic cells and macrophages (A, Downloaded from D–G), follicular B cells (B), and plasma cells (H) are detected. IF, In- terfollicular area; Ma, mantle zone; Fo, follicle. Objective, ϫ20 in A, ϫ40 in B–H. http://www.jimmunol.org/

PBMC and further separated into CD4ϩCD25Ϫ T cells and EBI3, was observed (Fig. 3C, lane 13). Thus, in vitro analysis of CD4ϩCD25ϩ T cells by magnetic cell separation. Treg cells can purified CD4ϩCD25ϩFoxp3ϩ Treg cells confirmed that they do be distinguished from activated Teff cells transiently expressing not constitutively express EBI3. CD25 by surface staining with CD127 (IL-7R␣), a marker ex- pressed at high levels on activated Teff cells, but at low levels on Analysis of constitutive EBI3 expression in various human Treg cells (20, 21). FACS analysis of purified CD4ϩCD25ϩ cells T cell subsets showed a Ն95% purity based on CD4 and CD25 expression and In mice, EBI3 and p35 mRNA have been shown to be constitu- the presence of 4–6% CD127high cells, indicating that at least 94% tively coexpressed, not only by Treg cells, but also, albeit at a by guest on September 25, 2021 of CD4ϩCD25ϩ T cells purified from blood have a Treg pheno- lower level, by CD8ϩ and ␥␦ T cells (4). Thus, to determine type (Fig. 3A). whether the discrepancy observed between humans and mice for In vitro functional studies confirmed that purified CD4ϩCD25ϩ IL-35 expression is restricted to Treg cells or also extends to other T cells were bona fide Treg cells. Indeed, consistent with a func- T cell subsets, we analyzed by Western blot the expression of EBI3 tional Treg phenotype, CD4ϩCD25ϩ T cells were anergic in re- in various human T cell subsets purified from peripheral blood by sponse to a polyclonal stimulation as shown by their proliferation magnetic cell separation (three to five donors tested for each sub- level that corresponded to 2–9% of that of Teff cells tested in set). In the cell lysates from ␥␦ T cells, CD8ϩ T cells, or naive or parallel, and inhibited Teff cell proliferation in a dose dependent- memory CD4ϩ T cells, no constitutive expression of EBI3 was manner, with very strong suppression (from to 81–92% inhibition) detected (Fig. 4, lanes 1–7). Thus, in contrast to the situation ob- when Teff cells and Treg cells were cocultured at a 1:1 ratio served in mice, none of the human T cell subsets we analyzed (Fig. 3B). (Treg and Teff, naive and memory CD4ϩ T, CD8ϩ T, and ␥␦ T Cell lysates from blood-purified Treg cells were then analyzed cells) expressed detectable amounts of EBI3 constitutively. by Western blot for Foxp3 and EBI3 expression, in parallel with ϩ cell lysates from whole CD4ϩ T and Teff cells (Fig. 3C) (three Analysis of EBI3 expression in in vitro-stimulated CD4 T cell different donors tested). As a positive control for EBI3, purified subsets tonsillar B cells, either freshly isolated or stimulated with anti- In a previous study, we showed that low levels of EBI3 could be CD40 Abs, were used (four different donors). B cells were chosen induced in human T cells upon in vitro stimulation with PHA and as a positive control, because in mice the level of EBI3 gene ex- IL-2 (14). Therefore, to further analyze the expression of EBI3 in pression in Treg cells is comparable to that observed in B cells (4). human T cells, whole CD4ϩ T cells or subsets of naive or memory In cell lysates from CD4ϩCD25ϩ T cells (lanes 5 and 10), a strong CD4ϩ T, Treg, and Teff cells were stimulated for 2 days with signal for Foxp3 was specifically observed further assessing their beads coated with anti-CD3 and anti-CD28 Abs (a more physio- Treg phenotype. Consistent with a previous study (15), EBI3 was logical stimulus than PHA) in the absence or presence of IL-2, and readily detected in the lysates of freshly isolated B cells (lanes 1 cell lysates were analyzed for EBI3 expression by immunoblotting and 6) and its expression increased upon in vitro B cell stimulation (Fig. 5) (six different donors tested for non-Treg CD4ϩ T cells and (lanes 2 and 7). In contrast, EBI3 was undetectable in cell lysates five for Treg cells). In all CD4ϩ T cell subsets, except Treg cells, from whole CD4ϩ T cells, Teff and Treg cells, even after a pro- CD3/CD28 stimulation resulted in EBI3 induction. In these cells, longed exposure of the blot (Fig. 3C, lanes 3–5, 8–10) (detection EBI3 levels were higher than those observed following PHA/IL-2 limit for EBI3 blot, 0.2 ng/lane). Similarly, when Treg cells were stimulation (Fig. 5A, lanes 12 and 13) and were not further en- purified from tonsils (two different donors tested) and their cell hanced by the addition of IL-2 (Fig. 5A, left panel and 5B, right lysates analyzed by Western blot, expression of Foxp3, but not of panel). EBI3 levels in CD3/CD28-activated CD4ϩ T cells were 6902 LACK OF IL-35 EXPRESSION BY HUMAN Treg CELLS

FIGURE 3. Analysis of EBI3 expression in purified Treg cells. A, Cell surface phenotype of purified Treg cells. CD4ϩ T cells were isolated from PBMC and sep- arated into CD25ϩ and CD25Ϫ fractions by magnetic cell sorting. Dot plot analysis (log fuorescence intensity on both axes) of the different fractions labeled with CD4-FITC, CD25-PE, and CD127-Alexa Fluor 647 Ab are shown. The percentage of cells in each quadrant or gate is indicated. A representative experiment is shown. B, Functional characterization of purified Treg cells. CD4ϩCD25Ϫ Teff cells and CD4ϩCD25ϩ Treg cells were cultured in triplicates in CD3 Ab-coated 96-well plates, either separately or mixed together at the indi- cated ratio, in the presence of irradiated PBMC. Prolif- eration was measured by adding [3H]thymidine for the last8hofthefourth day of incubation. Results are expressed in cpm (mean of triplicates Ϯ SD) and are representative of three experiments. C, Western blot analysis. Cell lysates from whole CD4ϩ T cells, Teff Downloaded from cells and Treg cells purified from blood (two different donors are shown) or tonsils were analyzed by immu- noblotting with anti-Foxp3 and anti-EBI3 mAbs. Anti- actin Abs were used as a control for equal total protein loading. As positive control for EBI3, cell lysates from tonsillar B cells (a different donor is shown on each http://www.jimmunol.org/ blot), either freshly isolated (B cells, lanes 1 and 6)or stimulated with anti-CD40 Abs (act. B cells, lanes 2 and 7) were used. For Foxp3, two major bands correspond- ing to the two isoforms present in human are detected. A barely detectable Foxp3 signal was occasionally seen in the cell lysate from total CD4ϩ T cells (data not shown). by guest on September 25, 2021

variable depending on the donor (Fig. 5A, left blot), and were but still no detectable EBI3 expression (Fig. 5B, lane 8). Thus, in consistently much lower than those observed in activated B cells both resting or CD3/CD28-stimulated Treg cells, no expression of (Fig. 5, A and B). In contrast, no EBI3 induction was observed in EBI3 was detected by Western blot. CD3/CD28-activated Treg cells from all donors tested (Fig. 5B, To further investigate EBI3 regulation in resting and in vitro lanes 2 and 7). Addition of IL-2 to Treg cell culture resulted in activated human Treg and Teff cells, EBI3 expression was ana- increased cell viability (data not shown) and enhanced expression lyzed by immunocytochemistry and by ELISA of the cell culture of Foxp3 (Fig. 5B, right blot), in keeping with previous data (22), supernatant. Intracellular staining for FACS analysis could not be performed because of the lack of suitable anti-human EBI3 Abs. By immunostaining of cytospin samples, EBI3 was detected only in CD3/CD28-activated Teff cells (Fig. 6b), but not in freshly pu- rified Teff or Treg cells (Fig. 6, a and d, respectively) nor in ac- tivated Treg cells (Fig. 6, e and f), in agreement with immuno- blotting data. By ELISA, levels of secreted EBI3 in the cell culture supernatants of T cells were undetectable, unless supernatants were concentrated by 8-fold. Under these conditions, from 3 to 4.4 ng/ml EBI3 (mean, 3.5 ng/ml) were detected in the supernatants of CD3/CD28-activated CD4ϩ T cells or Teff cells, but concentrated supernatants from activated Treg cells remained negative for EBI3 (two to three donors tested for each T cell subsets). Thus, at both intracellular or secreted levels, no EBI3 was detected in resting or activated human Treg cells. FIGURE 4. Analysis of constitutive EBI3 expression in various human In murine Treg cells, EBI3 expression was investigated both at ␮ T cell subsets. Cell lysates (40 g/lane) from various purified T cell subsets the mRNA and protein levels (4). Thus, to verify that the lack of (indicated at the top of each lane) were analyzed by Western blot with EBI3 detection in human Treg cells is not due to a technical issue anti-EBI3 mAb and subsequently with anti-actin Abs to verify equal total protein loading. For ␥␦ and CD8ϩ T cells, lysates from two different do- (lower sensitivity of the protein analysis compared with the mRNA nors are shown. As positive controls, lysates from tonsillar B cells, either analysis) or to dissociation between EBI3 mRNA and protein ex- ϩ freshly isolated (lane 8) or stimulated for 2 days with anti-CD40 Abs (lane pression in humans, we tested various CD4 T cells subsets for 9) were used. EBI3 expression by RT-PCR. In parallel, p35 gene expression was The Journal of Immunology 6903

FIGURE 7. RT-PCR analysis of EBI3 and p35 gene expression in var- ious human CD4ϩ T cell subsets. Purified CD4ϩ T cells (A), Treg and Teff cells (B), either freshly purified (Ϫ) or stimulated for 40 h with CD3/CD28 Ab coated beads (ϩ) were analyzed for EBI3 and p35 gene expression by Downloaded from RT-PCR. The lane denoted cDNA (Ϫ) corresponds to the negative control ␤ ␤ where no cDNA was added in the PCR. 2-microglobulin ( 2m) was used FIGURE 5. Analysis of EBI3 expression in activated CD4ϩ T cell sub- ϩ as an internal control for equal material in each reaction. cDNA from sets. A, Cell lysates from whole, naive, or memory CD4 T cells, either freshly isolated B cells (lane 1) was used as a positive control. Each braket Ϫ freshly isolated ( ) or stimulated for 2 days with beads coated with anti- indicates cells from a same donor processed in parallel (for example, two ϩ ϩ ϩ CD3 and anti-CD28 Abs ( ), alone or in presence of 20 U/ml IL-2 ( different donors are shown for resting and activated CD4 T cells on A, and http://www.jimmunol.org/ IL-2) or with PHA plus IL-2, were analyzed by Western blot with anti- for activated Teff and Treg cells, on B). The lower mobility band present EBI3 and anti-actin Abs. As positive controls, cell lysates from B cells, in p35 PCR in resting CD4ϩ T cells on A, was not consistently detected. either freshly isolated or stimulated for 2 days with anti-CD40 Abs, were ϩ On B, lane 7, a faint p35 band was visible on the gel, but is barely detect- used. For CD4 T cells, lysates obtained from two different donors are able on the scanned image. displayed to show the variable magnitude of EBI3 induction. B, Cell ly- sates from Teff and Treg cells purified from the same donor (two different donors are shown) and cultured in parallel as indicated in A, were analyzed stimulated CD4ϩ T cells (Fig. 7A). Also in agreement with the by Western blot with anti-EBI3, anti-actin, and in some cases anti-Foxp3 protein data, EBI3 transcripts were undetectable in freshly purified Abs. No induction of EBI3 was observed in activated Treg cells indepen- Teff or Treg cells (Fig. 7B, lanes 3 and 4) and were detected in dently of the donor. Similar results were observed when Treg cells were activated Teff cells, but not activated Treg cells (Fig. 7B, lanes by guest on September 25, 2021 cultured with higher doses of IL-2 (100 to 500 U/ml). Also, no EBI3 signal 6–9). Interestingly, p35 signal was much stronger in resting Treg was detected in Treg cells when they were treated with brefeldin A (10 cells than in resting Teff cells (Fig. 7B, compare lane 4 with lane ␮g/ml) for the last4hoftheculture (data not shown). 3), this up-regulation being specifically observed in resting but not activated Treg cells. This latter finding is in keeping with the data analyzed (Fig. 7). In CD4ϩ T cells, EBI3 gene expression was described in mouse. Indeed, constitutive expression of p35 gene undetectable before stimulation, but was consistently induced upon has been shown to be much higher in mouse Treg cells than in CD3/CD28 stimulation, in keeping with our protein data. In con- mouse Teff cells, and to be down-regulated in Treg cells upon trast, p35 gene expression was detected in both unstimulated or CD3/CD28 stimulation (4). Taken together, these data showed no dichotomy for EBI3 between mRNA and protein expression, and confirmed the lack of EBI3 expression in human Treg cells. They also pointed out that while EBI3 gene expression appeared to differ between human and mouse Treg cells, p35 gene seemed to be similarly regulated among human and mouse Treg cells.

Discussion In this study, we investigated whether human Treg cells can ex- press IL-35 protein. Because immunoassays to detect EBI3 ex- pression are more sensitive than those to detect p35, we focused our analysis on EBI3 expression and initially used immunohisto- chemistry and Western blotting, two assays that in our hands were the most sensitive to detect EBI3 expression. Although in this and previous studies (10–15), we could readily detect EBI3 expression in many normal or tumoral cell types, we failed to detect signifi- cant expression of EBI3 by human Treg cells in all compartments FIGURE 6. EBI3 immunostaining of purified Teff and Treg cells. Cy- analyzed (thymus, spleen, lymph node, tonsil, intestine, and tospin preparations of Teff (a—c) and Treg (d—f) cells, either freshly pu- rified or stimulated for two days with beads (appearing as small brown blood). Other protein assays (ELISA, immunocytochemistry) or spheres on the figure) coated with CD3/CD28 mAbs were stained with RT-PCR analysis similarly did not show EBI3 expression in hu- rabbit anti-EBI3 IgG or control rabbit IgG (Col) as indicated on each panel. man Treg cells, constitutively or upon stimulation. Therefore, Only activated Teff cells are positive for EBI3 (brown cytoplasmic stain- these cells are unlikely to express substantial amounts of the two ing). Objective: ϫ100. EBI3-associated heterodimeric , IL-35 or IL-27. 6904 LACK OF IL-35 EXPRESSION BY HUMAN Treg CELLS

For the in situ analysis of EBI3 expression by natural or periph- blast and extravillous trophoblast cells at the human fetal-maternal eral Treg cells, we used Foxp3 to identify these cells. In mice, interface (10), a site of strong immunomodulation. Further studies Foxp3 expression is strictly restricted to Treg cells and therefore of IL-35 expression profile in human and mouse, in both normal constitutes an exclusive marker of murine Treg cells (23). In con- and pathological conditions, will be necessary to delineate its role trast, in humans, Foxp3 is constitutively expressed at high levels in the regulation of the immune response and its potential interest by Treg cells, but is also transiently expressed at low levels by in for therapeutic manipulation. vitro-activated Teff cells (Fig. 5B, and Ref. 24–28). Although some authors observed a conversion to a Treg phenotype (24), Acknowledgments others did not (25–27). Therefore, although little information is We thank the Departments of Otorhinolaryngology, Pediatric Cardiac Sur- available on the extent of Foxp3 expression by in vivo activated gery, Hematology, and Pathology of Necker Hospital (Paris, France) for ϩ Teff cells in tissues, it is likely that all Foxp3 cells detected in situ tissues and blood samples, and the Department of Pathology of Cochin ϩ are not Treg cells. Notably, in human tonsils, Foxp3 cells have Hospital (Paris, France) for technical assistance. We also thank Ame´de´e been found not only among resting CD69ϪCD4ϩCD25ϩ T cells, Renan for technical advice. but also, with a lower frequency, among activated CD69ϩ CD4ϩCD25ϩ T cells comprising activated Teff cells (19). Never- Disclosures theless, because in the different tissues we analyzed, none of the The authors have no financial conflict of interest. Foxp3ϩ cells expressed EBI3, regardless of their intensity of ϩ Foxp3 staining, we can conclude that Foxp3 Treg cells do not References express EBI3 and therefore IL-35. Also, in the human thymus, 1. Devergne, O., M. Hummel, H. Koeppen, M. M. Le Beau, E. C. Nathanson, Downloaded from both CD4ϩCD25ϩ and CD8ϩCD25ϩ regulatory thymocytes ex- E. Kieff, and M. Birkenbach. 1996. A novel interleukin-12 p40-related protein pressing Foxp3 mRNA have been described (29–30). Our data induced by latent Epstein-Barr virus infection in B lymphocytes. J. Virol. 70: ϩ 1143–1153. suggest that neither of these two Foxp3 suppressive subsets 2. Pflanz, S., J. C. Timans, J. Cheung, R. Rosales, H. Kanzler, J. Gilbert, L. Hibbert, expressed EBI3. T. Churakova, M. Travis, E. Vaisberg, et al. 2002. IL-27, a heterodimeric cyto- kine composed of EBI3 and p28 protein, induces proliferation of naive CD4ϩ T This study also highlights the difference between humans and cells. Immunity 16: 779–790. mice regarding EBI3 expression in T cells. In mice, gene microar- 3. Devergne, O., M. Birkenbach, and E. Kieff. 1997. Epstein-Barr virus-induced http://www.jimmunol.org/ rays identified EBI3 as a gene selectively overexpressed in Treg gene 3 and the p35 subunit of form a novel heterodimeric hema- topoietin. Proc. Natl. Acad. Sci. USA 94: 12041–12046. cells (4, 31). Accordingly, EBI3 protein was specifically detected 4. Collison, L. W., C. J. Workman, T. T. Kuo, K. Boyd, Y. Wang, K. M. Vignali, by intracellular staining in mouse Treg but not Teff cells (4). In R. Cross, D. Sehy, R. S. Blumberg, and D. A. Vignali. 2007. The inhibitory addition, transduction of mouse Teff cells with a Foxp3 retroviral cytokine IL-35 contributes to regulatory T-cell function. Nature 450: 566–569. 5. Pflanz, S., L. Hibbert, J. Mattson, R. Rosales, E. Vaisberg, J. F. Bazan, vector resulted in strong EBI3 gene induction (4), while con- J. H. Phillips, T. K. McClanahan, R. de Waal Malefyt, and R. A. Kastelein. 2004. versely, conditional ablation of Foxp3 in mature peripheral Treg WSX-1 and glycoprotein 130 constitute a signal-transducing receptor for IL-27. cells resulted in the down-regulation of EBI3 gene expression (32), J. Immunol. 172: 2225–2231. 6. Kastelein, R. A., C. A. Hunter, and D. J. Cua. 2007. Discovery and biology of indicating that mouse EBI3 was a Foxp3 target gene. Our findings IL-23 and IL-27: related but functionally distinct regulators of inflammation. suggest that in humans, Foxp3 expression is not sufficient to in- Annu. Rev. Immunol. 25: 221–242. by guest on September 25, 2021 7. Batten, M., and N. Ghilardi. 2007. The biology and therapeutic potential of in- duce significant EBI3 expression, which further emphasizes the terleukin 27. J. Mol. Med. 85: 661–672. differences observed between human and mouse Foxp3 (23). 8. Niedbala, W., X. Q. Wei, B. Cai, A. J. Hueber, B. P. Leung, I. B. McInnes, and This differential expression of EBI3 between humans and mice F. Y. Liew. 2007. IL-35 is a novel cytokine with therapeutic effects against collagen-induced arthritis through the expansion of regulatory T cells and sup- is not restricted to Treg cells, but is also observed in other T cell pression of Th17 cells. Eur. J. Immunol. 37: 3021–3029. subsets. Whereas low levels of EBI3 transcripts were detected in 9. Trinchieri, G. 2003. Interleukin-12 and the regulation of innate resistance and mouse CD8ϩ and ␥␦ T cells, no EBI3 expression was detected in adaptive immunity. Nat. Rev. Immunol. 3: 133–146. 10. Devergne, O., A. Coulomb-L’Hermine´, F. Capel, M. Moussa, and F. Capron. both subsets in humans. Conversely, whereas CD3/CD28 stimu- 2001. Expression of Epstein-Barr virus-induced gene 3, an interleukin-12 p40- lation failed to induce EBI3 expression in murine Teff cells and related molecule, throughout human pregnancy: involvement of syncytiotropho- down-regulated its constitutive expression in Treg cells (4), it con- blasts and extravillous trophoblasts. Am. J. Pathol. 159: 1763–1776. 11. Coulomb-L’Hermine´, A., F. Larousserie, S. Pflanz, E. Bardel, R. Kastelein, and sistently induced EBI3 expression in human Teff cells. In contrast, O. Devergne. 2007. Expression of interleukin-27 by human trophoblast cells. similar to mouse Treg cells, human Treg cells expressed high Placenta 28: 1133–1140. amounts of p35 transcripts, constitutively. In addition, RT-PCR 12. Larousserie, F., S. Pflanz, A. Coulomb-L’Hermine´, N. Brousse, R. Kastelein, and O. Devergne. 2004. Expression of IL-27 in human Th1-associated granulomatous analysis showed that both EBI3 and p35 were expressed in diseases. J. Pathol. 202: 164–171. activated human Teff cells, which therefore may produce IL-35. 13. Niedobitek, G., D. Pa¨zolt, M. Teichmann, and O. Devergne. 2002. Frequent expression of the Epstein-Barr virus (EBV)-induced gene, EBI3, an IL-12 p40- However, coimmunoprecipitation experiments performed from the related cytokine, in Hodgkin and Reed-Sternberg cells. J. Pathol. 198: 310–316. cell lysate of activated Teff cells using our anti-EBI3 Abs and 14. Larousserie, F., E. Bardel, S. Pflanz, B. Arnulf, C. Lome-Maldonado, various commercial anti-p35 Abs failed so far to show substantial O. Hermine, L. Bre´gaud, M. Perennec, N. Brousse, R. Kastelein, and O. Devergne. 2005. Analysis of interleukin-27 (EBI3/p28) expression in Epstein- association between both (our unpublished observations). Barr virus- and human T-cell leukemia virus type 1-associated lymphomas: het- The recent identification of immunosuppressive functions for erogeneous expression of EBI3 subunit by tumoral cells. Am. J. Pathol. 166: IL-35 led to the hypothesis that this novel cytokine may represent 1217–1228. 15. Larousserie, F., E. Bardel, A. Coulomb-L’Hermine´, D. Canioni, N. Brousse, a new target for immunotherapy, notably in inflammatory diseases R. A. Kastelein, and O. Devergne. 2006. Variable expression of Epstein-Barr (4, 8). However, because of the differential regulation of EBI3 virus-induced gene 3 during normal B-cell differentiation and among B-cell lym- expression in T cells, IL-35 may play different roles in human and phomas. J. Pathol. 209: 360–368. 16. Larousserie, F., P. Charlot, E. Bardel, J. Froger, R. A. Kastelein, and mouse. In humans, various cell types other than Treg cells might O. Devergne. 2006. Differential effects of IL-27 on human B cell subsets. J. Im- constitute potential sources of IL-35. Indeed, strong expression of munol. 176: 5890–5897. 17. De Waal Malefyt, R., C. G. Figdor, R. Huijbens, S. Mohan-Peterson, B. Bennett, EBI3, in the absence of detectable expression of p28, has been J. Culpepper, W. Dang, G. Zurawski, and J. E. de Vries. 1993. Effects of IL-13 detected in tumoral cells of various forms of lymphomas (14, 15). on phenotype, cytokine production, and cytotoxic function of human monocytes: Production of IL-35 by these cells might constitute a mechanism to comparison with IL-4 and IFN-␥ or IL-10. J. Immunol. 151: 6370–6381. 18. Watanabe, N., Y. H. Wang, H. K. Lee, T. Ito, Y. H. Wang, W. Cao, and Y. J. Liu. evade the anti-tumoral immune response. Also, in previous studies, 2005. Hassall’s corpuscles instruct dendritic cells to induce CD4ϩCD25ϩ regu- we showed that EBI3 and p35 were expressed by syncytiotropho- latory T cells in human thymus. Nature 436: 1181–1185. The Journal of Immunology 6905

19. Lim, H. W., P. Hillsamer, A. H. Banham, and C. H. Kim. 2005. Cutting edge: 26. Gavin, M. A., T. R. Torgerson, E. Houston, P. deRoos, W. Y. Ho, direct suppression of B cells by CD4ϩCD25ϩ regulatory T cells. J. Immunol. A. Stray-Pedersen, E. L. Ocheltree, P. D. Greenberg, H. D. Ochs, and 175: 4180–4183. A. Y. Rudensky. 2006. Single-cell analysis of normal and FOXP-3 mutant human 20. Liu, W., A. L. Putnam, Z. Xu-Yu, G. L. Szot, M. R. Lee, S. Zhu, P. A. Gottlieb, T cells: FOXP3 expression without regulatory T cell development. Proc. Natl. P. Kapranov, T. R. Gingeras, B. Fazekas de St. Groth, et al. 2006. CD127 ex- Acad. Sci. USA 103: 6659–6664. pression inversely correlates with FoxP3 and suppressive function of human 27. Tran, D. Q., H. Ramsey, and E. M. Shevach. 2007. Induction of FOXP3 expres- CD4ϩ T reg cells. J. Exp. Med. 203: 1701–1711. sion in naive human CD4ϩFOXP3Ϫ T cells by T-cell receptor stimulation is 21. Seddiki, N., B. Santner-Nanan, J. Martinson, J. Zaunders, S. Sasson, A. Landay, transforming growth-factor-ß-dependent but does not confer a regulatory pheno- M. Solomon, W. Selby, S. I. Alexander, R. Nanan, et al. 2006. Expression of type. Blood 110: 2983–2990. interleukin (IL)-2 and IL-7 receptors discriminates between human regulatory 28. Roncarolo, M. G., and S. Gregory. 2008. The validity of Foxp3 to define human and activated T cells. J. Exp. Med. 203: 1693–1700. and mouse regulatory T cells. Eur. J. Immunol. 38: 925–927. 22. Zorn, E., E. A. Nelson, M. Mohseni, F. Porcheray, H. Kim, D. Litsa, R. Bellucci, 29. Annunziato, F., L. Cosmi, F. Liotta, E. Lazzeri, R. Manetti, V. Vanini, E. Raderschall, C. Canning, R. J. Soiffer, D. A. Frank, and J. Ritz. 2006. IL-2 P. Romagnani, E. Maggi, and S. Romagnani. 2002. Phenotype, localization, and regulates FOXP3 expression in human CD4ϩCD25ϩ regulatory T cells through mechanism of suppression of CD4ϩCD25ϩ human thymocytes. J. Exp. Med. a STAT-dependent mechanism and induces the expansion of these cells in vivo. 196: 379–387. Blood 108: 1571–1579. 30. Cosmi, L., F. Liotta, E. Lazzeri, M. Francalanci, R. Angeli, B. Mazzinghi, 23. Ziegler, S. F. 2006. FOXP3: of mice and men. Annu. Rev. Immunol. 24: 209–226. V. Santarlasci, R. Manetti, V. Vanini, P. Romagnani, et al. 2003. Human 24. Walker, M. R., D. J. Kasprowicz, V. H. Gersuk, A. Benard, M. Van Landeghen, CD8ϩCD25ϩ thymocytes share phenotypic and functional features with J. H. Buckner, and S. F. Ziegler. 2003. Induction of Foxp3 and acquisition of T CD4ϩCD25ϩ regulatory thymocytes. Blood 102: 4107–4114. regulatory activity by stimulated human CD4ϩCD25Ϫ T cells. J. Clin. Invest. 31. Gavin, M. A., J. P. Rasmussen, J. D. Fontenot, V. Vasta, V. C. Manganiello, 112: 1437–1443. J. A. Beavo, and A. Y. Rudensky. 2007. Foxp3-dependent programme of regu- 25. Allan, S. E., L. Passerini, R. Bacchetta, N. Crellin, M. Dai, P. C. Orban, latory T-cell differentiation. Nature 445: 771–775. S. F. Ziegler, M. G. Roncarolo, and M. K. Lewings. 2005. The role of the 2 32. Williams, L. M., and A. Y. Rudensky. 2007. Maintenance of the Foxp3-depen- FOXP3 isoforms in the generation of human CD4ϩ Tregs. J. Clin. Invest. 115: dent developmental program in mature regulatory T cells requires continued ex- 3276–3284. pression of Foxp3. Nat. Immunol. 8: 277–284. Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021