Combination Therapy Using IL-2/IL-2 Monoclonal Antibody Complexes, Rapamycin, and Islet Autoantigen Peptides Increases Regulatory T Cell Frequency and This information is current as Protects against Spontaneous and Induced of September 29, 2021. Type 1 Diabetes in Nonobese Diabetic Mice Jean N. Manirarora and Cheng-Hong Wei

J Immunol published online 19 October 2015 Downloaded from http://www.jimmunol.org/content/early/2015/10/17/jimmun ol.1402540

Supplementary http://www.jimmunol.org/content/suppl/2015/10/17/jimmunol.140254 http://www.jimmunol.org/ Material 0.DCSupplemental

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists by guest on September 29, 2021

• Fast Publication! 4 weeks from acceptance to publication

*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 All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published October 19, 2015, doi:10.4049/jimmunol.1402540 The Journal of Immunology

Combination Therapy Using IL-2/IL-2 Monoclonal Antibody Complexes, Rapamycin, and Islet Autoantigen Peptides Increases Regulatory T Cell Frequency and Protects against Spontaneous and Induced Type 1 Diabetes in Nonobese Diabetic Mice

Jean N. Manirarora and Cheng-Hong Wei

Regulatory T cells (Treg) play a crucial role in the maintenance of self-tolerance. In this study, we sought to expand Ag-specific Tregs in vivo and investigate whether the expanded Tregs can prevent or delay the development of type 1 diabetes (T1D) in the NOD mouse model. NOD mice were treated with a combination of IL-2/anti–IL-2 Ab complex, islet Ag peptide, and rapamycin. After the Downloaded from combined treatment, CD4+CD25+Foxp3+ Tregs were significantly expanded in vivo, they expressed classical Treg markers, exerted enhanced suppressive functions in vitro, and protected against spontaneous development of T1D in NOD mice. Moreover, treated mice were almost completely protected from the adoptively transferred, aggressive form of T1D caused by in vitro–activated cytotoxic islet Ag-specific CD8 T cells. Protection from T1D was transferrable by Tregs and could be attributed to reduced islet infiltration of immune cells as well as the skewing of the immune response toward a Th2 cytokine profile. This new method of peripheral immune regulation could potentially contribute to development of novel immunotherapeutic strategies to prevent the http://www.jimmunol.org/ development of T1D or to promote tolerance to islet transplants without using immunosuppressive drugs for long terms. The Journal of Immunology, 2015, 195: 000–000.

ype 1 diabetes (T1D) represents a failure of the immune autoimmunity-mediated destruction (11). b cell regeneration system to maintain self-tolerance, resulting in loss of therapies via genetic reprogramming have also been intensively T pancreatic islet b cells via T cell–mediated autoimmunity pursued as a therapy; however, a recent study shows that the ge- (1). T cells reactive to islet Ags such as insulin, the 65-kDa iso- netically reprogrammed, liver-derived insulin-producing cells are form of glutamic acid decarboxylase (GAD65), and islet-specific also susceptible to autoimmune destruction in settings of murine by guest on September 29, 2021 glucose-6-phosphatase catalytic subunit–related protein (IGRP) models of T1D (12). Therefore, it is critical to understand the play pivotal pathogenic roles during the development of T1D in recurrent autoimmune mechanisms leading to b cell destruction NOD mice (2, 3) and in humans (4–8). and to find immunotherapeutic strategies that specifically target There is currently no cure for T1D, and the disease is treated with autoimmune anti-islet responses to achieve tolerance to islet Ags life-long exogenous insulin replacement. Allogeneic islet trans- without long-term use of immunosuppressive drugs. plantation has been performed experimentally in animal models One effective mechanism of peripheral tolerance involves the and humans with promising results, but this investigational mea- control of autoreactive T cells by naturally occurring CD4+ sure is not yet practical in regular clinical practice. Moreover, CD25+Foxp3+ regulatory T cells (nTregs). Tregs can potently allogeneic islet transplants can provoke immune response, and suppress the activation and effector function of other immune long-term immunosuppressive drugs are needed to protect the cells, including CD4 and CD8 T cells, B cells, NK cells, mac- transplanted islets (9, 10). Even when the transplanted islets are rophages, and dendritic cells (13). Previous studies in mice have from monozygotic twins, they are still susceptible to the recurrent shown that nTregs play a major role in the control of a variety of autoimmune diseases (14, 15). Recently, Tregs have also been investigated in clinical trials to prevent/treat T1D, graft-versus- Gene Transfer and Immunogenicity Branch, Division of Cellular and Gene Thera- pies, Office of Cellular, Tissue, and Gene Therapies, Center for Biologics Evaluation host diseases, and organ transplant rejection (16). Notably, it is and Research, U.S. Food and Drug Administration, Silver Spring, MD 20993 well established in several models that Ag-specific Tregs are Received for publication October 6, 2014. Accepted for publication August 31, 2015. more effective than polyclonal Tregs in immunosuppression This work was supported by the Food and Drug Administration Modernizing Science in vitro and control of autoimmunity in vivo (e.g., T1D, arthritis, grant program, as well as by the Division of Cellular and Gene Therapies. J.N.M. was multiple sclerosis, arthritis, colitis) (17–24); particularly, in- supported through a fellowship administered by the Oak Ridge Institute for Science jection of small numbers of Ag-specific Tregs selectively ex- and Education. panded in vitro can reverse diabetes even after disease onset (17, Address correspondence and reprint requests to Dr. Cheng-Hong Wei, Food and Drug Administration, Center for Biologics Evaluation and Research, 10903 New Hamp- 18). However, whether Ag-specific Tregs can be expanded in vivo, shire Avenue, Building 52/72, Room 3104, Silver Spring, MD 20993-0002. E-mail whether these cells retain their immunosuppressive function fol- address: [email protected] lowing expansion, and the stability of their Foxp3 expression have The online version of this article contains supplemental material. not yet been elucidated. Abbreviations used in this article: GAD65, 65-kDa isoform of glutamic acid decar- IL-2 is an important factor for the survival, expansion, and boxylase; GITR, glucocorticoid-induced TNFR; IGRP, islet-specific glucose-6- phosphatase catalytic subunit–related protein; nTreg, naturally occurring regulatory function of Tregs in vivo (25–28). The administration of IL-2/anti– T cell; Tconv, conventional T cell; T1D, type 1 diabetes; Treg, regulatory T cell. IL-2 mAb (clone JES6-1) complexes in vivo has been shown to

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1402540 2 IN VIVO EXPANSION OF Tregs PREVENTS DIABETES IN NOD MICE increase polyclonal Treg numbers in lymphoid organs (29). It has T cell activation and CFSE assay also been reported that the mammalian target of rapamycin inhib- + For T cell proliferation assay, CD8 T cells were purified and labeled with itor rapamycin can selectively inhibit proliferation of CD4 effector CFSE (Molecular Probes, Eugene, OR) according to the manufacturer’s T cells, enhance the expansion of Tregs (30, 31), and maintain the instruction; the CFSE-labeled Thy1.1+CD8+ 8.3 T cells (1 3 105/well in stability of expanded Tregs (32, 33). Therefore, we hypothesized 96-well round-bottom plates) were cultured for 3 d at 37˚C in 5% CO2 in that a combined treatment comprising an islet Ag (BDC2.5 cognate the presence of anti-CD3/CD28 Dynabeads (Life Technologies, Grand Island, NY) (bead/T cell ratio 2:1) and murine IL-2 (PeproTech) at 2 3 103 mimetope), IL-2/IL-2 mAb complexes, and rapamycin could act in U/ml with or without treated CD4+CD25+ Tregs or untreated CD4+CD25+ synergy to expand Ag-specific Tregs in lymphoid organs. Tregs at a series of titrated ratios (from 1:1 to 1:32 of Treg/8.3 T cell ratio). In this study, we explored this new method to expand Ag-specific After 3 d, cells were harvested then labeled with fluorochrome-conjugated Tregs in NOD mice. The phenotype and function of expanded anti-CD4, anti-CD8, anti-CD90.1 (Thy1.1), and anti-CD90.2 (Thy1.2) Abs (BD Biosciences, San Diego, CA) for 20 min at 4˚C and then washed Tregs, as well as their capacity to protect against spontaneous and twice. Cells were analyzed by flow cytometry using a FACSCanto II induced diabetes development in NOD mice, were investigated. (Becton Dickinson, Palo Alto, CA). For T cell activation before adoptive transfer, CD8+ 8.3 T cells (2 3 6 10 /well in 24-well plates) were cultured for 3 d at 37˚C in 5% CO2 in Materials and Methods the presence of anti-CD3/CD28 Dynabeads (Life Technologies) at Mice a 1:2 cell/bead ratio and murine IL-2 (PeproTech) at 2 3 103 U/ml. Female NOD/ShiLtj, NOD SCID, CD8 TCR transgenic NOD 8.3, CD4 TCR After 3 d, activated T cells were harvested, separated from the beads transgenic NOD BDC2.5, and NOD.Thy1.1+/+ mice were purchased from using a magnet, and adoptively transferred (i.v.) into 6-wk-old recipient NOD mice at 2 3 106 Tcellspermouse. The Jackson Laboratory (Bar Harbor, ME). The BDC2.5 TCR transgenic

+ g7 Downloaded from CD4 T cells specifically recognize the I-A –restricted epitope derived from Flow cytometry analysis an islet Ag (34), whereas the 8.3 TCR transgenic CD8+ T cells specifically d recognize the K -restricted IGRP206–214 epitope derived from the islet Ag For T cell subset analysis, cells were labeled with fluorochrome-conjugated 2 2 IGRP (35). NOD Thy1.1+/ BDC2.5 and NOD Thy1.1+/ 8.3 mice were anti-CD4, anti-CD8, anti-CD69, anti-CD62L, anti-CD44, and anti– bred in a specific pathogen-free facility at the Food and Drug Administration glucocorticoid-induced TNFR (GITR) Abs (BD Biosciences) for 20 min at Center for Biologics Evaluation and Research (Bethesda, MD). All of the 4˚C and then washed twice. mouse colonies were maintained according to Institutional Animal Care and For non–T cell subset analysis, the cells were labeled with fluorochrome-

Use Committee guidelines, and all animal protocols and procedures were conjugated anti-TCRab, anti-DX5, anti-CD11b, anti-CD11c, and anti- http://www.jimmunol.org/ approved by the Institutional Animal Care and Use Committee. CD19 Abs (BD Biosciences) for 20 min at 4˚C and then washed + + twice. Expansion of CD4 CD25 regulatory cells in vivo For intracellular Foxp3 and CTLA-4 labeling, cells were labeled with IL-2/anti–IL-2 mAb (JES6-1) complexes were prepared as previously re- fluorochrome-conjugated anti-CD4, anti-CD8, and anti-CD25 Abs (BD Biosciences), fixed and permeabilized, and then labeled with either ported (36). Briefly, 100 mg anti–IL-2 mAb (eBioscience, San Diego, CA; fluorochrome-conjugated anti-Foxp3 (eBioscience) or fluorochrome- clone JES6-1, catalog no. 16-7022-85) was mixed in vitro with recombi- conjugated anti-CTLA-4 (eBioscience) Abs, according to the manu- nant murine IL-2 (20 mg) (PeproTech, Rocky Hill, NJ) in 200 ml HBSS (Mediatech, Herndon, VA) supplemented with 1% heat- inactivated FBS facturers’ instructions. (HyClone, Logan, UT), and incubated at 4˚C for 30 min. After incubation, For intracellular cytokine staining, splenocytes were stimulated for 6 h the volume of the mixture was brought to 2 ml with HBSS/1% FBS. with a ready-to-use polyclonal cell activation mixture containing the by guest on September 29, 2021 Six-week-old female NOD mice were treated with one i.v. injection of PMA, a calcium ionophore (ionomycin), and a protein transport inhibi- tor (brefeldin A) (BD Parmingen, San Diego, CA). Cells were labeled 100 mg BDC2.5 TCR mimetope peptide (37) (amino acid sequence with fluorochrome-conjugated anti-CD4 and anti-CD8 Abs, fixed and per- RVRPLWVRME) (Food and Drug Administration/Center for Biologics Evaluation and Research/Foundation for Biomedical Research core facil- meabilized, and then labeled with fluorochrome-conjugated anti–IL-10, g ity, Bethesda, MD) at day 0 and three daily i.p. injections (days 0, 1, and 2) anti–IL-4, anti–IL-2 (BD Biosciences), and anti–IFN- (eBioscience) Abs of IL-2/anti–IL-2 mAb complexes (1 mg IL-2 with 5 mg IL-2 mAb in according to the manufacturers’ instructions. Cells were analyzed using a volume of 100 ml) and 4 mg rapamycin (Sigma-Aldrich, St. Louis, MO). a FACSCanto II (Becton Dickinson) flow cytometer. For simplicity, this treatment with BDC2.5 mimetope peptide, IL-2/anti– BDC/I-Ag7 tetramer and tetramer staining IL-2 mAb complexes, and rapamycin will be referred to as standard combined treatment afterward. Each injection was a 100-ml volume. Mice The PE-labeled BDC/I-Ag7 tetramer (I-Ag7/RTRPLWVRME) was syn- were euthanized 3 d after the last injections (day 5) to assess the phenotype thesized by the National Institutes of Health Tetramer Core Facility at and functions of the expanded regulatory T cells. In some experiments, Emory University (Atlanta, GA). Cells from the spleens or pancreatic g7 other islet autoantigen peptides were used, such as I-A GAD65524–543 lymph nodes were stained with the BDC tetramer and then with surface (amino acid sequence SRLSKVAPVIKARMMEYGTT) (38) and I-Ag7 markers (CD4, CD25), followed by intracellular staining for Foxp3. insulin B chain 9–23 (amino acid sequence SHLVEALYLVCGERG) (39). Additionally, a control non-islet autoantigen peptide I-Ag7 myelin basic Assessment of diabetes incidence protein 74–85 (also called 1B3 cognate peptide) (amino acid sequence SRPGLCHMYKDS) (40) was also used. The NOD mice (females, 4–6 wk old) were randomly assigned into different groups using GraphPad software (http://www.graphpad.com/ Isolation of CD4+CD25+ Tregs and CD4+CD252 responder quickcalcs/randomize1.cfm). All mice were tested weekly for urine T cells glucose starting at 10 wk of age using Keto-Diastix reagent strips for urinalysis (Bayer, Elkhart, IN). When the urine glucose measurement Lymphocytes were obtained from lymph nodes and spleens and processed in reached 300 mg/dl the mice were tested for blood glucose until 32 wk of HBSS supplemented with 2% FBS. CD4+ T cells were first negatively se- age using blood glucose Accu-Check Aviva Plus test strips (Roche, lected and purified, then the CD4+CD25+ and CD4+CD252 subsets were Indianapolis, IL) and an Accu-Check blood glucose meter (Roche). Mice purified using PE–anti-CD25 Ab and anti-PE magnetic beads (Miltenyi were designated as diabetic when blood glucose measurement was .300 Biotec, Auburn, CA) according to the manufacturer’s instructions. CD4+ mg/dl for 2 consecutive weeks. CD25+ regulatory cell purity was consistently .90% as assessed by staining with anti-CD4, anti-CD25 Abs followed by flow cytometry analysis. Prevention of spontaneous diabetes Isolation of CD8+ 8.3 T cells Six-week-old female NOD mice were left untreated or treated with standard combined treatment. Urine/blood glucose was assessed weekly Lymphocytes were obtained from lymph nodes and spleens of NOD 8.3 starting at 10 wk. mice, or NOD Thy1.1+/2 8.3 mice, and processed in HBSS supplemented Alternatively, 6-wk-old female NOD mice (or NOD BDC2.5 mice where with 2% FBS. CD8+ T cells were negatively selected and purified using the indicated) were left untreated or treated with standard combined treatment. mouse CD8+ T cell isolation kit (Miltenyi Biotec) according to the man- Three days after the last injections, the mice were sacrificed and CD4+ ufacturer’s instructions. The purity of CD8+ 8.3 T cells was usually .90%, CD25+ Tregs or CD4+CD252 effector cells were purified from lymph as determined by flow cytometry. nodes and spleens. Then, 1 3 106 Tregs or effector T cells were adoptively The Journal of Immunology 3 transferred (i.v.) into 6-wk-old female NOD mice. Urine/blood glucose H&E staining and histological assessment was assessed weekly starting at 10 wk. Mice with two consecutive blood glucose reading of .300 mg/dl were defined as diabetic. Harvested pancreatic tissues were fixed in 10% formalin/PBS and em- bedded in paraffin, and 5-mm sections were prepared. Sections were stained Prevention of diabetes induced by adoptive transfer of with H&E and examined by microscopy. activated CD8+ 8.3 T cells Statistical analysis Six-week-old NOD mice were left untreated or treated with standard combined treatment. At day 5, the mice were injected i.v. through the tail Data are presented as means 6 SD and were analyzed using GraphPad vein with 2 3 106 CD8+ 8.3 T cells activated for 3 d in vitro at 37˚C in 5% Prism 5 software (GraphPad Software, La Jolla, CA). Statistical anal- CO2 in the presence of anti-CD3/CD28 Dynabeads (Life Technologies) ysis was performed using a Student t test for comparison between two (bead/T cell ratio 2:1) and murine IL-2 (PeproTech) at 2 3 103 U/ml. groups and ANOVA for more than two groups. A p value ,0.05 was Urine/blood glucose was assessed daily. considered significant. T1D disease incidence data were plotted using Downloaded from

FIGURE 1. The combined treat- ment significantly expands Tregs in NOD mice. CD4+ T cells were iso- lated from 6- to 8-wk-old Thy1.1+ BDC2.5 NOD mice, labeled with

CFSE, and i.v. injected into 6- to http://www.jimmunol.org/ 8-wk-old NOD mice at 3 3 106/mouse dosage. Mice were left untreated or treated with BDC2.5 peptide, IL-2/IL-2 mAb, and rapamycin for 3 d (days 0, 1, and 2). At day 5 cells were isolated from lymphoid organs and labeled with anti-CD4, anti-CD8, anti-Thy1.1, and anti-CD25 Abs, permeablized and fixed, and then stained with anti-Foxp3 by guest on September 29, 2021 Ab and analyzed by FACS gating on endogenous CD4+Foxp3+cells (A and B), transferred Thy1.1+CD4+Foxp3+ cells (C and D), or transferred and di- vided CFSElowThy1.1+CD4+Foxp3+ cells (E and F). Except the rapamycin group (n = 2), each group contained five to eight samples compiled from two independent experiments. Purified CD4+CD252 effector T cells (1 3 106) and CD4+CD25+ Tregs (1 3 106)from Thy1.1 mice were injected separately into two groups of NOD mice. Re- cipient NOD mice were treated with BDC2.5 peptide, IL-2/IL-2 mAb, and rapamycin (n = 3 mice/group) for 3 d. On day 5 cells were isolated from lymphoid organs and labeled with anti- CD4, anti-CD8, anti-Thy1.1, and anti- CD25 Abs and then stained with anti-Foxp3 Ab and analyzed by FACS (G and H). The results are the repre- sentative of two independent experi- ments with similar findings. *p , 0.05, **p , 0.01, ***p , 0.001. 4 IN VIVO EXPANSION OF Tregs PREVENTS DIABETES IN NOD MICE Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 2. Transferred effector cells do not covert to Tregs, and the treatment enriches Ag-specific Tregs. (A) Thy1.1+CD4+CD252 T cells (Tconvs) and Thy1.1+CD4+CD25+ T cells (Tregs) were purified from Thy1.1+ BDC2.5 NOD mice, CFSE labeled, and transferred to recipient NOD mice (n = 3/group); the recipient mice were treated with IL-2 complexes, rapamycin, and BDC2.5 cognate peptide for 3 d as described in Materials and Methods. On day 5, mice were sacrificed and the cells were isolated from spleen, stained with Abs against Thy1.1, CD4, and Foxp3, and analyzed by FACS. Cells are gated on CD4 and Thy1.1 double-positive cells to view only the transferred T cells. Shown are representative data of two independent experiments with similar results. (B) Female NOD mice were either left untreated or treated with standard combined treatment (n = 6/group). On day 5, mice were sacrificed and the cells were isolated from spleens, stained with Abs against CD4, CD25, Foxp3, and BDC/I-Ag7 tetramer, and analyzed by FACS. Shown are data compiled from two independent experiments. (C) Female NOD mice were either left untreated or treated with the indicated treatment (n = 3/group). On day 5, mice were sacrificed and the cells were isolated from pancreatic lymph nodes, stained with Abs against CD4, CD25, Foxp3, and BDC/I-Ag7 tetramer, and analyzed by FACS. **p , 0.01, ****p , 0.0001. The Journal of Immunology 5 the Kaplan–Meier plot, and the statistical analysis was performed using T cells was less marked in the spleen (Fig. 1A, 1C, 1E) compared the log rank analysis. with the pancreatic lymph node (Fig. 1B, 1D, 1F). Additionally, only the combination of BDC2.5 peptide, IL-2/IL-2 mAb com- Results plexes, and rapamycin yielded a significant difference in Treg ex- Tregs are expanded in recipients after the combined treatment pansion in the pancreatic lymph node compared with untreated To test the hypothesis that a combination of islet Ag peptide, IL-2/ control (Fig 1B, 1D, 1F). In the pancreas, the combined treatment anti–IL-2 mAb complexes, and rapamycin can selectively expand also significantly expanded Tregs compared with untreated control Ag-specific Tregs in vivo, we first evaluated the effect of different (Supplemental Fig. 1). Taken together, these data indicate the treatments on Treg expansion in various lymphoid organs. With one possibility of Ag-specific stimulation and expansion of Tregs by the i.v. injection of BDC2.5 peptide at day 0 and three daily i.p. combined treatment. We therefore decided to use this combination injections (days 0, 1, and 2) of IL-2 /anti–IL-2 mAb complexes and of IL-2/IL-2 mAb complexes, repamycin, and islet peptide Ag as rapamycin we observed substantial expansion of CD4+Foxp3+ standard combined treatment in subsequent experiments. T cells on day 5 after start of treatment. Both endogenous CD4+ To understand whether the observed increase of Tregs is due to Foxp3+ Tregs and transferred Thy1.1+ BDC2.5 TCR transgenic effector T cell conversion into Tregs or simply due to the expansion CD4+Foxp3+ T cells expanded equally in the spleen (Fig. 1A, 1C, of nTregs, or both, CFSE-labeled Thy1.1+CD4+CD25+ T cells 1E), pancreatic lymph nodes (Fig. 1B, 1D, 1F), and mesenteric (Tregs) and CFSE-labeled Thy1.1+CD4+CD252 T cells (effector lymph nodes (data not shown). The expansion of CD4+Foxp3+ T cells) were adoptively transferred into two separate groups of Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 3. Phenotype of the expanded Tregs in treated mice. Spleen cells were isolated from NOD mice untreated or treated with BDC2.5 peptide, IL-2/IL-2 mAb,andrapamycinfor3d(n = 2/group). Cells were labeled with anti-CD4, anti-CD25, anti-GITR, and anti-CD62L Abs then with anti-Foxp3 or anti–CTLA-4 Abs and analyzed by FACS (A–I). Results shown are representative of one of three independent experiments with similar results. *p , 0.05, **p , 0.01, ***p , 0.001. 6 IN VIVO EXPANSION OF Tregs PREVENTS DIABETES IN NOD MICE

Table I. Effect of treatment on the activation of conventional CD4+ T cells and CD8+ T cells

Activation/Memory Markers among Activation/Memory Markers among CD4+ T Cells CD8+ T Cells

Untreated Treated p Value Untreated Treated p Value % CD62Lhigh cells 58.98 6 5.65 53.6 6 0.85 .0.05 60.7 6 5.94 57.02 6 5.33 .0.05 % CD69+ cells 5.21 6 0.08 4.17 6 0.01 .0.05 1.38 6 0.52 1.71 6 0.49 .0.05 % CD44+ cells 15.12 6 2.42 29.88 6 1.88 ,0.05 18.03 6 5.3 50.83 6 2.96 ,0.05 (T memory) % CD44+CD62Lhigh cells 5.19 6 0.67 17.53 6 2.76 ,0.05 14.97 6 2.65 16.73 6 1.11 .0.05 (T central memory) Spleen cells were isolated from NOD mice untreated or treated with BDC2.5 peptide, IL-2/IL-2 mAb, and rapamycin (n = 3/group). Cells were labeled with anti-CD4, anti-CD8, anti-CD44, anti-CD62L, and anti-CD69 Abs and analyzed by FACS. The results are the representative of three different experiments with similar findings.

NOD mice, followed by the standard combined treatment. As CD62L (Fig 3I). After treatment, the expression level of Foxp3 is shown in Fig. 1G and 1H, the Tregs expanded more robustly in the increased in Tregs (Fig. 3B, 3E). Taken together, these data in- Treg transfer group than the group that received T effector cells, dicate that the expanded Tregs expressed classical markers asso- suggesting that the combined treatment preferentially leads to ciated with nTreg phenotype and function. substantial expansion of Tregs rather than the conversion of CD4+ Downloaded from 2 Effect of treatment on other immune cell subsets CD25 T conventional cells into Tregs. Furthermore, as shown in Fig. 2A, when CFSE-labeled Thy1.1+ It has been reported that injection of low doses of IL-2 in complex CD4+CD25+ BDC2.5 T cells (Tregs) were adoptively transferred with the anti–IL-2 neutralizing mAb JES6-1 leads to selective into NOD mice and treated with the combined treatment, all of expansion of Tregs and minimal or no change in other cell subsets the transferred CD4+CD25+ Tregs divided and remained Foxp3+. and that short-term treatment with IL-2/anti–IL-2 mAb (JES6-1) + + 2 In contrast, when CFSE-labeled Thy1.1 CD4 CD25 BDC2.5 complex increases Tregs up to 6-fold (29). The increase in Tregs is http://www.jimmunol.org/ T cells (conventional T cells [Tconvs]) were adoptively trans- specific to this IL-2/anti–IL-2 mAb complex, as complexing IL-2 ferred into NOD mice and treated with the combined treatment, with other anti–IL-2 neutralizing Ab clones, for example, S4B6, very few (∼5%) of the transferred CD4+CD252 T cells became drives expansion of CD8+ T cells and NK cells (36). We therefore Foxp3+, strongly suggesting that effector T cells do not convert tested whether our combined treatment has similar effects on into Foxp3+ cells under the condition of combined treatment in different immune cell subsets. As shown in Table I, the combined this model system. Therefore, we think that the increase of Tregs treatment did not affect the activation status of either CD4+ or both in number and percentage after the combined treatment (IL- CD8+ T cells. With regard to T cell memory pool, the combined 2/anti–IL-2 mAb complexes, rapamycin, and BDC2.5 peptide) treatment increased the percentage of both CD4 and CD8 T cells observed in the present study is mainly a result of the expansion of with memory phenotype (CD44+). Also, the combined treatment by guest on September 29, 2021 nTregs and not the conversion of effector T cells into Tregs. led to an increased percentage of CD44+CD62Lhi central memory To assess whether the combined treatment indeed enhances Ag- CD4+ T cells but had no effect on CD8+ T cells with central specific Tregs or not, after treatment, Tregs were stained with H-2g7 memory phenotype (Table I), suggesting a differential respon- BDC tetramer. As shown in Fig. 2B and 2C, the result confirmed that siveness to cytokine stimulation between CD4+ and CD8+ T cells. the combined treatment significantly enriched islet Ag-specific In agreement with previous reports (29), the combined treatment CD4+CD25+Foxp3+ Tregs in the spleens and the pancreatic had minimal or no effect on the percentage of non–T cell immune 2 2 lymph nodes. However, note that among the significantly ex- subsets, namely NK cells (TCRab DX5+), B cells (TCRab 2 panded Tregs in the spleens and pancreatic lymph nodes, the CD19+), dendritic cells (TCRab CD11c+), and macrophages 2 frequency of the BDC mimotope-specific Tregs was relatively low, (TCRab CD11b+) (Table II). only 0.4–0.5%. In other words, a large proportion of the expanded The suppressive function of the expanded Tregs is enhanced Tregs was not Ag specific. both in vitro and in vivo The expanded Tregs express classical Treg markers nTregs suppress the proliferation of CD4+CD252 T cells in re- Next, we investigated the phenotype of Tregs expanded by the sponse to CD3 ligation (41). In the present study, we tested the combined treatment 2 d after the last injections (day 5) using flow ability of Tregs expanded by the combined treatment to inhibit the cytometry analysis. As shown in Fig. 3, the expanded Tregs were proliferation of the diabetogenic CD8+ 8.3 T cells activated by found to express molecules that are crucial to Treg function such CD3/CD28 ligation for 3 d. The suppressive function of the ex- as Foxp3 (Fig. 3A–F), CTLA-4 (Fig. 3G), GITR (Fig. 3H), and panded Tregs in inhibiting 8.3 T cell proliferation (as measured by

Table II. Effect of treatment on some other immune cell subsets (NK cells, B cells, dendritic cells, and macrophages)

Percentage in the Spleen Absolute Numbers (3106) in the Spleen

Untreated Treated p Value Untreated Treated p Value NK cells (TCRab2DX5+) 2.5 6 0.5 5.5 6 0.5 .0.05 1.38 6 0.15 4.16 6 0.35 ,0.05 B cells (TCRab2CD19+) 42.0 6 6.0 40.5 6 0.5 .0.05 22.59 6 3.27 32.52 6 0.29 .0.05 Dendritic cells (TCRab2CD11c+) 3.2 6 0.2 4.95 6 0.15 .0.05 1.75 6 0.11 3.96 6 0.13 ,0.05 Macrophages (TCRab2CD11b+) 0.75 6 0.5 0.85 6 0.5 .0.05 0.41 6 0.05 0.71 6 0.03 .0.05 Spleen cells were isolated from NOD mice untreated or treated with BDC2.5 peptide, IL-2/IL-2 mAb, and rapamycin (n = 3/group). Cells were labeled with anti-CD3, anti-DX5, anti-CD19, anti-CD11b, and anti-CD11c Abs and analyzed by FACS. The results are representative of three different experiments with similar findings. The Journal of Immunology 7

CFSE dilution) was higher compared with Tregs from untreated mice (Fig. 4A). Similarly, the expanded Tregs were significantly more potent at blocking homeostatic proliferation of naive CD4+ CD25+ T cells in lymphopenic NOD SCID mice (Fig. 4B, 4C) compared with Tregs from untreated mice. Taken together, these data demonstrated that the combined treatment significantly en- hanced the immunosuppressive function of Tregs. Combined treatment significantly inhibits the development of spontaneous and adoptively transferred diabetes in NOD mice Previous studies have shown that Tregs control the progression of diabetes and that diabetes onset in NOD mice may correlate with impaired Treg functions and/or numbers (reviewed in Ref. 42). Daily injections of IL-2/anti–IL-2 mAb complexes into predia- betic NOD mice for 5 d followed by twice a week maintenance treatment for a total duration of 10 wk prevented the onset of diabetes in these mice (43). We tested the effect of a very short- term (3 d) treatment of NOD mice with the combined treatment on

the prevention of spontaneous diabetes development. As expected, Downloaded from 60% control mice progressively developed diabetes starting at 10 wk of age whereas ∼80% treated mice remained disease-free for .30 wk of age (Fig. 5A). Next, we tested the effect of the treatment on the development of diabetes induced by adoptive transfer of activated islet Ag-specific CD8 T cells (in this study the

islet Ag IGRP-specific 8.3 T cells). NOD mice were treated for 3 d http://www.jimmunol.org/ or left untreated with the combined treatment. At day 5 all the mice from both groups were injected with 2 3 106 in vitro–acti- vated CD8+ 8.3 T cells and the mice were monitored daily for blood glucose. As shown in Fig. 5B, .80% of treated mice remained disease-free up to 42 d after treatment whereas all control mice became severely sick (blood glucose . 600 mg/dl) and were sacrificed for humane reasons by day 7 after treatment. To investigate whether the observed protection in treated mice

was due to the expanded Tregs and was therefore transferable, by guest on September 29, 2021 NOD mice were injected with Tregs isolated from untreated or treated NOD BDC2.5 mice, or with Tregs from treated or untreated NOD mice, or effector T cells from untreated NOD mice. By 30 wk of age, we observed that only 10% of both the control mice and the mice that received untreated effector T cells from NOD mice remained diabetes-free (Fig. 5C). Polyclonal Tregs from untreated NOD mice provided temporary marginal protection, but Tregs from treated NOD mice provided significantly greater protection from T1D for a longer period (p , 0.05); treated and untreated FIGURE 4. Suppressive function of the expanded Tregs in vitro and Tregs from NOD BDC2.5 mice (clonal islet Ag-specific Tregs) in vivo. (A) In vitro suppression of CD8+ 8.3 T cell proliferation by treated were most potent at preventing spontaneous diabetes development and untreated Tregs: CFSE-labeled Thy1.1+CD8+8.3+ cells (1 3 105/well in 96-well round-bottom plates) were cultured for 3 d with or without compared with polyclonal Tregs from NOD mice (Fig. 5C). These treated CD4+CD25+ cells or with untreated CD4+CD25+ cells at a 1:1 ratio data corroborate previous findings that ex vivo–expanded islet Ag- of Tregs/8.3 cells. The number of 8.3 T cells was fixed, and the number of specific Tregs provide better protection of NOD mice from T1D CD4+CD25+ Tregs was titrated down to cover a ratio of Tregs/8.3 T cells than polyclonal natural Tregs (17, 18). Importantly, our data ex- from 1:1 to 1:32. After 3 d, cells were harvested then labeled with anti- tend the previous findings in that the islet Ag-specific Tregs are CD4, anti-CD8, anti-Thy1.1, and anti-Thy1.2 Abs. The percentage divided expanded in vivo. cells (CFSElo cells) was analyzed by FACS among the gated CD8+Thy1.1+ At 30 wk of age, ,20% of the mice that received untreated population (8.3 T cells). Shown are means and SE of duplicate samples. Tregs from NOD mice remained disease-free compared with 70% The differences between the untreated Treg and treated Treg groups are of the mice that received untreated Tregs from NOD BDC2.5 TCR statistically significant (p , 0.05, paired t test). (B) Homeostatic expansion 2 2 trangenic mice (Fig. 5C; p , 0.05). Overall, treated Tregs from of conventional CD4+CD25 T cells: 1 3 106 Thy1.1+CD4+CD25 cells (Tconvs) were purified from untreated Thy1.1+ NOD mice and the Tconvs NOD BDC2.5 mice provided the best protection; at 32 wk of age, were injected into female NOD SCID mice (n = 3 mice/group). At day 7, cells were isolated from spleens and labeled with anti-CD4, anti-CD8, and anti-Thy1.1 Abs and analyzed by FACS to quantitate the transferred were coinjected at 1:1 ratio into female NOD SCID mice (n = 3 mice/ Thy1.1+CD4+ Tconvs. (C) Inhibition of homeostatic expansion of con- group). At day 7, cells were isolated from spleens and labeled with anti- ventional CD4+ T cells by treated Tregs: 1 3 106 CD4+CD25+ Tregs were CD4, anti-CD8, and anti-Thy1.1 Abs and analyzed by FACS to quantitate purified from untreated NOD mice or NOD mice treated with the com- the transferred Thy1.1+CD4+ Tconvs. Shown are representative results of bined treatment for 3 d, and 1 3 106 Thy1.1+CD4+CD252 cells (Tconvs) two independent experiments with similar results. The groups were com- were purified from untreated Thy1.1+ NOD mice; the Tregs and Tconvs pared using ANOVA. *p , 0.05, **p , 0.01, ***p , 0.001. 8 IN VIVO EXPANSION OF Tregs PREVENTS DIABETES IN NOD MICE Downloaded from FIGURE 5. Combined treatment prevents diabetes development in NOD mice, and the protection can be transferred by Tregs. (A) Female NOD mice at6wkofage(n = 19/group) were left untreated or treated with BDC2.5 peptide, IL-2/IL-2 mAb, and rapamycin. Urine/blood glucose was assessed weekly starting at 10 wk. Shown are pooled results of three independent experiments. (B) NOD mice were left untreated (n = 15) or treated (n =15)with BDC2.5 peptide, IL-2/IL-2 mAb, and rapamycin. At day 5 the mice were injected with 2 3 106 CD8+ 8.3 T cells activated in vitro for 3 d with anti- CD3/CD28 beads. Urine/blood glucose was assessed daily. The data shown are pooled from two independent experiments. (C)CD4+CD252 cells (Tconvs) were isolated from untreated NOD mice, and CD4+CD25+ Tregs were purified from untreated NOD, treated NOD, and untreated and treated NOD BDC2.5 mice. One million Tregs or Tconvs were injected (i.v.) into female NOD mice at 6 wk of age. A group of untreated NOD mice (n =22) http://www.jimmunol.org/ was used as control (no cell transfer). Below are the total numbers of recipient NOD mice per group: n = 22 for treated BDC Treg group, n =7for untreated BDC Treg group, n = 20 for treated NOD Treg group, n = 35 for untreated NOD Treg group, and n = 15 for NOD untreated Tconv group. Urine/blood glucose was assessed weekly starting at 10 wk. The data shown are pooled from three independent experiments. The percentage diabetes- free mice were plotted using a Kaplan–Meier curve and log rank statistical analysis was performed. (D) NOD mice were left untreated (n = 10) or treated (n = 10) with BDC2.5 peptide (or insulin B9–23 peptide or GAD65524–543 peptide as indicated), IL-2/IL-2 mAb, and rapamycin for 3 d. At day 5 the mice were injected with 2 3 106 CD8+ 8.3 T cells activated in vitro for 3 d with anti-CD3/CD28 beads. Urine/blood glucose was assessed daily. *p , 0.05, **p , 0.01, ***p , 0.001.

90% of the mice that received treated Tregs from NOD BDC2.5 Treatment with IL-2/IL-2 mAb complexes or rapamycin alone by guest on September 29, 2021 mice remained disease-free (Fig. 5C). Taken together, these data does not provide optimal protection from T1D in NOD mice indicate that the observed protection is due to Tregs, and that Ag- We further evaluated whether the single component of the treatment specific–treated Tregs (whether from treated NOD mice or treated can achieve similar protection. As shown in Fig. 6A, rapamycin NOD BDC2.5 mice) are more potent in vivo compared with un- alone did not protect NOD mice from spontaneous T1D develop- treated polyclonal Tregs. ment; similarly, IL-2/anti–IL-2 mAb complexes alone provided Because in this study we used a mimotope peptide (BDC2.5) in- some protection, but it was not as efficient and long-lasting as the stead of a bona fide islet Ag peptide such as insulin B9–23 (39) or one provided by the combined treatment (Fig. 6B). Combination of GAD65524–543 (38), we tested whether, in our setting, natural islet Ag rapamycin and IL-2/anti–IL-2 mAb complexes provided efficient peptides would also be efficient at preventing the development of and long-lasting protection (Fig. 6C). Consistent with the previous induced T1D. NOD mice were left untreated or treated for 3 d with report (43), continuous, 10-wk-long treatment with IL-2/anti–IL-2 either one of the following treatments: 1) rapamycin, IL-2/anti–IL-2 mAb complexes alone provided 80% protection from T1D devel- mAb complexes; 2) rapamycin, IL-2/anti–IL-2 mAb complexes, opment in female NOD mice (Fig. 6D). BDC2.5 peptide; 3) rapamycin, IL-2/anti–IL-2 mAb complexes, in- Effects of treatment on histology of pancreas and cytokine sulin B9–23 peptide; and 4) rapamycin, IL-2/anti–IL-2 mAb com- profiles plexes, GAD65524–543 peptide. At day 5 all of the mice from all five groups (n = 10/group) were To understand the mechanism of protection in our treated groups, injected with 2 3 106 CD8+ 8.3 T cells activated in vitro for 3 d we tested the effect of the combined treatment on cytokine profiles and the mice were monitored daily for blood glucose. As shown in in both CD4+ and CD8+ T cells. As shown in Fig. 7A and 7B, Fig. 5D, the mimotope peptide (BDC2.5) and natural peptides the treatment skewed the immune profile toward a Th2 response, + (insulin B9–23 and GAD65524–543) provided similar levels of pro- mostly owing to an increased percentage of CD4 T cells producing tection (70–80% disease-free) that were significantly higher (p , IL-4 and IL-10, suggesting that the lower ratio of IFN-g/IL-10 and 0.01) compared with the level of protection in untreated mice IFN-g/IL-4 production in T cells (Table III) may play a role in the (10%). However, no significant difference was found between the observed protection from T1D mediated by treated Tregs. mimotope peptide and natural islet Ag peptides or between these Histology of pancreata from different groups was evaluated as three peptides and the group comprising rapamycin and IL-2/anti– shown in representative results in Fig. 7C, demonstrating that the IL-2 mAb complexes alone. Taken together, these data indicate mice that received treated Tregs had less severe islet infiltrates, that in our setting the combined treatment comprising either compared with the mice in untreated groups. a mimotope peptide (BDC2.5) or a natural islet Ag peptide (in- We also examined the expression of Foxp3 in the Tregs trans- sulin or GAD65) equally protects against induced T1D develop- ferred from treated NOD mice to other NOD mice. As shown in ment in NOD mice. Fig. 7D, even after 3 wk, the Tregs from the treated group still The Journal of Immunology 9 Downloaded from

FIGURE 6. IL-2/IL-2 Ab complexes or rapamycin alone does not provide optimal protection of NOD mice from T1D. (A) Female NOD mice at 6 wk of age (n = 10/group) were left untreated or treated with rapamycin. Urine/blood glucose was assessed weekly starting at 10 wk. (B) Female NOD mice at 6 wk http://www.jimmunol.org/ of age (n = 10/group) were left untreated or treated with IL-2/IL-2 mAb complexes daily for three times. Urine/blood glucose was assessed weekly starting at 10 wk. (C) Female NOD mice at 6 wk of age (n = 10/group) were left untreated or treated with IL-2/anti–IL-2 mAb complexes and rapamycin daily for three times. Urine/blood glucose was assessed weekly starting at 10 wk. (D) Following the previously reported protocol (43), female NOD mice at 6 wk of age (n = 10/group) were left untreated or treated with daily injections of IL-2/IL-2 mAb complexes for 5 d, subsequently followed by twice a week maintenance treatment for a total duration of 10 wk. Urine/blood glucose was assessed weekly starting at 10 wk. *p , 0.05, **p , 0.01. stably express Foxp3, even at a slightly higher percentage than the numbers of CD4+Foxp3+ Tregs observed in this study are mainly Tregs from untreated group, indicating that the in vivo–expanded due to substantial expansion of CD4+Foxp3+ Tregs, whereas by guest on September 29, 2021 Tregs are relatively stable phenotypically. In agreement with this, conversion of CD4+CD252 Tconvs into Tregs may only marginally following the combined treatment, on days 21 and 69 the ex- contribute to this increase (45). panded Tregs could still exert their function and protect the treated Some previous reports raised concerns about the stability of NOD mice from induced T1D (Supplemental Fig. 2). expanded Tregs upon their transfer in vivo, due to a loss of Foxp3 expression and suppressive function, especially in the context of Discussion autoimmune settings (46, 47). Therefore, we evaluated the in vivo In this study, we show that Tregs expanded in vivo using a com- stability of the expanded Tregs in this study and found that the bined treatment of islet Ag (BDC2.5 mimetope peptide), IL-2/IL-2 treated Tregs did not lose their expression of Foxp3 (Fig. 7D). mAb complexes, and mammalian target of rapamycin inhibitor Moreover, the short-term combined treatment resulted in protec- (rapamycin) displayed an enhanced suppressive function and pro- tion against both spontaneous and induced T1D development in tected NOD mice from diabetes. NOD mice, suggesting that expanded Foxp3 expressing Tregs The expanded CD4+Foxp3+ T cells expressed normal levels of were stable (for .30 wk in the spontaneous T1D model and for at molecules associated with Treg development and regulation, in- least 7 wk in the induced T1D model). Previous studies reported cluding CTLA-4 and GITR. The expanded Tregs suppressed the that IL-2/anti–IL-2 mAb complexes provided some protection in vivo proliferation of responder CD4+CD252 T cells and the against autoimmune diseases (experimental autoimmune enceph- in vitro proliferation of responder CD8+ 8.3 T cells more effi- alomyelitis) in mice. However, such protection required either ciently than did control Tregs. These data suggest that our com- continuous treatment for several days (48), weeks (43), or bined treatments allow in vivo expansion of regulatory cells whose a combined treatment of rapamycin and IL-2/anti–IL-2 mAb phenotype and suppressive function resemble those of naturally complexes (29). The dose of rapamycin and the length of treat- occurring CD4+CD25+ Tregs. ment should be considered carefully because higher doses and/or Rapamycin has been shown to enhance CD4+CD25+ Treg repeated treatments of rapamycin have been shown to have det- proliferation and function in vitro (30) and to promote conversion rimental effects in vivo (49–52). of induced Tregs from CD4+CD252 Tconvs (44, 45). In this study, Because administration of peptide Ag together with immuno- we have confirmed that the combination of IL-2/IL-2 mAb com- modulatory reagents favors the induction of Ag-specific Tregs in vivo plexes, rapamycin, and islet Ag provides the most optimal ex- (32, 53, 54), we assume it should be desirable to include islet Ag pansion of Tregs in the draining lymph nodes when compared peptide in the combination, and indeed such a combination pro- with treatment with each ingredient individually (Fig. 1B, 1D). moted significant Ag-specific Treg expansion in the spleens and Although we cannot completely exclude the possibility that pancreatic lymph nodes (Fig. 2B, 2C). Among the expanded Tregs, some CD4+Foxp3+ Tregs were induced from CD4+Foxp32 ef- a small proportion are Ag specific (∼0.4%), mostly due to the low fector cells in the spleen, the increase in frequency and total frequency of BDC2.5 peptide–specific Tregs in the endogenous 10 IN VIVO EXPANSION OF Tregs PREVENTS DIABETES IN NOD MICE Downloaded from http://www.jimmunol.org/

FIGURE 7. Treatment leads to less islet infiltrates and tilts the cytokine production to Th2 profile; treated Tregs maintain their Foxp3 expression in vivo. A B n ( and ) Spleen cells were isolated from treated or untreated 6- to 8-wk-old NOD mice ( = 3/group). Cells were stimulated with PMA/ionomycin with by guest on September 29, 2021 GolgiPlug for 6 h. Cells were labeled with anti-CD4 and anti-CD8 Abs and then with different anti-cytokine Abs and analyzed by FACS. The results are the representative of two independent experiments with similar findings. (C) Pancreas from untreated control group and group transferred with treated Tregs were stained with H&E. Islets were scored for severity of infiltrates according to the previously reported method (55). (D) Thy1.1+/2 NOD mice were treated with the combined regimen or left untreated. On day 5, Tregs were isolated from untreated and treated group and then adoptively transferred to two groups of NOD mice (3 3 106 Tregs/mouse, i.v.; n = 5/group). Three weeks later, both groups were sacrificed, and Foxp3 expression within the transferred Thy1.1+CD4+CD25+ Treg population in the spleens was analyzed and plotted; each dot represents data from one individual mouse. The expression of Foxp3 on freshly isolated Tregs before injection was also analyzed. *p , 0.05, **p , 0.01.

Treg pool before treatment. This also indicates that some of the Notably, our short-term combined treatment provided long- cells expanded with the combined treatment are not Ag specific, lasting protection against diabetes; mice can be protected for which explains why even the 3-d treatment with IL-2/IL-2 mAb .1 y without developing T1D (data not shown). Repetitive, alone provided some protection (Fig. 6B), and the 10-wk-long IL-2 multiple treatments were not needed for the protection. complexes treatment achieved similar protection (Fig. 6D). The IL- In conclusion, the combination of IL-2/IL-2 mAb complexes, 2 complexes with rapamycin (but without islet Ag peptide) achieved rapamycin, and islet Ag peptide can efficiently expand islet Ag- significant protection from T1D as well (Fig. 6C). It seems that the specific Tregs in vivo. The expanded Tregs exhibit a classical Ag nonspecific Tregs expanded can provide a good level of pro- Treg phenotype and are suppressive both in vitro and in vivo. tection, and for expansion of protective Tregs in vivo, the minimal Importantly, this combination therapy prevented both spontaneous requirement is the combination of IL-2 complexes plus rapamycin, and induced diabetes development in NOD mice. This protection because IL-2 complexes alone or rapamycin alone did not provide was transferrable by adoptive transfer of the expanded Tregs, and it optimal protection against T1D (Fig. 6A, 6B). could be attributed to reduced infiltration of immune cells into

Table III. Combined treatment leads to a decreased Th1/Th2 cytokine ratio

Cytokine Ratios for CD4+ Cells Cytokine Ratios for CD8+ Cells

Untreated Treated p Value Untreated Treated p Value IFN-g/IL-10 22.34 6 0.59 1.58 6 0.07 ,0.05 14.64 6 0.84 11.25 6 0.59 ,0.05 IFN-g/IL-4 85.64 6 16.29 7.5 6 0.67 ,0.05 24.67 6 1.76 15.67 6 1.86 ,0.05 Spleen cells were isolated from treated or untreated 6- to 8-wk-old NOD mice (n = 3/group) and stimulated with PMA/ionomycin with GolgiPlug for 6 h. Cells were labeled with anti-CD4 and anti-CD8 Abs first and then permeabilized and stained with different anti-cytokine Abs and analyzed by FACS. The ratios of IFN-g/IL-10 and IFN-g/IL-4 within the CD4 and CD8 T cell subsets were calculated. The results are the representative of two independent experiments with similar findings. The Journal of Immunology 11 pancreatic islets as well as polarization of the immune response 21. Stephens, L. A., K. H. Malpass, and S. M. Anderton. 2009. Curing CNS auto- + toward a Th2 cytokine profile. immune disease with myelin-reactive Foxp3 Treg. Eur. J. Immunol. 39: 1108– 1117. 22. Wright, G. P., C. A. Notley, S. A. Xue, G. M. Bendle, A. Holler, T. N. Schumacher, M. R. Ehrenstein, and H. J. Stauss. 2009. Adoptive therapy Acknowledgments with redirected primary regulatory T cells results in antigen-specific suppression We thank Drs. Graeme Price and Bharat Joshi of Center for Biologics of arthritis. Proc. Natl. Acad. Sci. USA 106: 19078–19083. Evaluation and Research, Food and Drug Administration for critically re- 23. Elinav, E., N. Adam, T. Waks, and Z. Eshhar. 2009. Amelioration of colitis by viewing this manuscript. We acknowledge the National Institutes of Health genetically engineered murine regulatory T cells redirected by antigen-specific g7 chimeric receptor. Gastroenterology 136: 1721–1731. Tetramer Core Facility for provision of BDC/I-A tetramers. 24. Blat, D., E. Zigmond, Z. Alteber, T. Waks, and Z. Eshhar. 2014. Suppression of murine colitis and its associated cancer by carcinoembryonic antigen-specific regulatory T cells. Mol. Ther. 22: 1018–1028. Disclosures 25. Malek, T. R., A. Yu, V. Vincek, P. Scibelli, and L. Kong. 2002. CD4 regulatory The authors have no financial conflicts of interest. T cells prevent lethal autoimmunity in IL-2Rb-deficient mice. Implications for the nonredundant function of IL-2. Immunity 17: 167–178. 26. Wolf, M., A. Schimpl, and T. Hunig.€ 2001. Control of T cell hyperactivation in IL-2-deficient mice by CD4+CD252 and CD4+CD25+ T cells: evidence for two References distinct regulatory mechanisms. Eur. J. Immunol. 31: 1637–1645. 27. Bayer, A. L., A. Yu, D. Adeegbe, and T. R. Malek. 2005. Essential role for 1. Rother, K. I. 2007. Diabetes treatment—bridging the divide. N. Engl. J. Med. + + 356: 1499–1501. interleukin-2 for CD4 CD25 T regulatory cell development during the neonatal 2. Nakayama, M., N. Abiru, H. Moriyama, N. Babaya, E. Liu, D. Miao, L. Yu, period. J. Exp. Med. 201: 769–777. 28. Setoguchi, R., S. Hori, T. Takahashi, and S. Sakaguchi. 2005. Homeostatic D. R. Wegmann, J. C. Hutton, J. F. Elliott, and G. S. Eisenbarth. 2005. Prime role + + + for an insulin epitope in the development of type 1 diabetes in NOD mice. maintenance of natural Foxp3 CD25 CD4 regulatory T cells by interleukin

Nature 435: 220–223. (IL)-2 and induction of autoimmune disease by IL-2 neutralization. J. Exp. Med. Downloaded from 3. DiLorenzo, T. P., and D. V. Serreze. 2005. The good turned ugly: immunopa- 201: 723–735. thogenic basis for diabetogenic CD8+ T cells in NOD mice. Immunol. Rev. 204: 29. Webster, K. E., S. Walters, R. E. Kohler, T. Mrkvan, O. Boyman, C. D. Surh, 250–263. S. T. Grey, and J. Sprent. 2009. In vivo expansion of T reg cells with IL-2-mAb 4. Roep, B. O. 2000. T-cell reactivity to b-cell antigens in human insulin-dependent complexes: induction of resistance to EAE and long-term acceptance of (type 1) diabetes mellitus. Implications for diagnosis and therapy. Clin. Rev. islet allografts without immunosuppression. J. Exp. Med. 206: 751–760. Allergy Immunol. 19: 265–275. 30. Battaglia, M., A. Stabilini, B. Migliavacca, J. Horejs-Hoeck, T. Kaupper, and + 5. Pinkse, G. G., C. Boitard, T. I. Tree, M. Peakman, and B. O. Roep. 2006. HLA M. G. Roncarolo. 2006. Rapamycin promotes expansion of functional CD4 class I epitope discovery in type 1 diabetes: independent and reproducible CD25+FOXP3+ regulatory T cells of both healthy subjects and type 1 diabetic http://www.jimmunol.org/ identification of proinsulin epitopes of CD8 T cells—report of the IDS T Cell patients. J. Immunol. 177: 8338–8347. Workshop Committee. Ann. N. Y. Acad. Sci. 1079: 19–23. 31. Battaglia, M., A. Stabilini, and M. G. Roncarolo. 2005. Rapamycin selectively 6. Pinkse, G. G., O. H. Tysma, C. A. Bergen, M. G. Kester, F. Ossendorp, P. A. van expands CD4+CD25+FoxP3+ regulatory T cells. Blood 105: 4743–4748. Veelen, B. Keymeulen, D. Pipeleers, J. W. Drijfhout, and B. O. Roep. 2005. 32. Daniel, C., K. Wennhold, H. J. Kim, and H. von Boehmer. 2010. Enhancement of Autoreactive CD8 T cells associated with b cell destruction in type 1 diabetes. antigen-specific Treg vaccination in vivo. Proc. Natl. Acad. Sci. USA 107: Proc. Natl. Acad. Sci. USA 102: 18425–18430. 16246–16251. 7. Mallone, R., V. Brezar, and C. Boitard. 2011. T cell recognition of autoantigens 33. Zhang, P., S. K. Tey, M. Koyama, R. D. Kuns, S. D. Olver, K. E. Lineburg, in human type 1 diabetes: clinical perspectives. Clin. Dev. Immunol. 2011: M. Lor, B. E. Teal, N. C. Raffelt, J. Raju, et al. 2013. Induced regulatory T cells 513210. promote tolerance when stabilized by rapamycin and IL-2 in vivo. J. Immunol. 8. Nagata, M., R. Kotani, H. Moriyama, K. Yokono, B. O. Roep, and M. Peakman. 191: 5291–5303. 2004. Detection of autoreactive T cells in type 1 diabetes using coded auto- 34. Haskins, K., M. Portas, B. Bradley, D. Wegmann, and K. Lafferty. 1988. antigens and an immunoglobulin-free cytokine ELISPOT assay: report from the T-lymphocyte clone specific for pancreatic islet antigen. Diabetes 37: 1444– by guest on September 29, 2021 fourth immunology of diabetes society T cell workshop. Ann. N. Y. Acad. Sci. 1448. 1037: 10–15. 35. Lieberman, S. M., A. M. Evans, B. Han, T. Takaki, Y. Vinnitskaya, 9. Ricordi, C., and T. B. Strom. 2004. Clinical islet transplantation: advances and J. A. Caldwell, D. V. Serreze, J. Shabanowitz, D. F. Hunt, S. G. Nathenson, et al. immunological challenges. Nat. Rev. Immunol. 4: 259–268. 2003. Identification of the b cell antigen targeted by a prevalent population of 10. Shapiro, A. M. J., C. Ricordi, B. J. Hering, H. Auchincloss, R. Lindblad, pathogenic CD8+ T cells in autoimmune diabetes. Proc. Natl. Acad. Sci. USA R. P. Robertson, A. Secchi, M. D. Brendel, T. Berney, D. C. Brennan, et al. 2006. 100: 8384–8388. International trial of the Edmonton protocol for islet transplantation. N. Engl. J. 36. Boyman, O., M. Kovar, M. P. Rubinstein, C. D. Surh, and J. Sprent. 2006. Se- Med. 355: 1318–1330. lective stimulation of T cell subsets with antibody-cytokine immune complexes. 11. Sibley, R. K., D. E. Sutherland, F. Goetz, and A. F. Michael. 1985. Recurrent Science 311: 1924–1927. diabetes mellitus in the pancreas iso- and allograft. A light and electron mi- 37. Judkowski, V., C. Pinilla, K. Schroder, L. Tucker, N. Sarvetnick, and croscopic and immunohistochemical analysis of four cases. Lab. Invest. 53: 132– D. B. Wilson. 2001. Identification of MHC class II-restricted peptide ligands, 144. including a glutamic acid decarboxylase 65 sequence, that stimulate diabeto- 12. Tang, D. Q., L. Shun, V. Koya, Y. Sun, Q. Wang, H. Wang, S. W. Li, Y. Sun, genic T cells from transgenic BDC2.5 nonobese diabetic mice. J. Immunol. 166: D. L. Purich, C. Zhang, et al. 2013. Genetically reprogrammed, liver-derived 908–917. insulin-producing cells are glucose-responsive, but susceptible to autoimmune 38. Kaufman, D. L., M. Clare-Salzler, J. Tian, T. Forsthuber, G. S. Ting, P. Robinson, destruction in settings of murine model of type 1 diabetes. Am. J. Transl. Res. 5: M. A. Atkinson, E. E. Sercarz, A. J. Tobin, and P. V. Lehmann. 1993. Sponta- 184–199. neous loss of T-cell tolerance to glutamic acid decarboxylase in murine insulin- 13. Shevach, E. M. 2009. Mechanisms of Foxp3+ T regulatory cell-mediated sup- dependent diabetes. Nature 366: 69–72. pression. Immunity 30: 636–645. 39. Daniel, D., R. G. Gill, N. Schloot, and D. Wegmann. 1995. Epitope specificity, 14. Sakaguchi, S., R. Setoguchi, H. Yagi, and T. Nomura. 2006. Naturally arising + + cytokine production profile and diabetogenic activity of insulin-specific T cell Foxp3-expressing CD25 CD4 regulatory T cells in self-tolerance and autoim- mune disease. Curr. Top. Microbiol. Immunol. 305: 51–66. clones isolated from NOD mice. Eur. J. Immunol. 25: 1056–1062. 15. Salomon, B., D. J. Lenschow, L. Rhee, N. Ashourian, B. Singh, A. Sharpe, and 40. Yadav, M., C. Louvet, D. Davini, J. M. Gardner, M. Martinez-Llordella, J. A. Bluestone. 2000. B7/CD28 costimulation is essential for the homeostasis of S. Bailey-Bucktrout, B. A. Anthony, F. M. Sverdrup, R. Head, D. J. Kuster, et al. the CD4+CD25+ immunoregulatory T cells that control autoimmune diabetes. 2012. Neuropilin-1 distinguishes natural and inducible regulatory T cells among regulatory T cell subsets in vivo. J. Exp. Med. 209: 1713–1722. Immunity 12: 431–440. + + 16. National Institutes of Health. 2014. ClinicalTrials.com database. Available at: 41. Shevach, E. M. 2002. CD4 CD25 suppressor T cells: more questions than http://www.clinicaltrials.gov/. Accessed October 3, 2014. answers. Nat. Rev. Immunol. 2: 389–400. 42. Sgouroudis, E., and C. A. Piccirillo. 2009. Control of type 1 diabetes by CD4+ 17. Tarbell, K. V., S. Yamazaki, K. Olson, P. Toy, and R. M. Steinman. 2004. + CD25+ CD4+ T cells, expanded with dendritic cells presenting a single Foxp3 regulatory T cells: lessons from mouse models and implications for autoantigenic peptide, suppress autoimmune diabetes. J. Exp. Med. 199: human disease. Diabetes Metab. Res. Rev. 25: 208–218. 1467–1477. 43. Tang, Q., J. Y. Adams, C. Penaranda, K. Melli, E. Piaggio, E. Sgouroudis, 18. Tang, Q., K. J. Henriksen, M. Bi, E. B. Finger, G. Szot, J. Ye, E. L. Masteller, C. A. Piccirillo, B. L. Salomon, and J. A. Bluestone. 2008. Central role of de- H. McDevitt, M. Bonyhadi, and J. A. Bluestone. 2004. In vitro-expanded fective interleukin-2 production in the triggering of islet autoimmune destruc- antigen-specific regulatory T cells suppress autoimmune diabetes. J. Exp. Med. tion. Immunity 28: 687–697. 199: 1455–1465. 44. Gao, W., Y. Lu, B. El Essawy, M. Oukka, V. K. Kuchroo, and T. B. Strom. 19. Jaeckel, E., H. von Boehmer, and M. P. Manns. 2005. Antigen-specific FoxP3- 2007. Contrasting effects of cyclosporine and rapamycin in de novo gen- transduced T-cells can control established type 1 diabetes. Diabetes 54: 306–310. eration of alloantigen-specific regulatory T cells. Am.J.Transplant.7: 1722–1732. 20. Sagoo, P., N. Ali, G. Garg, F. O. Nestle, R. I. Lechler, and G. Lombardi. 2011. 45. Shin, H. J., J. Baker, D. B. Leveson-Gower, A. T. Smith, E. I. Sega, and Human regulatory T cells with alloantigen specificity are more potent inhibitors R. S. Negrin. 2011. Rapamycin and IL-2 reduce lethal acute graft-versus-host of alloimmune skin graft damage than polyclonal regulatory T cells. Sci. Transl. disease associated with increased expansion of donor type CD4+CD25+Foxp3+ Med. 3: 83ra42. regulatory T cells. Blood 118: 2342–2350. 12 IN VIVO EXPANSION OF Tregs PREVENTS DIABETES IN NOD MICE

46. Zhou, X., S. L. Bailey-Bucktrout, L. T. Jeker, C. Penaranda, M. Martı´nez- in vivo: implications for everolimus and zotarolimus eluting stents. Eur. Heart J. Llordella, M. Ashby, M. Nakayama, W. Rosenthal, and J. A. Bluestone. 2009. 31: 236–242. Instability of the transcription factor Foxp3 leads to the generation of pathogenic 52. Tanemura, M., Y. Ohmura, T. Deguchi, T. Machida, R. Tsukamoto, H. Wada, memory T cells in vivo. Nat. Immunol. 10: 1000–1007. S. Kobayashi, S. Marubashi, H. Eguchi, T. Ito, et al. 2012. Rapamycin causes 47. Josefowicz, S. Z., and A. Rudensky. 2009. Control of regulatory T cell lineage upregulation of autophagy and impairs islets function both in vitro and in vivo. commitment and maintenance. Immunity 30: 616–625. Am. J. Transplant. 12: 102–114. 48. Grinberg-Bleyer, Y., A. Baeyens, S. You, R. Elhage, G. Fourcade, S. Gregoire, 53. Daniel, C., H. Ploegh, and H. von Boehmer. 2011. Antigen-specific induc- N. Cagnard, W. Carpentier, Q. Tang, J. Bluestone, et al. 2010. IL-2 reverses tion of regulatory T cells in vivo and in vitro. Methods Mol. Biol. 707: 173– established type 1 diabetes in NOD mice by a local effect on pancreatic regu- 185. latory T cells. J. Exp. Med. 207: 1871–1878. 54. Zhang, J., W. Gao, X. Yang, J. Kang, Y. Zhang, Q. Guo, Y. Hu, G. Xia, and 49. Valle, A., T. Jofra, A. Stabilini, M. Atkinson, M. G. Roncarolo, and M. Battaglia. Y. Kang. 2013. Tolerogenic vaccination reduced effector memory CD4 T cells 2009. Rapamycin prevents and breaks the anti-CD3-induced tolerance in NOD and induced effector memory Treg cells for type I diabetes treatment. PLoS One mice. Diabetes 58: 875–881. 8: e70056. 50. Kwon, Y. S., H. S. Hong, J. C. Kim, J. S. Shin, and Y. Son. 2005. Inhibitory effect 55. Krishnamurthy, B., N. L. Dudek, M. D. McKenzie, A. W. Purcell, of rapamycin on corneal neovascularization in vitro and in vivo. Invest. Oph- A.G.Brooks,S.Gellert,P.G.Colman,L.C.Harrison,A.M.Lew, thalmol. Vis. Sci. 46: 454–460. H. E. Thomas, and T. W. Kay. 2006. Responses against islet antigens in 51. Camici, G. G., J. Steffel, I. Amanovic, A. Breitenstein, J. Baldinger, S. Keller, NOD mice are prevented by tolerance to proinsulin but not IGRP. J. Clin. T. F. Luscher,€ and F. C. Tanner. 2010. Rapamycin promotes arterial thrombosis Invest. 116: 3258–3265. Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021