Humanized Mice as a Model for Aberrant Responses in Human Immunotherapy Nalini K. Vudattu, Frank Waldron-Lynch, Lucy A. Truman, Songyan Deng, Paula Preston-Hurlburt, Richard Torres, This information is current as Maurice T. Raycroft, Mark J. Mamula and Kevan C. Herold of September 23, 2021. J Immunol 2014; 193:587-596; Prepublished online 18 June 2014; doi: 10.4049/jimmunol.1302455

http://www.jimmunol.org/content/193/2/587 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2014/06/18/jimmunol.130245 Material 5.DCSupplemental http://www.jimmunol.org/ References This article cites 53 articles, 25 of which you can access for free at: http://www.jimmunol.org/content/193/2/587.full#ref-list-1

Why The JI? Submit online.

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

by guest on September 23, 2021 • No Triage! Every submission reviewed by practicing scientists

• 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 Copyright © 2014 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Humanized Mice as a Model for Aberrant Responses in Human T Cell Immunotherapy

Nalini K. Vudattu,*,1 Frank Waldron-Lynch,*,1 Lucy A. Truman,* Songyan Deng,* Paula Preston-Hurlburt,* Richard Torres,† Maurice T. Raycroft,‡ Mark J. Mamula,‡ and Kevan C. Herold*,‡

Immune-deficient mice, reconstituted with human stem cells, have been used to analyze human immune responses in vivo. Although they have been used to study immune responses to xenografts, allografts, and pathogens, there have not been models of in which the mechanisms of the pathologic process can be analyzed. We have found that reconstituted “humanized” mice treated with anti–CTLA-4 Ab (ipilimumab) develop autoimmune disease characterized by hepatitis, adrenal- itis, sialitis, anti-nuclear Abs, and weight loss. Induction of autoimmunity involved activation of T cells and cytokine production, and increased infiltration of APCs. When anti–CTLA-4 mAb–treated mice were cotreated with anti-CD3 mAb (teplizumab), Downloaded from hepatitis and anti-nuclear Abs were no longer seen and weight loss did not occur. The anti-CD3 blocked proliferation and activation of T cells, release of IFN-g and TNF, infiltration, and release of IP-10 that was induced with anti– CTLA-4 mAb. We also found increased levels of T regulatory cells (CD25+CD1272) in the spleen and mesenteric lymph nodes in the mice treated with both Abs and greater constitutive phosphorylation of STAT5 in T regulatory cells in spleen cells compared with mice treated with anti–CTLA-4 mAb alone. We describe a model of human autoimmune disease in vivo. Humanized mice may be useful for understanding the mechanisms of biologics that are used in patients. Hepatitis, lymphadenopathy, and other http://www.jimmunol.org/ inflammatory sequelae are adverse effects of ipilimumab treatment in humans, and this study may provide insights into this pathogenesis and the effects of immunologics on autoimmunity. The Journal of Immunology, 2014, 193: 587–596.

fforts to develop new treatments for human autoimmune participants (1, 2). Thus, an animal model system of autoimmunity diseases have been hampered by the lack of small-animal with human immune cells would be optimal. E models of the disease states and the inherent differences Many investigators have used “humanized” mice to study hu- between immune responses of human and murine cells. Studies man immune responses in vivo. These model systems have gen-

of human autoimmune diseases have relied on the analysis of ro- erally involved transfer of mature human T cells or alternatively by guest on September 23, 2021 dent models or characterization of human cells ex vivo. However, murine cells that express human genes, but these systems do not studies of human cells in vitro cannot re-create the cellular dy- permit the study of human immune cells and responses after namics that may be important in the initiation, progression, and perturbations of cells that develop in and are tolerant to the host regulation of disease. Murine models have been relied on to pro- (3–9). A model system that can be used to study human autoim- vide insights into the pathologic mechanisms, but there are many mune disease in vivo has not been described. shortcomings of these models because of intrinsic differences be- After activation, T cells express CTLA-4, a transmembrane tween and human immune systems (1). In some cases, the glycoprotein of the Ig superfamily. Its engagement delivers neg- failure to recognize the differences between rodent and human ative signals that circumvent the actions of the CD28-mediated immune responses has resulted in significant harm to clinical trial costimulatory pathway. CTLA-4 was first identified during a search for cytotoxic genes using subtractive hybridization, and subsequent studies in knockout mice showed its pivotal role in maintaining 2/2 *Department of Immunobiology, Yale University School of Medicine, New Haven, peripheral tolerance (10–13). CTLA-4 mice develop an auto- CT 06520; †Department of Laboratory Medicine, Yale University School of Medicine, ‡ immune lymphoproliferative disorder characterized by infiltration New Haven, CT 06520; and Department of Internal Medicine, Yale University + School of Medicine, New Haven, CT 06520 of CD4 T cells leading to multiorgan tissue destruction within 3– 1N.K.V. and F.W.-L. contributed equally to this work. 4 wk of age. A number of molecular mechanisms have been proposed for the functions of CTLA-4–mediated negative regu- Received for publication September 16, 2013. Accepted for publication May 15, 2014. lation in T cells, including competition for CD28 ligands, This work was supported by the Juvenile Diabetes Research Foundation (Grant 2006- disruption of CD28 localization in the immunological synapse, 1059 to K.C.H.) and the National Institutes of Health (Grants U19 AI082713, U01 decreased TCR signaling by PP2A and SHP-2, and limited en- AI102011, and R01 DK057846 to K.C.H. and Grant AR41032 to M.J.M.). gagement of T cells with APCs (14–18). These cell intrinsic Address correspondence and reprint requests to Dr. Kevan C. Herold, Yale Uni- mechanisms have been thought to account for the widespread versity, 300 George Street, #353E, New Haven, CT 06520. E-mail address: kevan. [email protected] activation and proliferation of T cells that is seen in CTLA-4– The online version of this article contains supplemental material. deficient mice, which may include cells with self-reactive TCRs Abbreviations used in this article: ALT, alanine aminotransferase; ANA, anti-nuclear leading to autoimmune disease (19–21). Ab; GVHD, graft-versus-host disease; hCD45+, human CD45+; hIg, human Ig; NSG, In addition, cell extrinsic factors may also account for autoim- 2/2 NOD/SCID IL-2gc ; SLE, systemic ; Tconv, T conventional munity in the setting of CTLA-4 blockade. CTLA-4 is constitutively cell; Teff, T effector cell; Treg, T regulatory cell. expressed on CD4+ T regulatory cells (Tregs) (22), whereas it is Copyright Ó 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00 rapidly induced on T conventional cells (Tconvs) after TCR en- www.jimmunol.org/cgi/doi/10.4049/jimmunol.1302455 588 A HUMANIZED MOUSE MODEL OF AUTOIMMUNE DISEASE gagement (23), suggesting that CTLA-4 has an important role in Flowery Branch, GA) and 0.01% sodium azide (Sigma-Aldrich, St. Louis, both Treg and Tconv function. Recently, dual function of CTLA-4 MO). The FACS data were analyzed with FlowJo software (version 7.6.3). in Tregs and T effector cells (Teffs) to prevent autoimmunity has Treatment of mice been proposed (24). Yamaguchi et al. (25) showed that a triad of IL-2 repression, CTLA-4 expression, and antigenic stimulation is Each reconstituted mouse was given 500 mg ipilimumab (Bristol-Myers Squibb, New York, NY) or human Ig (hIg; Sigma, St. Louis, MO) with or a minimalistic requirement for conferring Treg-like suppressive without coadministration of teplizumab (MacroGenics; 5 mg) beginning the activity on Tconvs. day before ipilimumab i.p. every 4 d for a period of 40 d (35, 36). The mice Because of its role in maintaining tolerance and the enhanced were weighed regularly, and those that lost 15% of their baseline body weight cellular immune responses in its absence, blockade of CTLA-4 were sacrificed. with an mAb has been developed for the treatment of malignan- Histology and immunohistochemistry cies to activate tumor-specific effector cells (26). The anti-human The experimental mice were sacrificed after they had lost 15% of their body CTLA-4 mAb, ipilimumab, has been approved for the treatment weight or after 4 wk. Paraffin-embedded sections were prepared as pre- of metastatic melanoma and is also being tested in combination viously described (36). For immunohistochemistry, frozen sections were with other agents for treatment of refractory tumors. Blockade incubated for 15 min with 1% goat serum in PBS followed by avidin/biotin of CTLA-4 pathway resulted in an increased ratio of effector to blocking as per the manufacture’s protocol (Vector Laboratories, Burlin- game, CA). Anti-human CD4 (4B12), CD8 (CD/144B), CD19 (LE-CD19), Tregs (27), which has been proposed to account for the enhanced and CD68 (LP1) from Dako (Glostrup, Denmark) were applied to con- killing of tumors in anti–CTLA-4 mAb–treated patients. However, tiguous slides overnight followed by staining with biotinylated goat anti- in addition to these antitumor effects, ipilimumab-treated patients mouse Ab (Dako) and staining with peroxidase substrate (Vector labora- also develop immune-mediated adverse events including derma- tories). Murine were stained with pan macrophage marker Downloaded from titis, colitis, hepatitis, pancreatitis, hypophysitis, and uveitis (28). F4/80 (BM8). Hematoxylin counterstain was applied. Immunofluorescent staining for human macrophages was performed on frozen sections. These immune-mediated adverse effects are also thought to be due In brief, tissue sections were treated with citrate buffer and then heated by to the release of T cell inhibition, but there have not been studies microwave. The slides were then blocked with 4% NGS and 5% BSA in that have analyzed the immune mechanisms at the sites of pa- PBS for 1 h. Anti-CD68 and F4/80 Abs were diluted in blocking buffer and thology. Moreover, hyperproliferation of CD4+ T cells is thought added to the slides for 2 h followed by washing and staining with goat anti- mouse Alexa flour 546 red (1:300) and DAPI. to drive autoimmune tissue pathology in other settings such as in http://www.jimmunol.org/ human systemic lupus erythematosus (SLE) (29). SLE is charac- Extraction, purification, and FACS analysis of terized by the presence of anti-nuclear Abs (ANAs). human lymphocytes We therefore have studied the effects of ipilimumab in hu- Cells were isolated from the spleen, peripheral and mesenteric lymph nodes, manized mice that were created by reconstitution of immune-defi- liver, and adrenal glands by tissue homogenization. The liver-infiltrating cells cient recipients with human CD34+ stem cells. We found that were purified by Ficoll density gradient centrifugation. The cells from spleen the drug induced autoimmune disease characterized by weight and blood were treated with ACK lysing buffer (Lonza, Walkersville, MD) to remove RBCs. Some splenocytes were studied immediately, and others loss, hepatitis, and detectable ANAs. To define the mechanisms of were frozen and analyzed later. The following conjugated mAbs were pur- the disease and tissue pathology, we treated mice with another chased from Biolegend (San Diego, CA) and BD (San Jose, CA) and used for human biologic, teplizumab, which is a non-FcR binding anti- flow cytometry analysis to phenotype lymphocytes: CD45 (HI30) allophy- by guest on September 23, 2021 CD3 mAb that has been used to treat patients with new-onset cocyanin, anti-mouse CD45 (30-F11) Pacific Blue, CD19 (SJ25C1) PE, CD3 type 1 diabetes (30–34). Coadministration of anti-CD3 with anti– (OKT-3) FITC, CD4 (RPA-T4) allophycocyanin-H7, CD8 (RPA-T8) PerCP- Cy5.5, CD45RO (UCHL-1), CD62L (DREG-56) PE-Cy7, CCR7 (GO43H7) CTLA-4 reversed all autoimmune manifestations induced by anti– PE, CD45RA (HI100) allophycocyanin-H7, CXCR3(GO25H7) Alexa 647, CTLA-4 Ab. Our studies suggest three mechanisms that account CCR6 (11A9), CD127(A019D5) Brilliant Violet 421, ICOS (C398.4A) PerCP- for the development of autoimmunity: T cell activation and pro- Cy5.5, CD25 (2A3) PE-Cy7, PD-1 (EH12.2H7) Alexa Fluor 647. CD4+ T cell liferation, impaired localization of Tregs to pathologic sites, and phenotype was determined using CD45RA, CD45RO, and CCR7 markers: naive: CD45RA+CCR7+; effector: CD45RA2CCR72;effectormemoryRA: recruitment and interaction with APCs. These findings were re- CD45RA+CCR72; central memory: CD45RA2CCR7+. To identify Tregs, we versed by anti-CD3 mAb treatment. fixed the cells with intracellular fix/perm buffer for 20 min at room tempera- ture, followed by permeabilization buffer and staining with anti-FOXP3 (259D) Materials and Methods Alexa Fluor 647 (Biolegend) and anti-CD3 (OKT-3) FITC, CD4 (RPA-T4) Mice allophycocyanin-H7, CD25(2A3)-PE-Cy7, CD127 (A019D5). For FACS anal- ysis, an electronic gate was placed on live hCD45 cells. Human T and NOD/SCID IL-2gc2/2 (NSG) mice were purchased from The Jackson lymphocyte numbers were enumerated and expressed as the total cell Laboratory, bred at our animal facility, and maintained under specific counts or percent of hCD45+ cells. Viability of cells was confirmed by use of pathogen-free conditions. Experiments were approved by the Institutional a Live/Dead fixable dead stain kit (Invitrogen, Grand Island, NY). FACS analysis Animal Care and Use Committee of Yale University. Human fetal liver was was performed on an LSRII (BD Biosciences). obtained from Advanced Bioscience Resources (Alameda, CA). Approval for the use of human cells was obtained from the Yale University Insti- Intracellular cytokine staining tutional Review Board. Frozen spleen cells were thawed and the cells were counted. The cells were then Isolation of human CD34+ stem cells and reconstitution of stimulated with PMA (10 ng/ml) and Ionomycin (500 ng/ml) in the presence of GolgiStop (BD Biosciences) for 5 h. The cells were washed with FACS stain- mice ing buffer and stained first with the live/dead marker followed by staining with CD34+ cells were isolated from human fetal liver from six donors on surface markers CD3 (OKT-3) Brilliant Violet 650, CD4 (RPA-T4), CD8 a discontinuous density gradient of Ficoll-Paque Plus (GE Healthcare, (HIT8a) for 30 min at 4˚C. Cells were fixed and permeabilized and stained with Pittsburgh, PA) and purified with magnetically labeled microbeads con- anti–IFN-g (4S.B3) and IL-5 (JES1-39D10) PE, TNF-a (MAb11) PerCP- jugated to anti-human CD34+ (Miltenyi Biotec, San Diego, CA). Prepa- Cy5.5, IL-10 (JES3-9D7) PE-Cy7, and IL-17A (SCPL-1362) Alexa Fluor rations of the CD34+ cells that showed at least 80% purity by FACS were 647 Abs. The cells were incubated at 4˚C for 30 min, washed, and resuspended used for reconstitution of mice. The CD34+ cells (2 3 105) were injected in the staining buffer. intrahepatically into 1- to 2-d-old NSG pups that had been irradiated Stat5 phosphorylation in T cells (1 Gy). The reconstitution of these mice with human CD45+ (hCD45+) cells was determined in peripheral blood at 8 wk by flow cytometry STAT5 phosphorylation was also measured by flow cytometry in CD3+, (FACSFortessa; Becton Dickinson, San Jose, CA). On average, 91% of the CD4+, CD8+, and Tregs from the spleen ex vivo and after stimulation with cells within the scatter gates were hCD45+. All FACS staining was performed IL-2 in vitro (100 U/well). Splenocytes were cultured overnight in serum- in PBS supplemented with 2% heat-inactivated FCS (Atlanta Biologicals, free medium (AIM-V from Invitrogen, Carlsbad, CA) followed by incu- The Journal of Immunology 589 bation with 100 U recombinant human IL-2 for 105 cells for 20 min at Expansion and activation of human T cells with anti–CTLA-4 mAb 37˚C. The cells were stained with anti-CD3 PECF594 (clone UCHT1), treatment. To determine the effects of the anti–CTLA-4 mAb treat- anti-CD4 Alexa700 (RPA-T4), and anti-CD8a allophycocyanin-Cy7 ment on immune cells, we analyzed the cellular composition of the (clone HIT8a). Cells were incubated for 15 min at 4˚C and immediately fixed with 2% PFA at 37˚C for 10 min followed by permeabilization lymphoid organs. In general, NSG mice that are reconstituted with procedure as described previously (37). The anti-STAT5 Ab (y694)-Alexa fetal liver CD34+ cells show a reduced number of splenocytes (∼18 3 488 was added and incubated for 1 h at room temperature in the dark, 106) compared with the number of murine lymphocytes in the followed by washing with staining buffer. Cells were analyzed by flow spleen of normal mice. However, in mice reconstituted with fetal cytometry using a FACSAria (BD Biosciences). liver cells and treated with ipilimumab, the spleens were enlarged Serum cytokine analysis (Fig. 2A, 2B), with a 2-fold increase in the number of hCD45+ Human cytokines were measured in serum samples with an ultrasensitive splenocytes. Moreover, the mesenteric and other peripheral lymph + + Milliplex xMAP assay (Millipore, Billerica, MA). The lower limit of de- nodes were enlarged and showed infiltration with CD4 ,CD8 ,and tection for this assay was 1 pg/ml. CD19+ lymphocytes (Fig. 2C). However, the lymph nodes did not ANA staining show evidence of organization (Fig. 2Ca, b). We analyzed activation markers on the lymphocytes in the ANA testing was performed with HEp-2 slides that were incubated with spleen, as well as the levels of cytokines in the circulation. The a 1:40 dilution of serum samples collected from both control and experi- CD4+ splenocytes in the ipilimumab-treated mice expressed mental mice. The serum was incubated with the slides for 1 h at room temperature in a humidifying chamber. Excess unbound Abs were washed higher levels of CXCR3, PD-1, ICOS, and CD25, and the levels of for 5 min in PBS. The slides were incubated with secondary Ab at a 1:100 IFN-g and TNF were increased in the serum of the mice, com- dilution in PBS for 1 h. They were washed in PBS for 5 min and mounted. pared with the hIg-treated mice (Fig. 3A, 3B). Furthermore, .12% Downloaded from Images were rendered using an Olympus BX40 microscope. They were of the CD4+ cells expressed both IFN-g and TNF after PMA/ scored from +1 to +4 depending on the intensity of staining for ANA. Ionomycin stimulation (Fig. 3C, 3D). Statistical analyses To understand the effects of the mAb on T cell expansion and Unless indicated, the mean 6 SEM is reported. The analyses were per- activation, we measured STAT5 phosphorylation (pSTAT5) in formed with GraphPad software. T cells after stimulation with IL-2. pSTAT5 levels were increased + + after culture with IL-2 in CD4 and CD8 (Fig. 3E) Teffs from http://www.jimmunol.org/ Results ipilimumab-treated mice. Weight loss in reconstituted humanized mice treated with anti–CTLA-4 mAb Autoimmunity in humanized mice treated with anti–CTLA-4 mAb. These findings showed that treatment with anti–CTLA-4 mAb + We screened the reconstituted mice for hCD45 cells in peripheral induced activation of T cells resulting in cytokine production and blood by flow cytometry at 8 wk of age. On average, 91% of the cells heightened cytokine signaling. To determine whether the T cell + + within the scatter gates were hCD45 ; 25.5% were CD3 and 60.6% activation observed after treatment of humanized mice with anti– + were CD19 . The reconstituted mice were treated with ipilimumab CTLA-4mAb led to organ-specific damage, we examined the liver or hIg every 4 d at a dose that has been used to treat patients

for inflammatory lesions because damage of this organ might have by guest on September 23, 2021 with malignant melanoma (20 mg/kg). Beginning at day 7, the contributed to the observed weight loss. We found inflammatory ipilimumab-treated mice began to lose weight; by day 30, 80% of the infiltrates in in the liver, which were composed of human CD4+ and mice had lost at least 15% of their pretreatment body weight or an CD8+ T cells, B cells (CD19+), and both human (CD68+; Fig. 4A) , + average of 14 g (Fig. 1; p 0.05 versus baseline weight). The me- and mouse (F4/80 ) macrophages (Fig. 4B). The serum levels of dian time of survival was 26 d. In contrast, the body weight in the alanine aminotransferase (ALT) were elevated in the ipilimumab- hIg-treated mice did not change significantly (p = 0.0002 versus hIg). treated mice, and the serum albumin was decreased, indicating Weight loss is a common manifestation of graft-versus-host dis- hepatocellular damage with decreased synthetic function (Fig. 4C). ease (GVHD), which may have been precipitated by the treat- We also found cellular infiltrates in the adrenal cortex and sali- ment with the anti–CTLA-4 mAb. However, the mice did not show vary glands of the ipilimumab-treated mice (Supplemental Fig. 2). signs of dermatitis or diarrhea. We examined the colon (Supple- However, the levels of serum cortisol were not reduced. mental Fig. 1A), skin, and oral mucosal surfaces (Supple- We then tested the mice for the development of human ANAs mental Fig. 1B), and did not find infiltrating lymphocytes, indicat- using a standard clinical assay with HEp-2 cells. Positive ANA ingthatGVHDhadnotcausedtheweightloss. staining (scale +1 to +4) of at least grade +2 was found more frequently in serum (7/11) of ipilimumab-treated mice compared with (1/10) hIg-treated mice (p = 0.02; Fig. 5B) and with higher average titer (p , 0.05; data not shown). The majority of auto- antibodies were IgM isotype, but we also found autoantibodies against IgG in 2 of 10 mice (Fig. 5A). Effects of ipilimumab on T cell differentiation and Tregs. In gen- eral, the development of T-dependent immune responses in humanized mice has been difficult to achieve, but the finding of IgG ANAs suggested this may be occurring in the ipilimumab-treated mice. We therefore determined whether there was differentiation of T cells in the organs with immune infiltrates. The proportion of CD3+ T cells among the total lymphocytes was increased in the liver and spleen from the anti–CTLA-4 mAb–treated compared with hIg-treated mice. The increase in T cells was accounted for primarily by an expansion of the + + FIGURE 1. Systemic effects of ipilimumab (anti-CTLA mAb) treatment CD4 T cells (Fig. 6A). Among the CD4 cells, the proportion of on humanized mice. Mice treated with ipilimumab (circles) had median naive cells was reduced and the Teff fraction was increased survival of 26 d (versus hIg [squares]; p = 0.0002, n = 12 in each group). (Fig.6B).CentralmemoryCD4+ T cells (CCR7+CD45RO+) 590 A HUMANIZED MOUSE MODEL OF AUTOIMMUNE DISEASE Downloaded from

FIGURE 2. Lymphoproliferation in humanized mice treated with ipilimumab. (A) Hypersplenism was seen in the ipilimumab-treated mice after 4 wk of treatment. http://www.jimmunol.org/ Asinglepairof11isshown.(B) The increase in the size of the spleen was associated with an increase in the total number of lymphocytes in the organ (37 6 7.1 3 106 [n = 11] compared with hIg [n = 12] 14.58 6 1.9 3 106; p = 0.0006). (C) The lymph nodes were also enlarged in the ipilimumab versus control Ig-treated mice. Axillary nodes are shown. Immunohistochemical staining of axillary lymph nodes from ipilimumab-treated humanized mice: (a) H&E staining showing lymph node without evidence of organization (original magnification 320). (b) H&E higher magnification image (original magnification 340). (c) Human CD4 within lymph node (brown stain, red arrows). (d)CD8.(e)CD19.(f) There was little staining for hCD68 (all original magnifications 340). An individual animal is shown that is representative of sixineachgroup.(D) Immunohistochemical staining of axillary lymph nodes from hIg-treated mice; (a)CD4,(b)CD8,and(c) human CD68. ***p , 0.001 (p,0.001). were significantly increased in the liver (p , 0.05). These Similar to the expansion of CD4+ Teff cells, there was an increase findings suggest that there had been T cell activation and mem- in the total number of Tregs (CD4+CD25+CD127loFOXP3+Helios+) by guest on September 23, 2021 ory cell development by CD4+ cells in the liver and spleen. overall. However, these cells showed different proportions and distri-

FIGURE 3. Phenotype of CD4+ T cells after ipilimumab treatment. (A) The expression of CXCR3, PD-1, ICOS, and CD25 are shown on CD4+ splenocytes after 4 wk of treatment. *p , 0.05, **p , 0.01, ***p , 0.001. (B) Proinflammatory cytokines IFN-g and TNF are elevated in the systemic circulation after treatment with ipilimumab (hIg: n = 14, ipilimumab: n =13).(C)IFN-g and TNF production in mice treated with hIg or anti–CTLA-4 mAb after stimulation with PMA/Ionomycin. There were more T cells that coexpressed IFN-g and TNF in CD4+ T cells of spleen in the ipilimumab-treated mice (*p , 0.05) (D). (E) Spleen cells stimulated with IL-2 (100 U/105 cells) showed enhanced phosphorylation of STAT5 in CD4+ and CD8+ T cells (*p , 0.05, **p , 0.01; each dot represents an individual mouse). The Journal of Immunology 591 Downloaded from

FIGURE 4. Ipilimumab-treated humanized mice show cellular infiltrates in the liver and development of autoimmune hepatitis. (A) Histology and immunohistochemistry of liver showing human lymphocytic infiltration of liver in ipilimumab-treated (c–h) but not hIg-treated (a, b) animals. The liver was heavily infiltrated with lymphocytes (arrows) in mice treated with ipilimumab seen with H&E staining [arrows; original magnification 320 (c); 340 (d) when compared with hIg-treated controls [original magnification 320 (a), 340 (b)]. There was lobar hepatitis accompanied by ballooning and rosetting of http://www.jimmunol.org/ periportal hepatocytes. There was infiltration of human monocytes (CD68) that had formed multinucleated giants cells (e, original magnification 340) and an infiltrate of human CD4 (f) and CD8 cells (g) with limited evidence of B lymphocytes (CD19) (h, original magnification all 340). (B) Murine macrophages (F4/80+) are infrequent in hIg-treated mice but were found (arrow) in the liver of anti–CTLA-4–treated mice. (C) Serum levels of liver enzyme (ALT) and albumin. Each point represents an individual mouse (*p = 0.01 and 0.009, respectively). (D) There were very few CD4+ (a), CD8+ (b), and macrophages (arrow) (c) in the of the hIg-treated animals. bution in secondary lymph structures and tissue. There was a 4.5-fold with the equivalent dose (5 mg) as that used to treat patients with increase in the percentage of Tregs in the lymph nodes, but not in the new-onset type 1 diabetes (30–34). The mice treated with both by guest on September 23, 2021 liver (Fig. 7A; p , 0.01). In addition, the ratio of the Treg/Teff was teplizumab and ipilimumab did not lose weight and showed im- .3.5-fold lower in the livers from mice that lost weight compared with proved survival compared with mice treated with anti–CTLA-4 those that did not (Fig. 7B). alone (p = 0.0005; Fig. 8A). Anti-CD3 treatment reduced the infiltration of T cells in the liver to levels that were similar to hIg- Effects of anti-CD3 mAb treatment on autoimmunity induced with treated mice (Fig. 8B). In addition, the increased levels of ALT anti–CTLA-4 mAb. Our findings suggested that the autoimmunity triggered by anti-CTLA4 were reduced and liver function was re- in the ipilimumab-treated humanized mice was due to the activation stored by anti-CD3, as measured by albumin levels (Fig. 8C and of the human T cells. To test this hypothesis, we cotreated mice 8D). In addition, the production of inflammatory mediators, IFN-g with ipilimumab and FcR nonbinding anti-CD3 mAb (teplizumab) and TNF, was reduced by anti-CD3 mAb treatment (p , 0.05;

FIGURE 5. Ipilimumab-treated humanized mice develop ANA. (A) IgM and IgG ANAs were tested using standard laboratory methods. Identification of IgM (a) and IgG (b) ANAs on HEP2 cells. (c) Staining for hIgM Abs in an hIg-treated mouse. Original mag- nification 340. (B) There was a significantly greater proportion of ANA scores .1 in the anti–CTLA-4 mAb–treated mice (p = 0.02). *p , 0.05. 592 A HUMANIZED MOUSE MODEL OF AUTOIMMUNE DISEASE

FIGURE 6. Number and pheno- type of infiltrating T cells after ipili- mumab treatment. (A)FACSanalysis of human lymphocytes isolated from liver and spleen. There was a relative increase in CD4+ T cells and a de- crease in B cells at the three sites (*p , 0.05, **p , 0.01, ***p , 0.001; n = 6 hIg and 11 anti–CTLA-4 mAb). (B) Ipilimumab treatment de- + creased the proportion of naive CD4 Downloaded from T cells and increased the proportion of Teffs in the liver and spleen. The pro- portion of central memory CD4+ T cells increased significantly in the liver. http://www.jimmunol.org/

Fig. 8E and 8F). There was also a reduction in the levels of anti- has led to its development as a biologic to enhance immune- hIgM ANAs (p , 0.05) with no detection of IgG ANAs (Fig. 8G). mediated killing of tumors such as malignant melanoma (28, 38, by guest on September 23, 2021 The frequency of hCD68+ cells was reduced in the livers of mice 39). At the same time, the frequency of immune-related adverse treated with both mAbs (Fig. 9A); in addition, the levels of circulating events has been reported as high as 64.2% in patients receiving IP-10 were reduced, consistent with reduced activation of mononu- ipilimumab. Common findings are enterocolitis, dermatitis, hep- clear cells (Fig. 9B). Finally, we compared the frequency of Tregs in atitis, hypophysitis, adenopathy, as well as adrenalitis (38, 40, 41). the livers of the mice treated with anti–CTLA-4 and anti–CTLA-4+ Similarly, our humanized mouse model treated with ipilimumab anti-CD3 mAbs. The frequency of CD4+CD25+CD1272 Tregs was recapitulated some of these pathologies found in treated humans. increased in the spleen in the mice treated with both mAbs, and these Thus, the present model is relevant to define immunologic path- cells showed increased constitutive phosphorylation of STAT5, sug- ways of untoward effects of this therapeutic modality. gesting in vivo activation (Fig. 9D). In our studies, we exploited the known adverse effects of anti– CTLA-4 to develop a model of autoimmune disease caused by Discussion human immune cells. The critical role of human T cells in this Efforts to develop new treatments for human autoimmune disease pathology is illustrated by cotreatment with teplizumab, a non- have been hampered by the lack of small-animal models that can FcR binding anti-CD3 mAb that has been shown to ameliorate the predict the safety and efficacy of interventions in humans. The progression of type 1 diabetes. The anti–CTLA-4 mAb caused ability of anti–CTLA-4 mAbs to prevent negative costimulation weight loss and autoimmune hepatitis associated with the devel-

FIGURE 7. Reduced Tregs in liver with weight loss. (A) The relative proportion of Tregs (CD4+CD25+ CD127loFOXP3+)inthemesenteric lymph nodes and liver is shown (**p , 0.01, ***p , 0.001, n =5hIgand 9 anti–CTLA-4 mAb–treated mice). (B) The ratio of Treg/Teff was rela- tively greater in the anti–CTLA-4 mAb–treated mice that did not lose weight. The Journal of Immunology 593 Downloaded from

FIGURE 8. Coadministration of anti-CD3 with anti–CTLA-4 modulates disease induced by anti–CTLA-4 mAb. (A) Mice treated with anti–CTLA-4 http://www.jimmunol.org/ mAb (n = 6) showed weight loss and had a median survival of 14.5 d, whereas mice injected with anti-CD3 and anti-CTLA-4 mAbs (n = 6) did not lose weight (p = 0.002). (B) Infiltration of T cells in to the liver was reduced when anti-CD3 mAb was given with anti–CTLA-4 mAb. Each point represents a single mouse. p = 0.05, hIg versus hIg+aCTLA-4 mAb; p = 0.07, hIg+aCTLA-4 versus aCD3+aCTLA-4 mAb). (C and D) Serum levels of ALT and albumin were reduced and increased, respectively, when anti-CD3 mAb was given with anti–CTLA-4 mAb (p , 0.05). (E and F) There was reduced frequency of CD4+IFN+ and CD4+IFN-g+/TNF+ T cells in the livers of the mice treated with both mAbs. (G) The titers of anti-IgM ANAs were also reduced (*p , 0.05, ***p , 0.001). opment of ANAs. We found no evidence of arthritis or renal lesions in the gastrointestinal tract. The weight loss in the hu- disease, which are also rare findings in humans treated with manized mice was most likely the result of cachexia from T cell by guest on September 23, 2021 anti-CTLA4. Both IgG and IgM ANAs were identified in the activation and cytokine production rather than colitis, and it was ipilimumab-treated mice, and this classical clinical autoimmune reversed with teplizumab treatment. manifestation appeared to be dependent on T cells because treat- CTLA-4–deficient mice develop lymphoproliferation, myocar- ment with anti-CD3 mAb prevented their induction. We also found ditis, and pancreatitis, and succumb by 3–4 wk of age (21). infiltration of human T cells in the adrenals and salivary glands. However, although both the anti–CTLA-4–treated humanized The increase in Th1 cytokines in the anti–CTLA-4 humanized was mice and CTLA-4–deficient mice exhibit lymphoproliferation and associated with a general enhancement in T cell responsiveness increased levels of cytokine production, the tissue pathology and indicated by increased STAT5 phosphorylation in response to IL-2. nature of immune responses differ dramatically. The CTLA-4– The human cells also displayed increased expression of CXCR3, deficient mice have increased production of Th2 cytokines, no- CD25, and PD-1 consistent with activation and a Th1 phenotype tably IL-4 and IL-5 (46). In contrast, we observed an increase in (42). These findings indicate generalized activation of T cells, but TNF and IFN-g in the humanized mice, which is more reflective the pathologic changes were limited to specific tissues, which of what is observed in human autoimmune diseases and possibly suggest that the development of autoimmune disease does not reflecting intrinsic differences between T cell activation patterns simply reflect T cell activation. Likewise, the pathology was not in murine and human cells (1, 2). The CTLA-4–deficient mice due to general loss of immune tolerance because we did not detect succumb to fatal multiorgan failure, unlike humans who have been signs of GVHD in the gut, which is a common finding in NSG treated with anti–CTLA-4 mAb. CTLA-4 deficiency in mice af- mice that receive mature human T cells (43). fects the development and activation of T cells throughout the life We uncovered three mechanisms involved in the development of of the mouse, unlike patients who are treated with anti–CTLA-4 autoimmunity: general activation of T cells leading to cytokine mAb after the development of a mature immune repertoire, or indi- release and lymphoproliferation with enlarged secondary lymph viduals who acquire autoimmune diseases after childhood. Thus, nodes, the association of lymphocytes with macrophages at the site this humanized model system more closely recapitulates human of tissue destruction, and failure to localize Tregs to the tissues. clinical disease than the CTLA-4–deficient mice. These studies are supported by related systemic autoimmune syn- The reduction in Treg/Teff cells that we observed may shift the dromes (SLE) in humans and mouse models that are marked by aber- balance to promote autoimmune pathology. This notion is sup- rant T cell hyperproliferation and ANAs, both diagnostic of this syn- ported by the finding that the ratio of Treg/Teff was reduced in mice drome (29, 44, 45). To our knowledge, this is the first described model that lost weight after ipilimumab treatment compared with those of autoimmune disease involving human cells in vivo. that did not. Moreover, there was an increase in the total number However, we also observed some differences in the responses of of Tregs and as a proportion of the total CD4+ cells in the lymph the humanized mice to ipilimumab compared with patients with nodes, but not in the liver, suggesting that the Tregs failed to treated with the drug including the absence of inflammatory migrate to the liver. Blockade of CTLA-4 signaling may interfere 594 A HUMANIZED MOUSE MODEL OF AUTOIMMUNE DISEASE Downloaded from http://www.jimmunol.org/

FIGURE 9. Anti-CD3 treatment reduced macrophage infiltration and activation, and improves Treg function in humanized mice treated with anti–CTLA-4 mAb: (A) human macrophage (CD68+, arrows) are fewer in the liver after anti-CD3 and anti–CTLA-4 treatment. The histology of a single mouse representative of 11 is shown. Original magnification 320, 340 (inset). (B) The levels of human IP-10 were reduced in the serum (**p , 0.01). (C) The frequency of Tregs (CD4+FOXP3+CD127lo cells/total CD4+) was increased in the liver (*p , 0.05). (D) There was enhanced constitutive phosphorylation of pSTAT5 in vivo in Tregs (*p , 0.05). by guest on September 23, 2021 with the generation of both CD4+ and CD8+ T cells (47, 48), as livers of ipilimumab-treated mice and a reduced frequency when well as the CD28-driven migration of memory T cells to extra- they were treated with teplizumab, suggesting there may be pro- lymphoid tissue (49). Our findings support a similar regulation of ductive interactions between human cells and murine macro- migration of Tregs, thus explaining their failure to migrate to the phages because many of the murine ligands can cross react with liver. In addition, it is possible that increases in the proin- their receptors on T cells such as B7.1 and B7.2 (56). flammatory cytokine TNF interfere with Treg suppressive function, In conclusion, we report the development of a small-animal as has been described in rheumatoid arthritis (50) and antibiotic model of human autoimmune hepatitis that could be controlled refractory Lyme arthritis (51). Finally, teplizumab treatment ap- by teplizumab. Our studies have identified several mechanisms peared to increase Treg number and restore function in a manner associated with disease, including T cell activation and expansion, not unlike we and others previously demonstrated in treated human cytokine release, a failure of Tregs to migrate to tissues, and par- patients (52–54). We found that there were increased numbers of ticipation of murine macrophages. These studies suggest potential Tregs and enhanced function reflected by the increase in constitu- therapeutic strategies to prevent tissue damage or enhance Treg tively phosphorylated STAT5. We previously have reported that activity in patients treated with ipilimumab, and the model may teplizumab induces Tregs in humanized mice via transit of the cells also be useful for testing agents that may be used for the treatment through the gut. Consistent with this mechanism, we found in- of human autoimmune hepatitis. Further studies in these mice may creased expression of CCR6 on CD4+ and CD8+ Teffs in the enhance our understanding of the mechanisms of human immune- periphery (data not shown). mediated autoimmune disease. Tissue destruction appeared to depend on the recruitment of macrophages to the sites of pathology and their interaction with the Disclosures human T cells. These macrophages are the likely source of CXCL10 The authors have no financial conflicts of interest. that we found was elevated in the ipilimumab-treated mice and reduced when these mice were treated with teplizumab. A limitation References for modeling human immune responses in NSG mice in general has 1. Seok, J., H. S. Warren, A. G. Cuenca, M. N. Mindrinos, H. V. Baker, W. Xu, been their failure to reconstitute human innate immune cells (55). D. R. Richards, G. P. McDonald-Smith, H. Gao, L. Hennessy, et al; Inflammation The recruitment of macrophages to the liver of the ipilimumab- and Host Response to Injury, Large Scale Collaborative Research Program. 2013. Genomic responses in mouse models poorly mimic human inflammatory treated mice is most likely due to products elaborated by acti- diseases. Proc. Natl. Acad. Sci. USA 110: 3507–3512. vated T cells. Indeed, we did not identify infiltration of human 2. Mestas, J., and C. C. Hughes. 2004. Of mice and not men: differences between CD68+ cells in the livers of mice that had not been treated with mouse and human immunology. J. Immunol. 172: 2731–2738. 3. Unger, W. W., T. Pearson, J. R. Abreu, S. Laban, A. R. van der Slik, S. M. der ipilimumab or were cotreated with teplizumab. Interestingly, we Kracht, M. G. Kester, D. V. Serreze, L. D. Shultz, M. Griffioen, et al. 2012. Islet- also found an increased proportion of murine macrophages in the specific CTL cloned from a type 1 diabetes patient cause beta-cell destruction The Journal of Immunology 595

after engraftment into HLA-A2 transgenic NOD/scid/IL2RG null mice. PLoS patterns, immune-related adverse events, and their management. Semin. ONE 7: e49213. Oncol. 37: 485–498. 4. Kuhn, C., S. You, F. Valette, G. Hale, P. van Endert, J. F. Bach, H. Waldmann, 29. Zielinski, C. E., S. N. Jacob, F. Bouzahzah, B. E. Ehrlich, and J. Craft. 2005. and L. Chatenoud. 2011. Human CD3 transgenic mice: preclinical testing of Naive CD4+ T cells from lupus-prone Fas-intact MRL mice display TCR- antibodies promoting immune tolerance. Sci. Transl. Med. 3: 68ra10. mediated hyperproliferation due to intrinsic threshold defects in activation. J. 5. Rathinam, C., W. T. Poueymirou, J. Rojas, A. J. Murphy, D. M. Valenzuela, Immunol. 174: 5100–5109. G. D. Yancopoulos, A. Rongvaux, E. E. Eynon, M. G. Manz, and R. A. Flavell. 30. Herold, K. C., S. E. Gitelman, M. R. Ehlers, P. A. Gottlieb, C. J. Greenbaum, 2011. Efficient differentiation and function of human macrophages in humanized W. Hagopian, K. D. Boyle, L. Keyes-Elstein, S. Aggarwal, D. Phippard, et al; CSF-1 mice. Blood 118: 3119–3128. AbATE Study Team. 2013. Teplizumab (anti-CD3 mAb) treatment preserves C- 6. Banuelos, S. J., L. D. Shultz, D. L. Greiner, L. M. Burzenski, B. Gott, peptide responses in patients with new-onset type 1 diabetes in a randomized B. L. Lyons, A. A. Rossini, and M. C. Appel. 2004. Rejection of human islets controlled trial: metabolic and immunologic features at baseline identify a sub- and human HLA-A2.1 transgenic mouse islets by alloreactive human lymphocytes group of responders. Diabetes 62: 3766–3774. in immunodeficient NOD-scid and NOD-Rag1(null)Prf1(null) mice. Clin. Immunol. 31. Herold, K. C., S. E. Gitelman, U. Masharani, W. Hagopian, B. Bisikirska, 112: 273–283. D. Donaldson, K. Rother, B. Diamond, D. M. Harlan, and J. A. Bluestone. 2005. 7. Libby, S. J., M. A. Brehm, D. L. Greiner, L. D. Shultz, M. McClelland, A single course of anti-CD3 monoclonal antibody hOKT3gamma1(Ala-Ala) K. D. Smith, B. T. Cookson, J. E. Karlinsey, T. L. Kinkel, S. Porwollik, et al. results in improvement in C-peptide responses and clinical parameters for at 2010. Humanized nonobese diabetic-scid IL2rgammanull mice are susceptible to least 2 years after onset of type 1 diabetes. Diabetes 54: 1763–1769. lethal Salmonella Typhi infection. Proc. Natl. Acad. Sci. USA 107: 15589–15594. 32. Herold, K. C., S. E. Gitelman, S. M. Willi, P. A. Gottlieb, F. Waldron-Lynch, 8. Shultz, L. D., F. Ishikawa, and D. L. Greiner. 2007. Humanized mice in trans- L. Devine, J. Sherr, S. M. Rosenthal, S. Adi, M. Y. Jalaludin, et al. 2013. lational biomedical research. Nat. Rev. Immunol. 7: 118–130. Teplizumab treatment may improve C-peptide responses in participants with 9. Shultz, L. D., T. Pearson, M. King, L. Giassi, L. Carney, B. Gott, B. Lyons, type 1 diabetes after the new-onset period: a randomised controlled trial. A. A. Rossini, and D. L. Greiner. 2007. Humanized NOD/LtSz-scid IL2 receptor Diabetologia 56: 391–400. common gamma chain knockout mice in diabetes research. Ann. N. Y. Acad. Sci. 33. Herold, K. C., W. Hagopian, J. A. Auger, E. Poumian-Ruiz, L. Taylor, 1103: 77–89. 10. Linsley, P. S., W. Brady, M. Urnes, L. S. Grosmaire, N. K. Damle, and J. A. Ledbetter. D. Donaldson, S. E. Gitelman, D. M. Harlan, D. Xu, R. A. Zivin, and 1991. CTLA-4 is a second receptor for the B cell activation antigen B7. J. Exp. Med. J. A. Bluestone. 2002. Anti-CD3 monoclonal antibody in new-onset type 1 di- Downloaded from 174: 561–569. abetes mellitus. N. Engl. J. Med. 346: 1692–1698. 11. Walunas, T. L., D. J. Lenschow, C. Y. Bakker, P. S. Linsley, G. J. Freeman, 34. Sherry, N., W. Hagopian, J. Ludvigsson, S. M. Jain, J. Wahlen, R. J. Ferry, Jr., B. J. M. Green, C. B. Thompson, and J. A. Bluestone. 1994. CTLA-4 can function Bode, S. Aronoff, C. Holland, D. Carlin, et al. 2011. Teplizumab for treatment of as a negative regulator of T cell activation. Immunity 1: 405–413. type 1 diabetes (Protege study): 1-year results from a randomised, placebo- 12. Brunet, J. F., F. Denizot, M. F. Luciani, M. Roux-Dosseto, M. Suzan, controlled trial. Lancet. 378: 487–497. M. G. Mattei, and P. Golstein. 1987. A new member of the immunoglobulin 35. Weber, J. S., S. O’Day, W. Urba, J. Powderly, G. Nichol, M. Yellin, J. Snively, superfamily—CTLA-4. Nature 328: 267–270. and E. Hersh. 2008. Phase I/II study of ipilimumab for patients with metastatic

13. Walunas, T. L., C. Y. Bakker, and J. A. Bluestone. 1996. CTLA-4 ligation blocks melanoma. J. Clin. Oncol. 26: 5950–5956. http://www.jimmunol.org/ CD28-dependent T cell activation. J. Exp. Med. 183: 2541–2550. 36. Waldron-Lynch, F., O. Henegariu, S. Deng, P. Preston-Hurlburt, J. Tooley, R. 14. van der Merwe, P. A., D. L. Bodian, S. Daenke, P. Linsley, and S. J. Davis. 1997. Flavell, and K. C. Herold. 2012. Teplizumab induces human gut-tropic regula- CD80 (B7-1) binds both CD28 and CTLA-4 with a low affinity and very fast tory cells in humanized mice and patients. Sci. Transl. Med. 4: 118ra112. kinetics. J. Exp. Med. 185: 393–403. 37. Vudattu, N. K., I. Magalhaes, M. Schmidt, V. Seyfert-Margolis, and 15. Ostrov, D. A., W. Shi, J. C. Schwartz, S. C. Almo, and S. G. Nathenson. 2000. M. J. Maeurer. 2007. Reduced numbers of IL-7 receptor (CD127) expressing Structure of murine CTLA-4 and its role in modulating T cell responsiveness. immune cells and IL-7-signaling defects in peripheral blood from patients with Science 290: 816–819. . Int. J. Cancer 121: 1512–1519. 16. Marenge`re, L. E., P. Waterhouse, G. S. Duncan, H. W. Mittrucker,€ G. S. Feng, 38. Della Vittoria Scarpati, G., C. Fusciello, F. Perri, F. Sabbatino, S. Ferrone, and T. W. Mak. 1996. Regulation of T cell receptor signaling by tyrosine C. Carlomagno, and S. Pepe. 2014. Ipilimumab in the treatment of metastatic phosphatase SYP association with CTLA-4. Science 272: 1170–1173. melanoma: management of adverse events. Onco. Targets Ther. 7: 203–209. 17. Chuang, E., T. S. Fisher, R. W. Morgan, M. D. Robbins, J. M. Duerr, 39. Hodi, F. S., S. J. O’Day, D. F. McDermott, R. W. Weber, J. A. Sosman, M. G. Vander Heiden, J. P. Gardner, J. E. Hambor, M. J. Neveu, and J. B. Haanen, R. Gonzalez, C. Robert, D. Schadendorf, J. C. Hassel, et al. 2010. by guest on September 23, 2021 C. B. Thompson. 2000. The CD28 and CTLA-4 receptors associate with the Improved survival with ipilimumab in patients with metastatic melanoma. N. serine/threonine phosphatase PP2A. Immunity 13: 313–322. Engl. J. Med. 363: 711–723. 18. Lee, K. M., E. Chuang, M. Griffin, R. Khattri, D. K. Hong, W. Zhang, D. Straus, 40. Min, L., and N. Ibrahim. 2013. Ipilimumab-induced autoimmune adrenalitis. L. E. Samelson, C. B. Thompson, and J. A. Bluestone. 1998. Molecular basis of Lancet Diabetes Endocrinol. 1: e15. T cell inactivation by CTLA-4. Science 282: 2263–2266. 41. Dillard, T., C. G. Yedinak, J. Alumkal, and M. Fleseriu. 2010. Anti-CTLA-4 19. Waterhouse, P., J. M. Penninger, E. Timms, A. Wakeham, A. Shahinian, antibody therapy associated autoimmune hypophysitis: serious immune related K. P. Lee, C. B. Thompson, H. Griesser, and T. W. Mak. 1995. Lymphoproli- adverse events across a spectrum of cancer subtypes. Pituitary 13: 29–38. ferative disorders with early lethality in mice deficient in Ctla-4. Science 270: 42. Antonelli, A., S. M. Ferrari, D. Giuggioli, E. Ferrannini, C. Ferri, and P. Fallahi. 985–988. 2014. Chemokine (C-X-C motif) ligand (CXCL)10 in autoimmune diseases. 20. Chambers, C. A., T. J. Sullivan, and J. P. Allison. 1997. Lymphoproliferation in Autoimm. Rev. 13: 272–280. CTLA-4-deficient mice is mediated by costimulation-dependent activation of 43. Pino, S., M. A. Brehm, L. Covassin-Barberis, M. King, B. Gott, T. H. Chase, CD4+ T cells. Immunity 7: 885–895. J. Wagner, L. Burzenski, O. Foreman, D. L. Greiner, and L. D. Shultz. 2010. 21. Tivol, E. A., F. Borriello, A. N. Schweitzer, W. P. Lynch, J. A. Bluestone, and Development of novel major histocompatibility complex class I and class II- A. H. Sharpe. 1995. Loss of CTLA-4 leads to massive lymphoproliferation and deficient NOD-SCID IL2R gamma chain knockout mice for modeling hu- fatal multiorgan tissue destruction, revealing a critical negative regulatory role of man xenogeneic graft-versus-host disease. Methods Mol. Biol. 602: 105– CTLA-4. Immunity 3: 541–547. 117. 22. Takahashi, T., T. Tagami, S. Yamazaki, T. Uede, J. Shimizu, N. Sakaguchi, 44. Zhu, J., X. Liu, C. Xie, M. Yan, Y. Yu, E. S. Sobel, E. K. Wakeland, and T. W. Mak, and S. Sakaguchi. 2000. Immunologic self-tolerance maintained by C. Mohan. 2005. T cell hyperactivity in lupus as a consequence of hyper- CD25(+)CD4(+) regulatory T cells constitutively expressing cytotoxic stimulatory antigen-presenting cells. J. Clin. Invest. 115: 1869–1878. T lymphocyte-associated antigen 4. J. Exp. Med. 192: 303–310. 45. Hansen, A., P. E. Lipsky, and T. Do¨rner. 2007. B-cell lymphoproliferation in 23. Brunner, M. C., C. A. Chambers, F. K. Chan, J. Hanke, A. Winoto, and chronic inflammatory rheumatic diseases. Nat. Clin. Pract. Rheumatol. 3: 561– J. P. Allison. 1999. CTLA-4-mediated inhibition of early events of T cell pro- liferation. J. Immunol. 162: 5813–5820. 569. 24. Jain, N., H. Nguyen, C. Chambers, and J. Kang. 2010. Dual function of 46. Khattri, R., J. A. Auger, M. D. Griffin, A. H. Sharpe, and J. A. Bluestone. 1999. CTLA-4 in regulatory T cells and conventional T cells to prevent multiorgan Lymphoproliferative disorder in CTLA-4 knockout mice is characterized by autoimmunity. Proc. Natl. Acad. Sci. USA 107: 1524–1528. CD28-regulated activation of Th2 responses. J. Immunol. 162: 5784–5791. 25. Yamaguchi, T., A. Kishi, M. Osaki, H. Morikawa, P. Prieto-Martin, K. Wing, 47. Zheng, J., Y. Liu, Y. Liu, M. Liu, Z. Xiang, K. T. Lam, D. B. Lewis, Y. L. Lau, T. Saito, and S. Sakaguchi. 2013. Construction of self-recognizing regula- and W. Tu. 2013. Human CD8+ regulatory T cells inhibit GVHD and preserve tory T cells from conventional T cells by controlling CTLA-4 and IL-2 general immunity in humanized mice. Sci. Transl. Med. 5: 168ra9. expression. Proc. Natl. Acad. Sci. USA 110: E2116–E2125. 48. Verhagen, J., R. Genolet, G. J. Britton, B. J. Stevenson, C. A. Sabatos-Peyton, 26.Phan,G.Q.,J.C.Yang,R.M.Sherry,P.Hwu,S.L.Topalian, J. Dyson, I. F. Luescher, and D. C. Wraith. 2013. CTLA-4 controls the thymic D. J. Schwartzentruber, N. P. Restifo, L. R. Haworth, C. A. Seipp, development of both conventional and regulatory T cells through modulation of L. J. Freezer, et al. 2003. Cancer regression and autoimmunity induced by the TCR repertoire. Proc. Natl. Acad. Sci. USA 110: E221–E230. cytotoxic T lymphocyte-associated antigen 4 blockade in patients with 49. Mirenda, V., S. J. Jarmin, R. David, J. Dyson, D. Scott, Y. Gu, R. I. Lechler, metastatic melanoma. Proc. Natl. Acad. Sci. USA 100: 8372–8377. K. Okkenhaug, and F. M. Marelli-Berg. 2007. Physiologic and aberrant regu- 27. Quezada, S. A., K. S. Peggs, M. A. Curran, and J. P. Allison. 2006. CTLA4 lation of memory T-cell trafficking by the costimulatory molecule CD28. Blood blockade and GM-CSF combination immunotherapy alters the intratumor bal- 109: 2968–2977. ance of effector and regulatory T cells. J. Clin. Invest. 116: 1935–1945. 50. Nie, H., Y. Zheng, R. Li, T. B. Guo, D. He, L. Fang, X. Liu, L. Xiao, X. Chen, 28. Kaehler, K. C., S. Piel, E. Livingstone, B. Schilling, A. Hauschild, and B. Wan, et al. 2013. Phosphorylation of FOXP3 controls regulatory T cell D. Schadendorf. 2010. Update on immunologic therapy with anti-CTLA-4 function and is inhibited by TNF-a in rheumatoid arthritis. Nat. Med. 19: antibodies in melanoma: identification of clinical and biological response 322–328. 596 A HUMANIZED MOUSE MODEL OF AUTOIMMUNE DISEASE

51. Vudattu, N. K., K. Strle, A. C. Steere, and E. E. Drouin. 2013. Dysregulation of 54. Bisikirska, B., J. Colgan, J. Luban, J. A. Bluestone, and K. C. Herold. 2005. TCR CD4+CD25(high) T cells in the synovial fluid of patients with antibiotic- stimulation with modified anti-CD3 mAb expands CD8+ T cell population and refractory Lyme arthritis. Arthritis Rheum. 65: 1643–1653. induces CD8+CD25+ Tregs. J. Clin. Invest. 115: 2904–2913. 52. Ablamunits, V., O. Henegariu, P. Preston-Hurlburt, and K. C. Herold. 2011. 55. Rongvaux, A., T. Willinger, J. Martinek, T. Strowig, S. V. Gearty, NKG2A is a marker for acquisition of regulatory function by human CD8+ L. L. Teichmann, Y. Saito, F. Marches, S. Halene, A. K. Palucka, et al. 2014. Development and function of human innate immune cells in a humanized mouse T cells activated with anti-CD3 antibody. Eur. J. Immunol. 41: 1832–1842. model. Nat. Biotechnol. 32: 364–372. 53. Belghith, M., J. A. Bluestone, S. Barriot, J. Me´gret,J.F.Bach,and 56. Lenschow, D. J., S. C. Ho, H. Sattar, L. Rhee, G. Gray, N. Nabavi, K. C. Herold, L. Chatenoud. 2003. TGF-beta-dependent mechanisms mediate restoration of and J. A. Bluestone. 1995. Differential effects of anti-B7-1 and anti-B7-2 self-tolerance induced by antibodies to CD3 in overt autoimmune diabetes. Nat. monoclonal antibody treatment on the development of diabetes in the non- Med. 9: 1202–1208. obese diabetic mouse. J. Exp. Med. 181: 1145–1155. Downloaded from http://www.jimmunol.org/ by guest on September 23, 2021