Manipulation of CD98 Resolves Type 1 Diabetes in Nonobese Diabetic Mice Gaojian Lian, Hideki Arimochi, Akiko Kitamura, Jun Nishida, Shigen Li, Kenji Kishihara, Yoichi Maekawa and This information is current as Koji Yasutomo of September 29, 2021. J Immunol 2012; 188:2227-2234; Prepublished online 30 January 2012; doi: 10.4049/jimmunol.1102586

http://www.jimmunol.org/content/188/5/2227 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2012/01/30/jimmunol.110258 Material 6.DC1 http://www.jimmunol.org/ References This article cites 27 articles, 2 of which you can access for free at: http://www.jimmunol.org/content/188/5/2227.full#ref-list-1

Why The JI? Submit online.

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

by guest on September 29, 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 © 2012 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Manipulation of CD98 Resolves Type 1 Diabetes in Nonobese Diabetic Mice

Gaojian Lian, Hideki Arimochi, Akiko Kitamura, Jun Nishida, Shigen Li, Kenji Kishihara, Yoichi Maekawa, and Koji Yasutomo

The interplay of CD4+ and CD8+ T cells targeting autoantigens is responsible for the progression of a number of autoimmune diseases, including type 1 diabetes mellitus (T1D). Understanding the molecular mechanisms that regulate T cell activation is crucial for designing effective therapies for autoimmune diseases. We probed a panel of Abs with T cell-modulating activity and identified a mAb specific for the H chain of CD98 (CD98hc) that was able to suppress T cell proliferation. The anti-CD98hc mAb also inhibited Ag-specific proliferation and the acquisition of effector function by CD4+ and CD8+ T cells in vitro and in vivo. Injection of the anti-CD98hc mAb completely prevented the onset of cyclophosphamide-induced diabetes in NOD mice. Treatment of diabetic NOD mice with anti-CD98hc reversed the diabetic state to normal levels, coincident with decreased proliferation of Downloaded from CD4+ T cells. Furthermore, treatment of diabetic NOD mice with CD98hc small interfering RNA resolved T1D. These data indicate that strategies targeting CD98hc might have clinical application for treating T1D and other T cell-mediated autoimmune diseases. The Journal of Immunology, 2012, 188: 2227–2234.

oth innate and adaptive immune responses control T cell diabetes is typically diagnosed in NOD mice older than 12 wk of ∼ responses to foreign and autoantigens (1, 2). The activa- age, at a time when 90% of the b cell mass has been destroyed. http://www.jimmunol.org/ B tion of innate immunity influences APCs to express T cell CD98/4F2 is a disulfide-linked complex composed of a glyco- costimulatory molecules, which are essential for effective T cell- sylated H chain of CD98 (CD98hc) and a nonglycosylated L chain of mediated adaptive immune responses. A variety of costimulatory CD98 (12). CD98hc is a type II transmembrane con- receptors expressed on T cells can positively or negatively regulate sisting of 529 aa, with many proposed biological functions, such as the proliferation, functional differentiation, and survival of T cells an transport when paired with the L chain of CD98, and (3–6). A number of reports investigating the experimental manip- a regulator of integrin-mediated cell adhesion, in addition to its ulation of these molecular interactions have demonstrated effective demonstrated role in regulating fusion (12–14). It has been reported modulation of T cell function, and this has led investigators to that an Ab specific for mouse CD98hc is able to suppress the Con generate panels of reagents, such as mAbs or recombinant soluble A-mediated activation of mouse T cells (15). Furthermore, a recent by guest on September 29, 2021 ligands, that target each costimulatory receptor/ligand pair, to es- article has indicated that genetic deletion of CD98hc in T cells tablish strategies for treating T cell-mediated autoimmune diseases. negatively impacts T cell clonal expansion and suppresses the de- Type 1 diabetes mellitus (T1D) is a progressive autoimmune velopment of experimental acute encephalomyelitis and T1D in- disease characterized by mononuclear cell infiltration into pan- duced by adoptively transferring pathogenic T cells (16). However, creatic islets, and the subsequent destruction of insulin-producing it remains unclear if CD98hc will be a good target molecule for b cells (2, 7). NOD mice recapitulate many features of T1D patho- treating spontaneously progressive autoimmune diseases. genesis in humans (8, 9). Many studies using these mice have In this study, we systematically investigated molecules that demonstrated that the early stages of the diabetogenic response modulate T cell function by establishing a panel of mAbs that re- involve effector T cells targeting b cell Ags, including glutamic cognize spleen cells and have the ability to modulate T cell func- acid decarboxylase 65, insulin, islet amyloid polypeptide, chro- tion. One Ab that highly suppressed T cell proliferation recognized mogranin A, islet-specific glucose-6-phosphatase catalytic subunit- CD98hc. The anti-CD98hc mAb suppressed proliferation and the related , and islet antigen 2 (10). Insulitis or islet inflam- functional differentiation of both CD4+ and CD8+ T cells. This Ab mation is detectable in NOD mice at ∼4 wk of age (7, 8, 11). Overt resolved T1D in NOD mice by inhibiting T cell responses. Fur- thermore, treatment of NOD mice with small interfering RNA (siRNA) against CD98hc also suppressed the progression of T1D in Department of Immunology and Parasitology, Institute of Health Biosciences, The NOD mice. These data indicate that CD98hc contributes to the University of Tokushima Graduate School, Tokushima 770-8503, Japan development of T1D and suggest that therapies targeting CD98hc, Received for publication September 7, 2011. Accepted for publication December 27, such as mAbs and siRNA, might be used clinically for preventing 2011. or treating T1D and other T cell-mediated autoimmune diseases. This work was supported by Knowledge Cluster programs from the Ministry of Education, Culture, Sports, Science and Technology, Japan. Materials and Methods Address correspondence and reprint requests to Prof. Koji Yasutomo, Department of Mice Immunology and Parasitology, Institute of Health Biosciences, The University of Tokushima Graduate School, 3-18-15 Kuramoto, Tokushima 770-8503, Japan. Female Wistar rats and C57BL/6, BALB/c, NOD, and NOD/SCID mice E-mail address: [email protected] were purchased from Japan SLC (Hamamatsu). OT-I or OT-II TCR The online version of this article contains supplemental material. transgenic mice were purchased from Taconic. All mice and rats were Abbreviations used in this article: CD98hc, H chain of CD98; siRNA, small inter- maintained under specific pathogen-free conditions in the animal research fering RNA; T1D, type 1 diabetes mellitus. center of the University of Tokushima, and all experiments were performed in accordance with institutional guidelines for animal care and use. Diabetes Copyright Ó 2012 by The American Association of Immunologists, Inc. 0022-1767/12/$16.00 incidence in NOD mice was monitored by weekly measurement of urine www.jimmunol.org/cgi/doi/10.4049/jimmunol.1102586 2228 MANIPULATION OF CD98 INHIBITS DIABETES glucose concentrations and venous blood glucose concentrations in non- (Miltenyi Biotec).CD4+or CD8+ T cells were purified from total spleen or fasting mice. Mice with blood glucose concentrations $250 mg/dl were lymph node cells by incubating with anti-CD4 (GK1.5) or anti-CD8 mAb considered diabetic. T1D developed in .80% of female NOD mice be- (53-6.72), followed by positive selection of CD4+or CD8+cells with anti- tween 20 and 25 wk of age in our colony. rat IgG MicroBeads (Miltenyi Biotec). CD4+ T cells were isolated from single-cell suspensions of spleens using the CD4+ T Cell Isolation Kit II. In Establishment and selection of hybridomas each case, purity of the cell separation was verified to be $95% by im- A total of 2 3 107 BALB/c mouse spleen cells were injected i.p. into munostaining and FACS analysis. Wistar rats three times at 10-d intervals. Total spleen cells from Wistar rats Ab treatment 5 d after the last immunization were fused with SP2 hybridoma parent cells (RIKEN Cell Bank) with PEG 1500 (Sigma-Aldrich) and cultured in hy- NOD mice 12 wk old were injected i.p. with cyclophosphamide (200 mg/kg poxanthine–aminopterin–thymidine medium. Clones with positive staining body weight) two times at 7-d intervals. The day of the second cyclo- for mouse T cells or dendritic cells were selected. Clones that were able to phosphamide injection was designated day 0. Anti-CD98hc mAb (100 mg/ suppress the proliferation of BALB/c T cells in an MLR with C57BL/6 mouse) was injected on days 0, 3, 6, 9, 12, and 15. In some experiments, splenocytes were selected for further study. diabetic mice showing glycosuria and fasting glycemia .250 mg/dl were Flow cytometry given anti-CD98hc (100 mg/mouse) or control IgG three times every other day, beginning the day the mice developed diabetes. Remission was de- Lymphocytes obtained from spleen, pancreatic lymph nodes, and pancreas fined by normal glycemia and the absence of glycosuria. C57BL/6 mice were stained with fluorochrome-conjugated Abs to CD4, CD8, TCRVa2, were immunized with OVA emulsified in CFA and were treated with anti- and TCRVb5 (eBioscience). In some experiments, cells were stimulated CD98hc or control IgG (100 mg/mouse) on days 21, 2, 5, and 8. The day with 250 ng/ml PMA (Sigma-Aldrich) and 1 mg/ml ionomycin (Sigma- of OVA immunization was designated day 0. C57BL/6 mice received Aldrich) for 5 h in the presence of monensin, followed by surface stain- CFSE-labeled T cells from OT-I or OT-II TCR transgenic mice and were ing for CD4 or CD8 and intracellular staining for granzyme B, IFN-g,or immunized with OVA protein (100 mg) or OVA peptide (257–264; 75 mg) Downloaded from IL-17 (eBioscience). 7-Aminoactinomycin D (Sigma-Aldrich) was used to emulsified in CFA immediately after T cell transfer. Recipient mice were exclude dead cells. CD25+CD4+Foxp3+ cells were identified using a mouse treated with anti-CD98hc or control IgG (200 mg/mouse) on days 1 and 2. regulatory T cell staining from eBioscience according to the manu- The day of OVA immunization was designated day 0. facturer’s instructions. Peripheral mononuclear cells from mice were stained with the I-142 mAb, followed by PE-conjugated anti-rat IgG (Jackson siRNA treatment ImmunoResearch Laboratories). Data were collected on a FACSCanto II A mixture of three siRNAs against mouse CD98hc or negative control (BD Biosciences) and analyzed using FlowJo software (Tree Star). siRNA without any target sequence (4 nmol per mouse; B-Bridge) (Sup- http://www.jimmunol.org/ T cell proliferation plemental Table I) was mixed with AteloGene Systemic (Koken), according to the manufacturer’s protocol. The siRNA transfection mix was injected Cells were stimulated with soluble or plate-bound anti-CD3 mAb (1 mg/ml; i.p. into NOD/SCID mice that had received total T cells from diabetic eBioscience) in the presence of I-143 or a control rat IgG (10 mg/ml; NOD mice. The siRNA was administered on days 2, 5, and 8 (the day when Jackson ImmunoResearch Laboratories) for 3 d. IL-12 (10 ng/ml; eBio- T cells were transferred was designated day 0). science) and anti–IL-4 (eBioscience) or TGF-b (1 ng/ml; Roche), IL-6 (10 ng/ml; Miltenyi Biotec), and anti–INF-g (4 mg/ml) were added to the Histopathology cultures for Th1- or Th17-skewing conditions, respectively. Pancreata from Tissues were fixed in buffered 10% formalin and embedded in paraffin, and diabetic NOD mice were dissected by scissors into small pieces in ice-cold then 8-mm sections were stained with H&E and scored for immune infiltrates. PBS, homogenized, and sonicated with a Sonifire sonicator (Branson Ul- For autoimmune pancreatitis, 0 = no pancreatitis; 1 = periductal infiltrates

trasonics) for 10 min. After centrifugation at 15,000 rpm for 20 min, the by guest on September 29, 2021 and focal acinar infiltrates with limited tissue destruction; 2 = periductal supernatants were used as a source of pancreatic Ags. Spleen cells from infiltrates and moderate acinar infiltrates (,50% of acinar tissue destroyed in NOD mice were stimulated with pancreatic Ags (20% of total volume) for one or more pancreatic lobes); and 3 = widespread periductal and acinar 3 d, and [3H]-thymidine uptake by the spleen cells was determined during infiltrates (.50% of acinar tissue in several pancreatic lobes destroyed). The the final 16 h of culture. Splenocytes were resuspended at a concentration number of islets in 10 fields was counted, and the mean 6 SE was calculated. of 1 3 107 cells/ml in serum-free DMEM at 37˚C. An equal volume of CFSE stock (5 mM in DMSO) in 37˚C serum-free DMEM was added to Statistical analysis the cell preparation and incubated for 5 min at 37˚C. CFSE labeling was quenched by adding an equal volume of heat-inactivated FCS. CFSE- The p values were derived using the two-tailed Student t test. Statistical labeled lymph node cells from OT-I or OT-II TCR transgenic mice were differences in survival rates between two groups with T1D development adoptively transferred into C57BL/6 mice. For T cell stimulation, spleno- were analyzed by the Kaplan–Meier log- test. Differences were con- cytes were stimulated with soluble or plate-bound anti-CD3 mAb (1 mg/ sidered significant when p , 0.05. ml). Cells were incubated for 2 d, harvested, and stained for flow cy- tometry with anti-CD8, anti-CD4, and 7-aminoactinomycin D. Results ELISA Establishment of mAbs that modulate T cell proliferation OVA (100 mg) emulsified in CFA was injected at the tail base of C57BL/6 To identify molecules that control T cell function, we first im- mice. Serum samples were obtained from peripheral blood. The anti-OVA munized Wistar rats with BALB/c spleen cells and established IgM and IgG titers were measured by mouse anti-OVA IgM and mouse . anti-OVA IgG kits (Alpha Diagnostic International) according to the 200 hybridoma cell lines that recognized mouse spleen cells. We manufacturer’s protocols. selected 43 hybridoma cells lines producing mAbs that were able to augment or inhibit T cell proliferation in an MLR assay. One Transfection hybridoma cell line was selected from the 43 because the mAb A total of 293T cells were transfected with mouse full-length CD98hc (cell line I-142) produced by this cell line was the best at inhib- cDNA in pcDNA3.1 using GeneJuice (Merck), and stable cells expressing iting T cell proliferation in an MLR assay (data not shown). CD98hc were obtained after Zeocin treatment. We first determined whether the I-142 mAb directly suppressed Immunoprecipitation T cell proliferation induced by TCR/CD3-mediated stimulation. Spleen cells from C57BL/6 mice were lysed with lysate buffer. After centrifu- I-142 addition suppressed T cell proliferation when total spleen gation, the supernatants were incubated with I-142 for 2 h at 4˚C and further cells were stimulated with soluble anti-CD3 mAb (Fig. 1A) or incubated with protein G agarose for 1 h at 4˚C. After washing the protein G when purified CD4+ or CD8+ T cells (Fig. 1B) were stimulated agarose with lysis buffer, boiled immunoprecipitants in sample buffer were with plate-bound anti-CD3 mAb. I-142 inhibited Th1 and Th17 subjected to SDS-PAGE. Gels were stained with a silver staining kit (Invitrogen). differentiation in CD4+ T cells (Fig. 1C) and suppressed granzyme Isolation of cells B-producing cytotoxic T cell differentiation in CD8+ T cells (Fig. Total T cells were purified from spleen cells by incubating with anti-FcR 1D). We next ascertained whether I-142 could suppress T cell- and were depleted of FcR-positive cells using anti-rat IgG MicroBeads mediated immune responses in vivo. I-142 (100 mg) was injected The Journal of Immunology 2229 Downloaded from http://www.jimmunol.org/

FIGURE 1. Characterization of the I-142 mAb. Total spleen cells (A) or purified CD4+ or CD8+ T cells (B) were stimulated with anti-CD3 mAb (1 mg/ ml) in the absence or presence of I-142 or control rat IgG (20 mg/ml) for 3 d. [3H]-thymidine incorporation during the final 8 h was measured. Data are shown as the mean 6 S.D. **p , 0.01. (C) Total spleen cells were stimulated with anti-CD3 mAb (1 mg/ml) in the absence or presence of I-142 or control rat IgG (20 mg/ml) for 3 d under Th1- or Th17-inducing conditions. The expression of IFN-g or IL-17 in CD4+ T cells was measured by flow cytometry. Data are shown as the mean 6 S.D. *p , 0.05, **p , 0.01. (D) Total spleen cells were stimulated with anti-CD3 mAb (1 mg/ml) in the absence or presence of I-142 or control IgG (20 mg/ml) for 3 d. The cytotoxic activity of CD8+ T cells was measured by the expression of granzyme B in CD8+ T cells. Data are shown as the mean 6 S.D. **p , 0.01. (E) C57BL/6 mice (n = 6 per group) were immunized with OVA in CFA and injected with I-142 or control IgG, 100 by guest on September 29, 2021 mg per mouse, on days 0, 2, 4, and 6 (the day on which mice were immunized was designated day 0). Anti–OVA-specific IgG or IgM on day 10 was measured by ELISA. For control, the serum titers of IgG or IgM from unimmunized mice were measured. Data are shown as the mean 6 S.D. The data shown in this figure are representative of at least three experiments. *p , 0.05, **p , 0.01. four times into mice immunized with OVA protein emulsified in subjected to SDS-PAGE. Silver staining showed a specific band CFA. We observed that I-142 inhibited the production of OVA- around 75 kDa (Fig. 2A). Liquid chromatography–mass spec- specific IgM and IgG, indicating that the I-142 mAb is able to trometry analysis of this band revealed that this band corresponds suppress Ag-specific Ab production (Fig. 1E). to the H chain of CD98 (CD98hc) (data not shown). 293T cells transfected with control vector or vector encoding the cDNA of I-142 mAb recognizes CD98hc mouse CD98hc were stained with I-142 to confirm that I-142 mAb We next identified the protein recognized by I-142. Cell lysates from recognizes CD98hc (Fig. 2B). Positive staining was observed in total spleen cells were immunoprecipitated with the I-142 mAb and 293T cells overexpressing CD98hc, whereas 293T cells transfected

FIGURE 2. I-142 recognizes the H chain of CD98. (A) Cell lysates from total spleen cells were immu- noprecipitated with control rat IgG or the I-142 mAb. The immunoprecipitants were subjected to SDS- PAGE, and gels were silver stained. (B) Mouse CD98hc cDNA was transfected into 293T cells. Control vector-transfected 293T cells (293/mock) or mouse CD98hc cDNA-transfected 293T cells (293/ mCD98hc) were stained with I-142, followed by PE- conjugated anti-rat IgG, and analyzed by flow cy- tometry. Green or pink lines indicate the cells stained with secondary Ab alone or I-142 and secondary Ab, respectively. The data shown in this figure are representative of at least three experiments. 2230 MANIPULATION OF CD98 INHIBITS DIABETES

FIGURE 3. Anti-CD98hc inhibits T cell proliferation in vivo. CFSE-labeled CD4+ T cells from OT-II TCR trans- genic mice (A) or CD8+ T cells from OT-I TCR transgenic mice (B) were transferred into C57BL/6 mice (n =5 per group) that were subsequently im- munized with OVA protein (OT-II) or OVA peptide (OT-I) in CFA and treated with control rat IgG or the I-142 mAb. CFSE dilution in CD4+TCRVa2+Vb5+ or CD8+TCRVa+2Vb5+ 4 d after im- munization was evaluated by flow cy- tometry. The number indicates percent of proliferating cells. Data are shown as the mean 6 S.D. The data shown in this figure are representative of at least three experiments. *p , 0.05. Downloaded from

with control vector were not stained by the I-142 mAb (Fig. 2B). with OVA peptide emulsified by CFA and treated with control IgG

Taken together, these data indicate that I-142 recognizes mouse or I-142. Treatment with I-142 reduced cell division (Fig. 3B). http://www.jimmunol.org/ CD98hc, although we cannot deny the possibility that I-142 rec- These data demonstrate that the I-142 mAb is capable of sup- ognizes CD98hc as a complex with other molecules. pressing both CD4+ and CD8+ T cell division in vivo. Anti-CD98hc suppresses CD4+ and CD8+ T cell proliferation I-142 administration prevents the progression of T1D in NOD in vitro and in vivo mice We determined whether or not the I-142 mAb could inhibit T cell To examine whether blockade of CD98 with I-142 suppresses proliferation and effector functions in vivo. CD4+ T cells from OT- T cell-mediated autoimmune pathological states, we made use of II TCR transgenic mice that recognize OVA peptide presented by NOD mice because they develop T1D owing to aberrant effector + + H-2 IAb were labeled with CFSE. The CFSE-labeled cells were functions of both CD4 and CD8 T cells. We first tested whether by guest on September 29, 2021 transferred into C57BL/6 mice that were subsequently immunized I-142 has a protective role for T1D in NOD mice by administering with OVA emulsified in CFA and administered the I-142 mAb. the I-142 mAb into NOD mice that had been injected with cy- Injection of I-142 inhibited Ag-specific proliferation by OT-II clophosphamide. Cyclophosphamide induces T1D in ∼70% of cells (Fig. 3A). We conducted similar experiments using CD8+ NOD mice in our colony. Of interest, all mice treated with I-142 T cells from OT-I TCR transgenic mice that recognize OVA remained free of disease (Fig. 4A). Histopathological analysis peptide presented by H-2Db. C57BL/6 mice were adoptively showed that most of the islets in control diabetic mice exhibited transferred with OT-I T cells that were subsequently immunized intrainsulitis (Fig. 4B), whereas the majority of islets in treated

FIGURE 4. Anti-CD98hc prevents development of T1D. (A) NOD mice (n = 6 per group) were injected with cyclophosphamide twice at 14-d intervals. The final injection of cyclophosphamide is day 0. I-142 (100 mg per mouse) was injected on days 0, 3, 6, 9, 12, and 15. The de- velopment of T1D was evaluated by measuring urine and blood glucose levels. (B) The histological section at day 10 was stained with H&E. Scale bars, 100 mm. As the negative control, the pancreata from cyclo- phosphamide-untreated age-matched NOD mice were used (Non-CPM). (C) Histological data at day 10 were scored. (D) The number of islets in four fields at day 10 was calculated. Data are shown as the mean 6 S.D. The data shown in this figure are representative of at least three experi- ments. *p , 0.05. The Journal of Immunology 2231

FIGURE 5. The I-142 mAb does not induce regu- latory T cells. (A) NOD mice were injected with cy- clophosphamide twice at 14-d intervals (n = 5 per group). The final injection of cyclophosphamide is day 0. I-142 (100 mg per mouse) was injected on days 0, 2, 4, and 6. CD4+CD25+Foxp3+ cells in the spleen, pancreas, or pancreatic lymph nodes in control IgG- or I-142–treated mice were examined by flow cytometry on day 10. Data are shown as the mean 6 S.D. (B) NOD mice were injected with cyclophosphamide twice at 14-d intervals. The final injection of cyclophospha- mide is day 0. I-142 (100 mg per mouse) was injected on days 0, 2, 4, and 6. Total spleen cells from the NOD mice on day 10 were transferred into NOD mice (n =5 per group) that had been treated with cyclophospha- mide twice at 14-d intervals. Cells were transferred 6 d after cyclophosphamide injection of recipient mice.

The data shown in this figure are representative of at Downloaded from least three experiments.

mice were not inflamed (Fig. 4B) or had slight peri-insulitis (Fig. that the I-142 mAb is able to protect against T1D when admin-

4B, 4C). The total number of islets was larger in the I-142–treated istered during an early stage of the disease. http://www.jimmunol.org/ group than in the control IgG-treated group, although the number As increased numbers of regulatory T cells are able to inhibit the was less than that in cyclophosphamide-untreated mice (Fig. 4D). development of T1D in NOD mice (17), we tested the effect of I- These data suggest that the I-142 mAb protects the islets from 142 on the number of Foxp3+ CD4+ T cells in cyclophosphamide- damage, which helps maintain their number. Such results indicate treated NOD mice. The number of Foxp3+ T cells in the spleen, by guest on September 29, 2021

FIGURE 6. Anti-CD98hc resolves T1D. After female NOD mice developed T1D, mice (n = 10 per group) were treated with control IgG or I-142 (100 mg per mouse). Each Ab was injected on days 0, 3, 5, and 7 (the day on which mice developed T1D was designated day 0). The restoration to normoglycemia was evaluated by monitoring urine and blood glucose levels. (A) This panel shows the representative course of blood glucose for one mouse, each treated with control rat IgG or I-142. (B) Effective treatment is defined by a blood glucose level ,250 mg/dl just after the final Ab treatment and thereafter normoglycemia for .14 d. Data are shown as the mean 6 S.D. **p , 0.01. (C and D) Histological sections after treatment with control IgG or I-142 were stained with H&E 10 d after initial treatment. Scale bars, 100 mm. Insulitis scores were calculated as described in Materials and Methods.(E) Total islets were counted 10 d after the final treatment. Data are shown as the mean 6 S.D. **p , 0.01. (F) Total T cells from pancreatic lymph nodes were obtained from NOD mice treated with I-142 or control rat IgG 2 d after the final Ab treatment and were stimulated with pancreatic Ags for 3 d. T cell proliferation in pancreatic lymph nodes was measured by incorporation of [3H]-thymidine. As the negative control, lymph node cells from rat IgG-treated mice were cultured without any exogenous Ags (No Antigen). Data are shown as the mean 6 S.D. *p , 0.05. 2232 MANIPULATION OF CD98 INHIBITS DIABETES

nificantly greater number of total islets than did the control IgG- treated mice (Fig. 6E). Overall, treatment with I-142 led to a sig- nificant increase in the number of noninflamed islets. Because T1D in NOD mice is caused by aberrant T cell re- sponses, we sought to determine whether the restoration of normo- glycemia by I-142 treatment involved interfering with autoantigen- specific T cell proliferation and effector T cell differentiation. We assessed T cell proliferation and cytokine secretion in T cells from NOD mice treated with I-142 or control IgG. T cells proliferated against pancreatic Ags more vigorously in control IgG-treated NOD mice than in I-142 mAb-treated NOD mice (Fig. 6F). These data indicate that I-142 suppressed diabetogenic T cell FIGURE 7. Anti-CD98hc resolves T1D independent of B cells. Total proliferation. 3 7 T cells (1 10 ) from diabetic NOD mice (n = 4) were transferred into Anti-CD98hc inhibits the progression of T1D independent of NOD/SCID mice (n = 10 per group) that were treated with I-142 or control B cell suppression IgG (100 mg per mouse). Ab was injected 0, 3, 6, 9, 12, and 15 d after T cell transfer, and the development of diabetes was monitored by mea- Previous articles have indicated that B cells are involved in T1D in suring blood glucose. Five mice in each group were used in this experi- NOD mice (18). As CD98 deficiency in B cells perturbs the dif- ment. The experiments in this figure are representative of three ferentiation of Ab-producing cells (19), we investigated whether Downloaded from independent experiments. *p , 0.05. T cells are involved in the anti-CD98hc mAb-mediated suppres- sion of T1D in NOD mice. We injected NOD/SCID mice with pancreatic lymph nodes, and pancreas was comparable between total T cells from diabetic NOD mice and treated the mice with control Ig- and I-142–treated mice (Fig. 5A). To further examine I-142. NOD/SCID mice injected with T cells from diabetic NOD whether treatment with I-142 induces the differentiation of cells mice began to develop diabetes 2 wk after transfer of the cells. + with T cell suppressive activity, regardless of Foxp3 expression, Treatment with I-142 suppressed the development of diabetes http://www.jimmunol.org/ we transferred total spleen cells from both cyclophosphamide- and (Fig. 7). These data indicate that I-142 is able to suppress T1D I-142–treated mice into cyclophosphamide-treated NOD mice and through direct inhibition of T cells. monitored blood glucose levels. We did not observe any inhibitory CD98 siRNA inhibit the progression of T1D effect of the spleen cells from I-142–treated mice compared with control IgG-treated mice (Fig. 5B). These data suggest that I-142 To determine if inhibition of CD98 function is involved in sup- is directly suppressive for effector T cells in vivo and does not pression of T1D, we treated NOD/SCID mice adoptively trans- induce the differentiation of cells with suppressive activity. ferred with T cells from diabetic NOD mice with siRNA against CD98hc. The siRNA against CD98hc or control siRNA was in- Anti-CD98hc inhibits the progression of T1D jected together with atelocollagen into NOD/SCID mice that by guest on September 29, 2021 We evaluated if the I-142 mAb is able to reverse T1D during the had been adoptively transferred with diabetogenic T cells. CD98hc advanced hyperglycemic stage. Accordingly, blood glucose levels expression on PBMCs was evaluated. The expression of CD98hc were monitored beginning at 15 wk, and mice displaying hyper- was efficiently downregulated by treatment with siRNA against glycemia between the ages of 20 and 30 wk were injected with CD98hc, but not control siRNA (Fig. 8A). Intraperitoneal injection I-142 or control IgG every 3 d. We found that .80% of the I-142– of siRNA against CD98hc prevented the development of T1D treated mice were protected from diabetes throughout the third (Fig. 8B). Even in mice that developed T1D, the development was treatment (Fig. 6A, 6B). To examine whether the restoration of delayed by treatment with CD98hcsiRNA. These data indicate that normoglycemia by I-142 suppressed cell infiltration around islets, downregulation of CD98hc in vivo is able to suppress T1D de- we performed histopathological analyses 10 d after the initial in- velopment in NOD mice. jection of I-142. Most of the islets in hyperglycemic and diabetic mice exhibited intrainsulitis (Fig. 6C, 6D), but the majority of Discussion islets in treated mice were not inflamed or had only mild peri- We identified a mAb that recognizes CD98hc by probing a panel of insulitis (Fig. 6C, 6D). Moreover, the treated animals had a sig- mAbs that modulate T cell proliferation. The I-142 mAb exhibits

FIGURE 8. Treatment with CD98hc siRNA sup- presses T1D development. Total T cells from diabetic NOD mice were injected into NOD/SCID mice (n =10 per group). The mice were treated with CD98hc siRNA (4 nmol per mouse) or control siRNA on days 2, 5, and 8 (the day on which T cells were transferred is designated day 0). (A) Expression of CD98hc on peripheral mononuclear cells with forward scat- ter/side light scatter gate was evaluated by staining cells with I-142, followed by anti-rat IgG for flow cytometry on day 2. Shadow, secondary Ab alone; dotted line, control siRNA; bold line, CD98hc siRNA. (B) The development of T1D was evaluated by mea- suring blood glucose levels. Ten mice in each group were used in the experiments. The experiments in this figure are representative of three independent experi- ments. *p , 0.05. The Journal of Immunology 2233 a strong inhibitory effect for both CD4+ and CD8+ T cell prolif- and that these cells prevent damage to b cells. In any case, it eration and functional differentiation. Injection of I-142 in cyclo- would be important to establish NOD mice lacking the CD98hc phosphamide-induced NOD mice completely prevented the onset gene in T cells, or other types of cells, to directly analyze the of T1D. Treatment of diabetic NOD mice with I-142 restored the major target cells for the I-142 mAb in terms of inhibiting T1D diabetic state to a normal level by suppressing T cell responses. development. Furthermore, treatment with siRNA against CD98hc inhibited the Currently, considerable interest has arisen in the potential of development of T1D. These data indicate that blocking CD98hc immunotherapy for treating T1D. The pathways that break down with an Ab or siRNA may have clinical applications for the immunological tolerance against b cells are unclear but include treatment of T1D and other T cell-mediated autoimmune diseases a significant contribution by T cells, through either CTL killing or in humans. cytokine release. Several attempts to treat T1D using Abs have CD98hc has a variety of functions, including amino acid been reported (18, 24, 25). Anti-CD3 mAb has shown therapeutic transport and participation in integrin signaling (12, 20, 21). potential in NOD mice by modulating regulatory T cells (26), and Previous articles have indicated that CD98 induces aggregation in the efficacy of anti-CD3 mAb to treat human T1D is currently in human T cells, which is blocked by soluble b1 integrin Abs (22), clinical trials (27). In contrast, treatment with I-142 did not affect and an anti-CD98 mAb suppresses Con A-mediated murine T cell the number of CD4+Foxp3+ regulatory T cells, and did not induce proliferation (15). Recently, Cantor et al. (16) reported that dele- suppressive activity to inhibit T1D in NOD mice, which suggests tion of the mouse CD98hc gene disturbs T cell function, sug- that anti-CD3 and anti-CD98 treatments use different mechanisms gesting that CD98hc controls Ag-specific T cell proliferation. To to suppress T1D in NOD mice. Therefore, it is important to de- ascertain whether the effect of I-142 administration on the reso- termine the best schedules for treatment with each Ab in relation Downloaded from lution of T1D in NOD mice is attributed to the inhibition of to the disease stages of T1D. In addition, because CD98hc is ex- CD98hc function, we downregulated CD98hc expression by in- pressed on various types of cells, anti-CD98hc mAb treatment might jecting CD98hc siRNA. Treatment with CD98hc siRNA sup- induce unappreciated adverse effects. Nevertheless, we have not pressed T1D development induced by diabetogenic T cells observed any macroscopic or microscopic alterations in mice after transferred into NOD/SCID mice. This treatment reduced the .10 injections of anti-CD98hc mAb. Furthermore, in future it

surface expression of CD98hc on PBMCs. Although we did not would be important to examine if I-142 tolerizes T cells or how http://www.jimmunol.org/ address which cells are the actual target cells for this treatment long the effect of I-142 lasts after initial treatment. In any case, it in vivo, this approach would also be useful in treating T1D by will be necessary to carefully apply anti-CD98hc mAb to develop inhibiting CD98hc function. Taken together, these data indicate therapeutic strategies for T1D. that inhibition of CD98hc function in vivo is able to suppress T1D In conclusion, these data indicate that anti-CD98hc mAb has a development in NOD mice, which suggests that the effect of the potent ability to resolve T1D in NOD mice. The I-142 mAb is able I-142 mAb on development of T1D is attributable to the inhibition to suppress both CD4+ and CD8+ T cell proliferation and func- of CD98hc function. Indeed, our anti-CD98hc mAb was able to tional differentiation. The inhibition of CD98hc by siRNA in vivo suppress Ag-specific proliferation and effector functions of both also exhibited a therapeutic effect on T1D in NOD mice. These CD4+ and CD8+ T cells in vitro and in vivo. A previous study data show that CD98hc is crucial for the progression of T1D in by guest on September 29, 2021 using CD98hc-deficient T cells showed reduced T cell prolifera- NOD mice and suggest that inhibition of CD98hc function could tion, but no defect in acquiring effector functions (16). This article be a useful strategy for the treatment of T cell-mediated autoim- indicated that defective T cell proliferation in CD98hc-deficient mune diseases, including T1D. T cells is attributed to lack of the integrin-binding domain of CD98hc (16). Therefore, the suppressive effect of I-142 on T cell Acknowledgments effector functions in our studies might suggest that the I-142 mAb We thank Chiemi Kinouchi and Chiyo Tomari for technical and editorial influences other functions of CD98hc besides integrin activity. For assistance. instance, we cannot rule out the possibility that I-142 recognizes complexes of CD98hc with other unidentified CD98hc-associated Disclosures molecules because I-142 does not recognize CD98hc in Western The authors have no financial conflicts of interest. blots (data not shown). It is possible that I-142 also affects the function of such associated molecules. Therefore, it would be important to test whether I-142 affects T cell functions, using References CD98hc-deficient T cells. In any case, these data suggest that 1. Castellino, F., and R. N. Germain. 2006. Cooperation between CD4+ and CD8+ I-142 administration or inhibition of CD98hc with siRNA would T cells: when, where, and how. Annu. Rev. Immunol. 24: 519–540. 2. Lehuen, A., J. Diana, P. Zaccone, and A. Cooke. 2010. Immune cell crosstalk in be a good strategy for suppressing the development or progression type 1 diabetes. Nat. Rev. Immunol. 10: 501–513. of T1D, which would at least partly be attributed to the inhibition 3. Carreno, B. M., and M. Collins. 2002. The B7 family of ligands and its recep- of T cell proliferation. tors: new pathways for costimulation and inhibition of immune responses. Annu. Rev. Immunol. 20: 29–53. Multiple factors might contribute to the therapeutic effect of the 4. Keir, M. E., M. J. Butte, G. J. Freeman, and A. H. Sharpe. 2008. PD-1 and its I-142 mAb on T1D because CD98hc is expressed in various tissues ligands in tolerance and immunity. Annu. Rev. Immunol. 26: 677–704. 5. Watts, T. H. 2005. TNF/TNFR family members in costimulation of T cell in addition to the immune system (14, 23). Even in immune cells, responses. Annu. Rev. Immunol. 23: 23–68. CD98hc is expressed on B cells, dendritic cells, and macrophages 6. Sugamura, K., N. Ishii, and A. D. Weinberg. 2004. Therapeutic targeting of the (data not shown). Nevertheless, as I-142 suppressed the prolifer- effector T-cell co-stimulatory molecule OX40. Nat. Rev. Immunol. 4: 420–431. + + 7. Stadinski, B., J. Kappler, and G. S. Eisenbarth. 2010. Molecular targeting of islet ation of both CD4 and CD8 T cells in vitro and in vivo, the autoantigens. Immunity 32: 446–456. therapeutic effect of I-142 on T1D, at least partly, would likely 8. Shoda, L. K., D. L. Young, S. Ramanujan, C. C. Whiting, M. A. Atkinson, depend on suppressing T cell proliferation. Indeed, we have evi- J. A. Bluestone, G. S. Eisenbarth, D. Mathis, A. A. Rossini, S. E. Campbell, et al. 2005. A comprehensive review of interventions in the NOD mouse and impli- dence that deletion of the CD98hc gene in T cells suppresses cations for translation. Immunity 23: 115–126. T cell-mediated pathological responses in vivo (data not shown). 9. Tsai, S., A. Shameli, and P. Santamaria. 2008. CD8+ T cells in type 1 diabetes. Adv. Immunol. 100: 79–124. This finding leads to the intriguing possibility that the I-142 mAb 10. van Belle, T. L., K. T. Coppieters, and M. G. von Herrath. 2011. Type 1 diabetes: interacts with cells located in the pancreas, including islet b cells, etiology, immunology, and therapeutic strategies. Physiol. Rev. 91: 79–118. 2234 MANIPULATION OF CD98 INHIBITS DIABETES

11. Bach, J. F., and D. Mathis. 1997. The NOD mouse. Res. Immunol. 148: 285–286. 20. Petty, H. R., R. G. Worth, and R. F. Todd, III. 2002. Interactions of integrins with 12. Cantor, J. M., M. H. Ginsberg, and D. M. Rose. 2008. Integrin-associated pro- their partner in leukocyte membranes. Immunol. Res. 25: 75–95. teins as potential therapeutic targets. Immunol. Rev. 223: 236–251. 21. Tsurudome, M., and Y. Ito. 2000. Function of fusion regulatory proteins (FRPs) 13. Verrey,F.,E.I.Closs,C.A.Wagner,M.Palacin,H.Endou,andY.Kanai.2004.CATs in immune cells and virus-infected cells. Crit. Rev. Immunol. 20: 167–196. and HATs: the SLC7 family of amino acid transporters. Pflugers Arch. 447: 532–542. 22. Miyamoto, Y. J., J. S. Mitchell, and B. W. McIntyre. 2003. Physical association 14. Yan, Y.,S. Vasudevan, H. T. Nguyen, and D. Merlin. 2008. Intestinal epithelial CD98: and functional interaction between beta1 integrin and CD98 on human an oligomeric and multifunctional protein. Biochim. Biophys. Acta 1780: 1087–1092. T lymphocytes. Mol. Immunol. 39: 739–751. 15. Komada, H., A. Imai, E. Hattori, M. Ito, H. Tsumura, T. Onoda, M. Kuramochi, 23. Nguyen, H. T., G. Dalmasso, L. Torkvist, J. Halfvarson, Y. Yan, H. Laroui, M. Tani, K. Yamamoto, M. Yamane, et al. 2006. Possible activation of murine D. Shmerling, T. Tallone, M. D’Amato, S. V. Sitaraman, and D. Merlin. 2011. T lymphocyte through CD98 is independent of interleukin 2/interleukin 2 re- CD98 expression modulates intestinal homeostasis, inflammation, and colitis- ceptor system. Biomed. Res. 27: 61–67. associated cancer in mice. J. Clin. Invest. 121: 1733–1747. 16. Cantor, J., M. Slepak, N. Ege, J. T. Chang, and M. H. Ginsberg. 2011. Loss of 24. Bach, J. F., and L. Chatenoud. 2001. Tolerance to islet autoantigens in type 1 T cell CD98 H chain specifically ablates T cell clonal expansion and protects diabetes. Annu. Rev. Immunol. 19: 131–161. from autoimmunity. J. Immunol. 187: 851–860. 25. Chatenoud, L., and J. A. Bluestone. 2007. CD3-specific antibodies: a portal to 17. Tarbell, K. V., L. Petit, X. Zuo, P. Toy, X. Luo, A. Mqadmi, H. Yang, the treatment of autoimmunity. Nat. Rev. Immunol. 7: 622–632. M. Suthanthiran, S. Mojsov, and R. M. Steinman. 2007. Dendritic cell-expanded, 26. Dromey, J. A., B. H. Lee, H. Yu, H. E. Young, D. J. Thearle, K. P. Jensen, islet-specific CD4+ CD25+ CD62L+ regulatory T cells restore normoglycemia S. I. Mannering, and L. C. Harrison. 2011. Generation and expansion of regu- in diabetic NOD mice. J. Exp. Med. 204: 191–201. latory human CD4(+) T-cell clones specific for pancreatic islet autoantigens. 18. Hu, C. Y., D. Rodriguez-Pinto, W. Du, A. Ahuja, O. Henegariu, F. S. Wong, J. Autoimmun. 36: 47–55. M. J. Shlomchik, and L. Wen. 2007. Treatment with CD20-specific antibody pre- 27. Herold, K. C., S. Gitelman, C. Greenbaum, J. Puck, W. Hagopian, P. Gottlieb, vents and reverses autoimmune diabetes in mice. J. Clin. Invest. 117: 3857–3867. P. Sayre, P. Bianchine, E. Wong, V. Seyfert-Margolis, et al; Immune Tolerance 19. Cantor, J., C. D. Browne, R. Ruppert, C. C. Fe´ral, R. Fa¨ssler, R. C. Rickert, and Network ITN007AI Study Group. 2009. Treatment of patients with new onset M. H. Ginsberg. 2009. CD98hc facilitates B cell proliferation and adaptive Type 1 diabetes with a single course of anti-CD3 mAb Teplizumab preserves humoral immunity. Nat. Immunol. 10: 412–419. insulin production for up to 5 years. Clin. Immunol. 132: 166–173. Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021