Lipopolysaccharide-Activated B Cells Down-Regulate Th1 Immunity and Prevent Autoimmune Diabetes in Nonobese Diabetic Mice

Transcription

1 This information is current as of July 20, References Subscription Permissions Alerts Lipopolysaccharide-Activated B Cells Down-Regulate Th1 Immunity and Prevent Autoimmune Diabetes in Nonobese Diabetic Mice Jide Tian, Dan Zekzer, Lorraine Hanssen, Yuxin Lu, Angelica Olcott and Daniel L. Kaufman J Immunol 2001; 167: ; ; doi: /jimmunol This article cites 40 articles, 23 of which you can access for free at: Why The JI? Submit online. Rapid Reviews! 30 days* from submission to initial decision No Triage! Every submission reviewed by practicing scientists Fast Publication! 4 weeks from acceptance to publication *average Information about subscribing to The Journal of Immunology is online at: Submit copyright permission requests at: Receive free -alerts when new articles cite this article. Sign up at: Downloaded from by guest on July 20, 2018 The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD Copyright 2001 by The American Association of Immunologists All rights reserved. Print ISSN: Online ISSN:

2 Lipopolysaccharide-Activated B Cells Down-Regulate Th1 Immunity and Prevent Autoimmune Diabetes in Nonobese Diabetic Mice 1 Jide Tian, 2 Dan Zekzer, Lorraine Hanssen, Yuxin Lu, Angelica Olcott, and Daniel L. Kaufman B cells can serve dual roles in modulating T cell immunity through their potent capacity to present Ag and induce regulatory tolerance. Although B cells are necessary components for the initiation of spontaneous T cell autoimmunity to cell Ags in nonobese diabetic (NOD) mice, the role of activated B cells in the autoimmune process is poorly understood. In this study, we show that LPS-activated B cells, but not control B cells, express Fas ligand and secrete TGF-. Coincubation of diabetogenic T cells with activated B cells in vitro leads to the apoptosis of both T and B lymphocytes. Transfusion of activated B cells, but not control B cells, into prediabetic NOD mice inhibited spontaneous Th1 autoimmunity, but did not promote Th2 responses to cell autoantigens. Furthermore, this treatment induced mononuclear cell apoptosis predominantly in the spleen and temporarily impaired the activity of APCs. Cotransfer of activated B cells with diabetogenic splenic T cells prevented the adoptive transfer of type I diabetes mellitus (T1DM) to NOD/scid mice. Importantly, whereas 90% of NOD mice treated with control B cells developed T1DM within 27 wk, <20% of the NOD mice treated with activated B cells became hyperglycemic up to 1 year of age. Our data suggest that activated B cells can down-regulate pathogenic Th1 immunity through triggering the apoptosis of Th1 cells and/or inhibition of APC activity by the secretion of TGF-. These findings provide new insights into T-B cell interactions and may aid in the design of new therapies for human T1DM. The Journal of Immunology, 2001, 167: B cells are not only central to the production and amplification of humoral immune responses, they also play an important role as APCs in the generation of T cell-mediated immune responses (1, 2). Activated B cells express high levels of MHC class II and T cell activation-associated molecules, allowing them to effectively present Ag to T cells (3, 4). Although dendritic cells and macrophages promote Ag-specific Th1 cell differentiation, B cell presentation of Ag usually induces T cell anergy and tends to promote naive T cell differentiation toward an anti-inflammatory Th2 phenotype (5 8). Thus, B cells have dual roles in modulating T cell immunity, which may contribute to the maintenance of immunological homeostasis. The nonobese diabetic (NOD) 3 mouse spontaneously develops type I diabetes mellitus (T1DM), which is thought to be mediated by both CD4 and CD8 autoreactive T lymphocytes (9 13). Although proinflammatory Th1 cells are considered to be the functional mediators of cell destruction in the pancreatic islets, the role of B cells in the disease process of T1DM is still unclear. B cells appear to be a necessary component for the development of cell-specific autoimmunity, as NOD Ig null mice fail to develop Department of Molecular and Medical Pharmacology, University of California, Los Angeles, CA Received for publication January 30, Accepted for publication May 9, The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by grants from the National Institute of Health and Juvenile Diabetes Foundation International. 2 Address correspondence and reprint requests to Dr. Jide Tian, Department of Molecular and Medical Pharmacology, School of Medicine, University of California, Le Conte Avenue, Los Angeles, CA address: jtian@mednet.ucla.edu 3 Abbreviations used in this paper: NOD, nonobese diabetic mouse; T1DM, type 1 diabetes mellitus; AICD, activation-induced cell death; GAD, glutamic acid decarboxylase; HSP, heat shock protein; HEL, hen egg lysozyme; FasL, Fas ligand; 7-AAD, 7-amino-actinomycin D. autoimmune diabetes (14, 15). Recent studies have demonstrated that B cells are crucial APCs in the pathogenic T cell response to glutamic acid decarboxylase (GAD) and for overcoming a checkpoint in T cell tolerance to cell Ags (16, 17). Furthermore, B cells activated by engaging CD40 and IgM can rescue activated T cells from activation-induced cell death (AICD) (18). Therefore, administering activated B cells to prediabetic NOD mice may promote cell-reactive T cell responses and exacerbate the disease process. Alternatively, recent studies have shown that LPS-activated B cells (activated B cells) express high levels of membraneassociated Fas ligand (FasL) and trigger apoptosis of Fas target cells in vitro (19). This suggests that transfusion of activated B cells into prediabetic NOD mice may inhibit disease progression by triggering apoptosis of diabetogenic T cells through the Fas- FasL interaction. To further delineate the role of activated B cells in modulating T cell effector functions, we investigated the impact of treatment with activated B cells on diabetes incidence and determined the mechanism(s) underlying the action of B cell-based immunotherapy in prediabetic NOD mice. We show that in addition to the expression of membrane-associated FasL, activated B cells secrete high levels of TGF-, an anti-inflammatory cytokine associated with down-regulating pathogenic responses in autoimmune diseases. Interaction of diabetogenic T cells with activated B cells, but not control B cells, leads to the apoptosis of both T and B lymphocytes in vitro. Transfusion of activated B cells, but not control B cells, into prediabetic NOD mice inhibits the spontaneous Th1 immunity to cell Ags and disease progression in prediabetic NOD mice. Coadoptive transfer of activated B cells with diabetogenic splenic T cells isolated from newly diabetic NOD mice prevents the adoptive transfer of T1DM to NOD/scid mice. The therapeutic effects of treatment with activated B cells are likely to be mediated by triggering diabetogenic T cells to apoptose and/or temporarily impairing APC function. Copyright 2001 by The American Association of Immunologists /01/$02.00

3 1082 LPS-ACTIVATED B CELLS INHIBIT AUTOIMMUNITY IN NOD MICE Materials and Methods Mice NOD H-2 NOD (Taconic Farms, Germantown, NY) were bred under specific pathogen-free conditions. Female NOD mice ( 85%) spontaneously develop autoimmune diabetes by 35 wk of age in our NOD colony. Only female mice were used in these studies. Antigens Mouse GAD65 and control -galactosidase were prepared as previously described (20). Heat shock protein peptide (HSP277) and hen egg lysozyme peptide (HEL11-25) were synthesized at 95% purity by Multiple Peptide Systems (San Diego, CA). The amino acid sequences of HSP277 and HEL11-25 have been reported elsewhere (21, 22). Insulin B chain, HEL, chicken OVA, and LPS from Escherichia coli 0111:B4 were purchased from Sigma (St. Louis, MO). Isolation and stimulation of B cells B cells were purified by negative selection. Briefly, splenic mononuclear cells were isolated from 2- to 3-wk-old or adult female NOD mice and incubated in petri dishes at 37 C for 1 h. The nonadherent lymphocytes were harvested and then T cells were removed by two successive treatments with complement (Pel-Freez Biologicals, Brown Deer, WI), and anti-cd4 and anti-cd8 Abs (BD PharMingen, San Diego, CA). After washing twice with medium, the remaining B cell-enriched fraction was loaded onto a B cell purification column (Biotex Laboratories, Edmonton, Canada). The eluted B cells were stained with fluorescently labeled Abs (anti-cd3, anti-cd4, anti-cd8, anti-cd19, anti-br220, anti-mac-1, and anti-cd11c) and analyzed by FACS scanning. The isolated B cells, whose purity was 95%, were directly injected into mice or activated by stimulation with 20 g/ml LPS for 48 h in X-vivo 20 medium (BioWhittaker, Walkersville, MD). FACS analysis of FasL expression The isolated B cells from individual NOD mice were incubated in the presence or absence of different concentrations of LPS in X-vivo 20 medium for different time periods (24 96 h). The activated B cells and control B cells (without LPS stimulation), as well as freshly isolated B cells, were costained with PE-anti-mouse FasL (MFL3) and FITC-anti-Br220 or FITC-anti-CD19 (BD PharMingen). The expression of membrane-associated FasL on B cells was determined by FACS analysis. ELISA for TGF- detection The isolated B cells from individual NOD mice were stimulated with (or without) 20 g/ml LPS in X-vivo 20 medium for 48 h and the supernatants were harvested. The amount of TGF- in the supernatants was determined by ELISA using recombinant human TGF- 1 as the standard (23). Induction of lymphocyte apoptosis in vitro Isolated B cells from 2- to 3-wk-old NOD mice were stimulated with 20 g/ml LPS in X-vivo 20 medium for 48 h. B cells that were incubated without LPS, as well as freshly isolated B cells, were used as control B cells. The activated B cells, or control B cells, were incubated with freshly isolated splenic T cells from individual newly diabetic NOD mice or ZK35 T cells (a Th1 T cell clone recognizing GAD ), at different ratios of T:B cells (1:1 1:10) in 96-well round-bottom plates (10 wells/ group) for 16 h in a 1:1 mixture of HL-1 and X-vivo 20 medium. The incubated T cells alone, or B cells alone, were used as controls to determine spontaneous T or B cell apoptosis. After incubation, the cells were harvested, stained with 7-amino-actinomycin D, PE-annexin V, FITC-anti- CD3, or FITC-anti-B220 (BD PharMingen), and analyzed by FACS scanning. The T and B lymphocyte apoptosis induced by T:B cell interactions was expressed as the mean number, or percentage, of apoptotic lymphocytes, respectively. T cell proliferation assay and ELISPOT analysis Female NOD mice were treated with 10 7 activated B cells or freshly isolated B cells (control) at 4 and 10 wk of age. Two weeks later, their splenic mononuclear cells as well as the splenic mononuclear cells from agematched unmanipulated NOD mice were characterized for spontaneous T cell immunity to a panel of cell Ags. To test T cell proliferation, mononuclear cells ( cell/well) were incubated with optimal concentrations of cell autoantigens (at 20 g/ml for whole protein and 7 M for peptides) in triplicate in FCS-free HL-1 medium in 96-well microtiter plates at 37 C with 5% CO 2 for 96 h. Medium alone (without any Ag) or anti-cd3 (1 g/ml) were used as the negative and positive controls, respectively, for each mouse. During the last h of the 96-h culture period, 1 Ci [ 3 H]thymidine was added into each well. Incorporation of label was measured by liquid scintillation counting. T cell proliferation against specific Ag was expressed as the stimulation index. The frequency of Ag-specific T cells secreting IFN-, IL-4, and IL-5 was determined by using a modified ELISPOT assay as previously described (24). Briefly, 10 6 splenic mononuclear cells per well were added (in duplicate) to an ELISPOT plate that had been coated with cytokine capture Abs and incubated with peptide (20 M) or whole protein (100 g/ml) for 24-h (for IFN- ) or 40-h (for IL-4 and IL-5) detection. After washing, biotinylated detection Abs were added and the plates were incubated at 4 C overnight. Bound secondary Abs were visualized using HRP-streptavidin (Dako, Carpinteria, CA) and 3-amino-9-ethylcarbazole. The Abs R4-6A2/ XMG 1.2-biotin, 11B11/BVD6-24G2-biotin and TRFK5/TRFK4-biotin (BD PharMingen) were used for capture and detection of IFN-, IL-4, and IL-5, respectively. TUNEL assay At 12 wk of age, groups of mice that were treated twice at 4 and 10 wk of age with 10 7 activated B cells or control B cells, as well as unmanipulated NOD mice, were sacrificed and their pancreata and spleens were paraffin embedded and sliced into 5- m sections. The number of apoptotic mononuclear cells was determined by TUNEL assay using an in situ cell death detection kit following the instructions of the supplier (Boehringer Mannheim, Indianapolis, IN). The data were presented as the average number of apoptotic cells per islet or per view in the marginal zone area of the spleen from 25 slides of 5 mice for each group. Ag presentation assay Female NOD mice were treated with 10 7 activated or control B cells at 4 and 10 wk of age. Two or 4 wk later, their splenic mononuclear cells were isolated. To prepare APCs, we depleted CD3 T cells by incubating splenic mononuclear cells with anti-cd3 Ab plus complement and loading the remaining cells on a CD3-enriching column (R&D Systems, Minneapolis, MN). The unbound CD3 mononuclear cells were used as APCs. CD3 mononuclear cells from age-matched and unmanipulated NOD mice were used as control APCs. To prepare responder T cells, 6-wk-old female NOD mice were immunized with 100 g OVA or HEL in 50% CFA in their footpads. Nine days later, CD3 lymph node T cells were isolated from the draining lymph node using an Ab mixture against Br220, CD11c, and Mac1 plus complement. The remaining cells were loaded on a T cell purification column and the unbound CD3 T cells were used as responders. Both the purified CD3 lymph node T cells and CD3 splenic mononuclear cells were analyzed by FACS analysis, and only isolated cells with purity 95% were used in the Ag-presenting assay. Responder lymph node T cells (10 5 ), along with CD3 splenic mononuclear cells, were stimulated with different concentrations of the injected Ag (in triplicate) in 96-well plates for 96 h. During the last h of the 96-h culture period, 1 Ci [ 3 H]thymidine was added to each well. Incorporation of label was measured by liquid scintillation counting. Adoptive transfer of diabetes Adoptive transfer of diabetes was performed as previously described (25). Briefly, 10 7 splenic CD3 T cells or splenic mononuclear cells isolated from unmanipulated and newly diabetic NOD mice were mixed with an equal number of activated B cells, or control B cells (from 2- to 3-wk-old prodiabetic or 6- to 8-wk-old prediabetic NOD mice), and injected i.v. into 6-wk-old female NOD/scid mice. Control groups of age-matched NOD/ scid mice received 10 7 isolated splenic T cells or mononuclear cells from newly diabetic NOD mice. To test for a transferable regulatory response, we transfused NOD/scid mice with 10 7 diabetogenic splenic cells along with an equal number of splenic mononuclear cells from 15-wk-old NOD mice that had been treated with activated B cells or control B cells. Control groups received diabetogenic splenic cells alone, splenic mononuclear cells from NOD mice treated with control or activated B cells. Following the adoptive transfer of splenic cells, urine glucose levels were monitored twice weekly for diabetes onset by Test-tape (Eli Lilly, Indianapolis, IN). After we detected abnormal glucose levels in the urine, the blood glucose levels were measured every other day. Two consecutive blood glucose levels 13 mmol/l was considered as T1DM onset. Spontaneous T1DM incidence Female NOD mice were treated at 4 wk of age with 10 7 activated B cells or control B cells. Following the first treatment, NOD mice received the same dosage of activated or control B cells every 6 wk until they reached

4 The Journal of Immunology wk of age. Another group of unmanipulated NOD mice served as a control. All mice were monitored for the development of diabetes as described above, up to 52 wk of age. The effect of treatments on T1DM incidence was statistically analyzed by life table analysis. Results Transfusion of activated B cells inhibits spontaneous autoimmune diabetes Because activated B cells express membrane-associated FasL and can trigger Fas target cells to apoptose, we first examined whether administering activated B cells to prediabetic NOD mice could inhibit disease progression. Female NOD mice were treated at 4 wk of age with 10 7 activated B cells or control B cells. This treatment was repeated every 6 wk until the mice were 34 wk of age. Another group of unmanipulated NOD mice served as a control. Approximately 90% of the mice that received control B cell treatments developed diabetes by 27 wk of age, paralleling the disease course observed in unmanipulated control NOD mice (Fig. 1). In contrast, 80% of mice that received activated B cells remained free of hyperglycemia at 52 wk of age ( p 0.012). The few mice that did become diabetic following treatment with activated B cells had a greatly delayed disease onset. Histological examination of the pancreata from 12-wk-old mice treated with activated B cells revealed that most of the islets had infiltrating lymphocytes. The severity of insulitis in the experimental group of mice was slightly reduced, but not significantly, when compared with that of mice treated with control B cells or unmanipulated control mice (data not shown). Thus, treatment with LPS-activated B cells greatly inhibited the progression of spontaneous autoimmune diabetes in prediabetic NOD mice, but did not significantly alter the extent of their insulitis. Activated, but not control, B cells express membrane-associated FasL and secrete high levels of TGF- Fas-FasL interactions usually lead to the AICD of Ag-primed T cells (26), which may contribute to the protective effects of activated B cell treatment in prediabetic NOD mice. Because LPSactivated B cells, which are isolated from normal strains of mice, express high levels of membrane-associated FasL (19), we examined whether B cells from these autoimmune-prone mice also express membrane-associated FasL following stimulation by LPS. B cells were isolated from 2- to 3-wk-old female NOD mice and incubated in the presence or absence of different concentrations of LPS in vitro for different time periods. After incubation, the B cells were harvested and stained with PE-anti-FasL, and FITC-anti- Br220 or anti-cd19, and then analyzed by FACS scanning. In the absence of LPS stimulation, the incubated B cells did not express detectable levels of FasL on their surface, nor did freshly isolated B cells from age-matched NOD mice (Fig. 2). In contrast, 65% of B cells that were stimulated with 20 g/ml LPS for 48 h were positive for both anti-fasl and anti-br220 staining. A similar pattern of staining was observed using both anti-fasl and anti-cd19 staining (data not shown). These data demonstrate that LPS stimulates B cells isolated from NOD mice to express membrane-associated FasL. Since activated APCs secrete cytokines which can modulate T cell autoimmunity, we also investigated whether activated B cells secrete anti-inflammatory cytokines. Isolated B cells were incubated in the presence or absence of 20 g/ml LPS for 48 h and the level of IL-10 and TGF- in their supernatants was determined by ELISA. We found that the concentration of IL-10 in the supernatants of LPS-stimulated B cells was slightly higher than that of B cells cultured without LPS, but not significantly (data not shown). Although a very low level of TGF- was found in the supernatants of incubated B cells in the absence of LPS stimulation, a high concentration of TGF- was detected in the supernatants of B cells stimulated with LPS (Fig. 3). The high level of FasL expression and TGF- secretion of activated B cells suggest that the therapeutic effects of treatment with activated B cells could be mediated by induction of T cell apoptosis through T-B cell interactions and/or by secretion of anti-inflammatory TGF-. Interaction of diabetogenic T cells with activated B cells leads to the apoptosis of both T and B lymphocytes in vitro To determine whether the T-B cell interaction triggers lymphocyte apoptosis, we incubated diabetogenic splenic T cells isolated from FIGURE 1. Transfusion of activated B cells inhibits diabetes progression in prediabetic NOD mice. Female NOD mice that were treated with activated B cells or control B cells, as well as unmanipulated control mice, were followed up to 1 year of age to determine the effect of treatment on long-term disease incidence. Two consecutive blood glucose levels of 13 mmol/l were considered as disease onset (n 16 for each group). FIGURE 2. Activated B cells express membrane-associated FasL. Isolated B cells from 2-wk-old NOD mice were incubated in the presence or absence of 20 g/ml LPS in X-vivo 20 medium for 48 h. The cells were then costained with PE-anti-mouse FasL, and FITC-anti-Br220 or FITCanti-CD19. The expression of membrane-associated FasL on B cells was determined by FACS analysis. The data represent one of three independent experiments with similar outcomes. Control B cells incubated in the absence of LPS showed undetectable FasL expression similar to that of freshly isolated B cells from age-matched unmanipulated NOD mice (data not shown). LPS-activated B cells, but not control B cells, isolated from a nondiabetic mouse strain (C57BL/6) mice showed a similar level of FasL expression (data not shown).

5 1084 LPS-ACTIVATED B CELLS INHIBIT AUTOIMMUNITY IN NOD MICE FIGURE 3. Activated B cells, but not control B cells, secrete TGF-. Isolated B cells from individual NOD mice were stimulated with 20 g/ml LPS in X-vivo 20 medium for 48 h, and the incubated B cells (without stimulation) were used as the control. The amount of TGF- in the supernatants was determined by an ELISA using recombinant human TGF- 1as the standard. Data are presented as the average concentrations of TGF- SEM (nine mice for each group). newly diabetic NOD mice, or ZK35 T cells ( /well), with different numbers of activated or control B cells (at T:B ratios of 1:1 1:10) for 16 h in vitro. The cells were then stained with 7-AAD, PE-annexin V, and FITC-anti-CD3 or FITC-anti-B220 to characterize the apoptotic lymphocytes by FACS scanning. Although few apoptotic T cells were detected among the diabetogenic T cells that had been incubated with control B cells, the number of apoptotic T cells was significantly increased (15- to 30-fold) after incubating T cells with activated B cells. Indeed, activated B cells induced the diabetogenic T cells to undergo apoptosis in a dose-dependent manner (Fig. 4A). Notably, the proportion of apoptotic activated B cells, but not control B cells, was also greatly increased following the interaction with diabetogenic T cells in vitro (Fig. 4B). A similar pattern of apoptotic T and B cells resulted after interactions between cloned autoantigen-specific T cells (ZK35) and activated B cells, but not control B cells (Fig. 4). Thus, our findings demonstrate that the interaction of diabetogenic T cells with activated B cells leads to the apoptosis of both T and B lymphocytes in vitro. Treatment with activated B cells, but not control B cells, inhibits spontaneous Th1 responses to cell Ags in prediabetic NOD mice Next, we characterized T cell autoimmunity in NOD mice that were treated with activated B cells or control B cells. Female NOD mice were treated twice with activated B cells or control B cells at 4 and 10 wk of age. At 12 wk of age, splenic mononuclear cells from B cell-treated or age-matched unmanipulated NOD mice were tested for T cell proliferative responses to a panel of cell autoantigens. We found that splenic mononuclear cells from NOD mice treated with control B cells developed strong proliferative responses to GAD65, HSP277, and insulin B chain similar in magnitude to those of unmanipulated NOD mice (Fig. 5A). This observation suggests that the transfusion of freshly isolated B cells into prediabetic NOD mice does not affect the spontaneous development of T cell autoimmunity. In contrast, treatment of prediabetic NOD mice with activated B cells almost abolished the spontaneous T cell autoreactivity to all of the tested autoantigens. These findings suggest that activated B cells inhibit the development, or cause the inactivation, of autoreactive T cells in NOD mice. FIGURE 4. Coincubation of diabetogenic T cells with activated B cells leads to apoptosis of lymphocytes. A, The number of apoptotic T cells was expressed as the mean apoptotic T cells (over background of spontaneously apoptotic T cells) SEM. The background for incubated diabetogenic T cells and ZK35 T cells (alone) was or per T cells, respectively. B, The extent of B cell apoptosis was expressed as the mean percentage of apoptotic B cells (over control of spontaneously apoptotic B cells) SEM. The average percentage of apoptotic B cells in controls was 1.4 or 0.5% for activated B cells and freshly isolated B cells, respectively. Data are from two independent assays. We further analyzed whether the ablation of the spontaneous proliferative T cell responses in NOD mice that had been treated with activated B cells resulted from the absence of cell Agprimed T cells or from the priming of regulatory Th2 responses. We performed ELISPOT assays to determine the frequency of IFN- -, IL-4-, and IL-5-secreting splenic T cells responding to cell autoantigens in B cell-treated and unmanipulated NOD mice. Two weeks after the final B cell treatment, we found that NOD mice receiving control B cells developed a high frequency of IFN- responses to the tested cell autoantigens, which were indistinguishable in frequency from that of age-matched unmanipulated NOD mice (Fig. 5B). In contrast, the frequency of IFN- secreting T cells responding to GAD65 in mice treated with activated B cells was reduced dramatically to only 18% of that in control NOD mice. Furthermore, no Th1 responses to HSP277 and insulin B chain were detected in mice treated with activated B cells. Importantly, there were no detectable IL-4- or IL-5-secreting T cells responding to cell autoantigens in any group of mice (data not shown). Consistent with this observation, analysis of Abs

6 The Journal of Immunology 1085 FIGURE 5. Treatment with activated B cells inhibits spontaneous Th1 responses to cell Ags in prediabetic NOD mice. A, The mean Ag-induced T cell proliferation over background was expressed as the stimulation index SEM. The background for medium alone ranged from 2240 to 2825 cpm. B, The frequency of cell Ag-specific IFN- -, IL-4-, and IL-5-secreting T cells in B cell-treated and unmanipulated NOD mice was determined by the ELISPOT assay. Data were presented as the mean number of spot-forming colonies (SFC) per 10 6 splenic cells SEM. Mice from control and experimental groups (n 6 for each group) were tested simultaneously in two independent experiments (using triplicate cultures). All of the mice had undetectable levels of IL-4- and IL-5-secreting T cell responses to all of the tested Ags (data not shown). None of the control Ags ( -galactosidase and HEL11-25) induced significant splenic T cell proliferation or IFN- responses (data not shown). NOD mice that were treated with activated B cells may stem from T cell apoptosis through T-B cell interactions. To test this possibility, we examined the spleens and islets of mice treated with activated or control B cells as well as unmanipulated age-matched NOD mice for the extent of apoptosis by TUNEL assay. Although a few islets in the pancreatic tissue sections showed a small number of apoptotic cells, the average number of apoptotic cells were indistinguishable among the tested groups of mice (Fig. 6). In contrast, the average number of apoptotic cells in splenic sections from mice treated with activated B cells were 12-fold higher than that in sections from mice treated with control B cells and unmanipulated control mice. Notably, those apoptotic cells were predominantly observed in the marginal zone area near to the periarterial lymphatic sheath, a site where splenic T-B cell interactions commonly take place. This observation suggests that the T-B cell interactions leading to the development of substantial mononuclear cell apoptosis might be located predominantly in the spleen. Because both activated T and B cells express high levels of Fas and they are highly sensitive to apoptotic induction, these data suggest that the interaction of activated T and B cells may promote the apoptosis of both T and B lymphocytes in the peripheral lymph tissues. Treatment with activated B cells, but not control B cells, temporarily impairs the function of APCs Besides inducing T cell apoptosis, activated B cells may also exert their protective effects through the secretion of TGF-, which has been demonstrated to act directly by down-regulating T cell autoimmunity or indirectly by modulating the function of APCs (27, 28). To test this hypothesis, we used HEL- or OVA-primed CD3 lymph node T cells as the responder cells and CD3 splenic cells from mice that had been treated with activated or control B cells as APCs, and characterized the Ag-presenting activity of APCs. APCs isolated from NOD mice 2 wk after treatment with control B cells presented Ag and stimulated responder T cell proliferation in a dose-dependent manner (Fig. 7). Their Ag-presenting activities were similar to those of APCs from unmanipulated NOD mice, contained in the sera of these mice revealed that mice treated with activated B cells developed only low levels of Abs against GAD65 and insulin B chain, similar to that of control mice (data not shown). Thus, our data suggest that treatment of prediabetic NOD mice with activated B cells, but not control B cells, abolishes spontaneous Th1 responses to cell autoantigens, which was not associated with the induction of regulatory Th2 responses. Treatment with activated B cells, but not control B cells, triggers apoptosis of mononuclear cells in the spleen Given that 1) activated T cells express high levels of membraneassociated Fas and are sensitive to apoptosis triggered by FasL, 2) activated B cells express membrane-associated FasL, and 3) the interaction of diabetogenic T cells with activated B cells leads to the apoptosis of both T and B lymphocytes in vitro (Fig. 4), we hypothesized that the ablation of Th1 immunity to cell Ags in FIGURE 6. Treatment with activated B cells, but not control B cells, triggers apoptosis of mononuclear cells in the spleen. NOD mice were treated twice at 4 and 10 wk of age with 10 7 LPS-activated B cells or control B cells. At 12 wk of age, B cell-treated and unmanipulated NOD mice were sacrificed and their pancreata and spleens were paraffin embedded and sliced into sections. The number of apoptotic mononuclear cells was determined by TUNEL assay. Data are presented as the average number SEM of apoptotic cells per islet or per view in the marginal zone area of the spleen from 25 slides of 5 mice for each group.

7 1086 LPS-ACTIVATED B CELLS INHIBIT AUTOIMMUNITY IN NOD MICE FIGURE 7. Treatment with activated B cells temporarily impairs the function of APCs. Female NOD mice were treated with 10 7 activated or control B cells at 4 and 10 wk of age. Two or 4 wk later, CD3 splenic mononuclear cells (APCs) isolated from B cell-treated or age-matched unmanipulated NOD mice were incubated with Ag-primed CD3 lymph node T cells in the presence of different concentrations of Ag (in triplicate) for 96 h to determine Ag-presenting activity by T cell proliferation assay. Ag-induced T cell proliferation was expressed as the mean cpm over background (n 6 for each group). A, T cell response to HEL. B, T cell response to OVA. Data shown were obtained using the APCs isolated from NOD mice 2 wk after the final B cell treatment. The background for medium alone ranged from 950 to 1320 cpm. The T cell response to 1 g/ml anti-cd3 showed a similar level of proliferation in all groups of mice (data not shown). Mice from control and experimental groups were simultaneously tested in two separate experiments for each Ag. suggesting that the treatment of prediabetic NOD mice with control B cells had little or no effect on modulating Ag-presenting activity of APCs. In contrast, the ability of APCs (from NOD mice 2 wk after the final treatment with activated B cells) to stimulate responder T cell proliferation to OVA and HEL was reduced by 60 and 85%, respectively, as compared with controls. However, 4 wk after treatment with activated B cells or control B cells, the ability of APCs from these mice to present OVA or HEL was indistinguishable from that of APCs isolated from unmanipulated mice (data not shown). Thus, the treatment of prediabetic NOD mice with activated B cells, but not control B cells, temporarily impaired the activity of APCs, which may have contributed to the ablation of spontaneous T cell autoimmunity to cell Ags. Activated B cells, but not control B cells, prevent the adoptive transfer of T1DM to NOD/scid mice Next, using a coadoptive transfer experimental system, we tested whether activated B cells can down-regulate an established diabetogenic T cell response. Ten million CD3 T cells isolated from newly diabetic NOD mice were mixed with an equal number of activated B cells or control B cells (from 2-wk-old NOD mice) and transferred to NOD/scid mice. Control groups of mice received 10 7 CD3 T cells or the same number of diabetogenic splenic mononuclear cells from newly diabetic mice. NOD/scid mice that received CD3 T cells plus control B cells developed diabetes within 3 4 wk, as did control mice which had received diabetogenic splenic cells from diabetic NOD mice (Fig. 8A). Most of the mice that had received CD3 T cells alone displayed a slight delay in T1DM onset, suggesting that B cells are unnecessary for the destruction of cells in the pancreatic islets, but may contribute to the amplificatory cascade of autoimmune responses. In contrast, all of the mice that had received CD3 T cells in combination with activated B cells remained disease free throughout the 10-wk experimental period, demonstrating that activated B cells can prevent the adoptive transfer of T1DM to NOD/scid mice. Importantly, coadoptive transfer of whole splenic mononuclear cells from newly diabetic NOD mice with activated B cells isolated from 6- to 8-wk-old prediabetic NOD mice also prevented the adoptive transfer of T1DM to NOD/scid mice (Fig. 8B). These findings suggest that activated B cells can down-regulate autoreactive effector T cells, even when the B cells are isolated from mice which already have an established diverse autoimmune response. However, coadoptive transfer of diabetogenic splenic cells with splenic mononuclear cells from NOD mice that had been treated with activated B cells, or control B cells, failed to protect the recipients from T1DM, suggesting the lack of a transferable regulatory cell response in the B cell-treated mice (Fig. 8C). Although 40% of NOD/scid mice that received splenic mononuclear cells from 15- wk-old (nondiabetic) NOD mice treated with control B cells developed a delayed onset of diabetes, NOD/scid mice that received splenic mononuclear cells from NOD mice that had been treated with activated B cells were diabetes free throughout the experimental period. This finding suggests that treatment with activated B cells inactivates effector T cells and/or inhibits the development of diabetogenic T cells in vivo. Our data provide the first demonstration that activated B cells in vivo can prevent effector T cells from mediating cell destruction. Discussion T1DM is thought to be an autoimmune disease mediated by proinflammatory Th1-type cells. Administration of autoantigens, such as GAD, insulin, HSP, as well as their peptides, can induce regulatory T cell responses, such as Th2, Tr1, and Th3, which are associated with an inhibition of autoimmune disease progression (25, 29 31). Down-regulation of autoimmunity to cell Ags can also be achieved by transfusion of CD4 CD25 T cells into prediabetic NOD mice (32). We observed that treatment of prediabetic NOD mice with activated B cells, but not control B cells, inhibited the spontaneous Th1 immunity to GAD65, HSP277, and insulin B chain. Furthermore, these treatments significantly reduced the incidence of spontaneous diabetes in prediabetic NOD mice. Although 90% of control mice developed diabetes by 27 wk of age, only 20% of mice treated with LPS-activated B cells became hyperglycemic over the 1-year observation period. Thus, treatment with activated B cells effectively down-regulates Th1 immunity to cell autoantigens and inhibits diabetes progression in prediabetic NOD mice.

8 The Journal of Immunology 1087 FIGURE 8. Activated B cells prevent the adoptive transfer of T1DM to NOD/scid mice. A, B cells were isolated from 2-wk-old NOD mice and activated by LPS. Activated B cells or control B cells were cotransferred with splenic T cells from newly diabetic NOD mice to NOD/scid mice. Positive control groups received T cells or splenic mononuclear cells only from newly diabetic NOD mice. B, B cells were isolated from 6-to 8-wkold NOD mice and activated by LPS. Activated B cells, or control B cells, were cotransferred with splenic mononuclear cells from newly diabetic NOD mice to NOD/scid mice. Positive controls received splenic cells only from newly diabetic NOD mice. C, Splenic mononuclear cells from 15- wk-old mice that had been treated with control B cells (F) or activated B cells (ƒ) were coadoptively transferred with splenic mononuclear cells isolated from newly diabetic NOD mice to NOD/scid mice. Control groups received diabetogenic splenic cells alone (E), splenic mononuclear cells from mice treated with control B cells alone ( ), or activated B cells alone ( ). Blood glucose levels were monitored frequently, and animals were considered diabetic after two consecutive blood glucose levels of 13 mmol/l (n 16 for A, n 12 for B, and n 10 for C). Several independent lines of evidence provide insights into the mechanism(s) underlying the efficacy of B cell-based immunotherapy. First, we observed that following activation, B cells expressed high levels of membrane-associated FasL. Notably, activated dendritic cells express FasL on their surface and induce activated T cells to apoptose during T-dendritic cell interactions in vivo (26). Second, we observed that the incubation of diabetogenic T cells with activated B cells in vitro led to the apoptosis of both T and B lymphocytes. Third, coadoptive transfer of activated B cells, but not control B cells, with diabetogenic T cells prevented the adoptive transfer of T1DM to NOD/scid mice. Furthermore, NOD mice that had received activated B cell treatment lacked adoptively transferable diabetogenic T cells. Finally, treatment of prediabetic NOD mice with activated B cells induced massive mononuclear cell apoptosis in the spleen and inhibited cell-specific T cell autoimmunity and diabetes progression. Collectively, these findings suggest that like activated dendritic cells, activated B cells trigger the apoptosis of diabetogenic T cells in vivo, contributing to the therapeutic action of activated B cells. However, our observations contrast with recent findings that B cells activated through the engagement of CD40 and IgM can protect activated T cells from AICD in vitro (18). These contrasting observations may stem from the different agents used to activate the B cells. Although both freshly isolated and activated B cells have the capacity to traffic into the pancreatic islets, treatment of prediabetic NOD mice with B cells did not significantly increase apoptotic signals in the pancreatic islets. The majority of both fluorescent dye DiI-labeled activated and control B cells migrated to the spleen and resided there at least 12 days after B cell transfusion (our unpublished observations). Treatment of prediabetic NOD mice with activated B cells, but not control B cells, induced massive mononuclear cell apoptosis predominantly in the marginal zone area near the periarterial lymphatic sheath of the spleen, where the T-B cell interactions are known to take place. Thus, the action of activated B cells in prediabetic NOD mice may predominantly take place in the peripheral lymph tissues. We also found that activated B cells secreted high levels of TGF- and that transfusion of activated B cells into prediabetic NOD mice inhibited spontaneous Th1 immunity to cell Ags and temporarily impaired the activity of APCs. These findings support the notion that TGF-, an anti-inflammatory cytokine, can inhibit T cell immunity by directly suppressing cytokine production of T cells and/or indirectly down-regulating the activity of APCs. Notably, TGF- can promote the expansion of Ag-specific Th2 cells and cause APCs to promote immune deviation (28). Although we administered a large number of activated B cells that secrete high levels of TGF-, the therapeutic effect of activated B cells is unlikely to be mediated by priming a regulatory T cell response. We did not detect any promotion of Ab responses to GAD65 or insulin B chain, nor did we detect Th2 responses to any cell Ags following B cell treatments. Furthermore, we did not detect a transferable regulatory T cell response from NOD mice that had received B cell treatment. Collectively, our findings suggest that the therapeutic effect of activated B cell treatment may be mediated by triggering the spontaneously primed autoreactive T cells to undergo apoptosis through Fas-FasL interactions and/or by inhibiting T cell and APC activities through the secretion of TGF-. The immunoregulatory functions of activated B cells may also contribute to the maintenance of self-tolerance (28, 33 35). Ag-based immunotherapies may cause unexpected side effects by promoting high Ab responses, and autoantigen-primed Th2 cells may cause autoimmune disease in immunocompromised mice (36 38). We observed that coadoptive transfer of diabetogenic T cells with splenic cells from NOD mice that had been previously treated with B cells failed to protect the recipients from T1DM, suggesting that treatment did not induce anti-inflammtory regulatory responses. Furthermore, splenic mononuclear cells from mice that had received activated B cells, but not control B cells, failed to adoptively transfer diabetes to NOD/scid mice. The lack of adoptively transferable diabetogenic T cells in NOD mice treated with activated B cells, but not control B cells, suggests that

9 1088 LPS-ACTIVATED B CELLS INHIBIT AUTOIMMUNITY IN NOD MICE activated B cells inactivate effector T cells and/or inhibit the development of diabetogenic T cells in vivo. Notably, activated B cells have a short life span as they express high levels of Fas on their surface and are also sensitive to apoptosis induction (39). Indeed, we found that activated B cells also undergo apoptosis after interaction with diabetogenic T cells in vitro and treatment with activated B cells caused the massive mononuclear cell apoptosis in the T-APC cell interaction area of the spleen in prediabetic NOD mice. Importantly, treatment of prediabetic NOD mice with activated B cells effectively inhibited disease progression. Collectively, these findings suggest that activated B cells may selectively inactivate or suppress spontaneously activated T cells in vivo. This suggests that B cells could be taken from a patient, activated in vitro, and transfused back into the patient. Thus, B cell-based immunotherapy may provide a new, safe, and effective means to prevent autoimmune disease progression. Notably, while most of the female NOD mice maintained under specific pathogen-free conditions develop spontaneous autoimmune diabetes, only a small percentage of NOD mice become hyperglycemic when those mice are raised in a natural environment. Our findings that activated B cells down-regulate pathogenic T cell responses may explain why NOD mice exposed to environmental pathogens have a greatly reduced disease incidence. Conceivably, B cells activated by microbial infection, or by microbial LPS alone, may inhibit spontaneous T cell immunity to cell Ags through a mechanism similar to that which we observed in prediabetic NOD mice, thereby preventing the development of diabetes. Our observations also suggest that during Gram-negative bacterial infection, bacterial LPS could activate B cells, which then promote T and B cell apoptosis. Thus, this finding may provide a new explanation for the development of lymphopenia which is associated with some cases of Gram-negative bacterial infections when the apoptosis of lymphocytes is not limited in scope (40). In summary, we have shown that activated B cells express high levels of membrane-associated FasL and secrete anti-inflammatory TGF-. Interaction of diabetogenic T cells with activated B cells, but not control B cells, leads to the apoptosis of both types of lymphocytes in vitro. Coadoptive transfer of activated B cells with diabetogenic T cells prevents the adoptive transfer of T1DM to NOD/scid mice. Treatment with activated B cells inhibits spontaneous T cell autoimmunity to cell autoantigens, enhances mononuclear cell apoptosis in the peripheral lymph tissue, and temporarily impairs the function of APCs in prediabetic NOD mice. Furthermore, this treatment inhibits progression of diabetes in prediabetic NOD mice. These data suggest that the down-regulatory effect of activated B cells on activated T cells may contribute to the maintenance of peripheral self-tolerance and promote lymphopenia during microbial infections. These findings also introduce a new explanation for the differential incidence of T1DM when NOD mice are raised in the natural environment vs one that is specific pathogen free. Overall, our findings provide new insights into the understanding of T cell and APC interactions and may aid in the design of immunotherapies for human autoimmune diseases. Acknowledgments We thank Dr. Michael Clare-Salzler (University of Florida, Gainesville, FL) for his critical comments. References 1. Mamula, M. J., and C. A. Janeway, Jr Do B cells drive the diversification of immune responses? Immunol. Today 14: Kurt-Jones, E. A., D. Liano, K. A. HayGlass, B. Benacerraf, M. S. Sy, and A. K. Abbas The role of antigen-presenting B cells in T cell priming in vivo: studies of B cell-deficient mice. J. Immunol. 140: Janeway, C. A., Jr., and K. Bottomly Signals and signs for lymphocyte responses. Cell 76: Constant, S., N. Schweitzer, J. West, P. Ranney, and K. Bottomly B lymphocytes can be competent antigen-presenting cells for priming CD4 T cells to protein antigens in vivo. J. Immunol. 155: Rossi-Bergmann, B., I. Muller, and E. B. Godinho TH1 and TH2 T-cell subsets are differentially activated by macrophages and B cells in murine leishmaniasis. Infect. Immun. 61: Hollsberg, P., V. Batra, A. Dressel, and D. A. Hafler Induction of anergy in CD8 T cells by B cell presentation of antigen. J. Immunol. 157: Adorini, L., J. C. Guery, F. Ria, and F. Galbiati B cells present antigen to CD4 T cells, but fail to produce IL-12: selective APC for Th2 cell development? Ann. NY Acad. Sci. 815: Croft, M., S. B. Joseph, and K. T. Miner Partial activation of naive CD4 T cells and tolerance induction in response to peptide presented by resting B cells. J. Immunol. 159: Castaño, L., and G. S. Eisenbarth Type-I diabetes: a chronic autoimmune disease of human, mouse, and rat. Annu. Rev. Immunol. 8: Atkinson, M. A., and N. K. Maclaren The pathogenesis of insulin-dependent diabetes mellitus. N. Engl. J. Med. 331: Delovitch, T. L., and B. Singh The nonobese diabetic mouse as a model of autoimmune diabetes: immune dysregulation gets the NOD. Immunity 7: Haskins, K., and D. Wegmann Diabetogenic T-cell clones. Diabetes 45: Tisch, R., and H. McDevitt Insulin-dependent diabetes mellitus. Cell 85: Serreze, D. V., H. D. Chapman, D. S. Varnum, M. S. Hanson, P. C. Reifsnyder, S. D. Richard, S. A. Fleming, E. H. Leiter, and L. D. Shultz B lymphocytes are essential for the initiation of T cell-mediated autoimmune diabetes: analysis of a new speed congenic stock of NOD.Ig null mice. J. Exp. Med. 184: Serreze, D. V., S. A. Fleming, H. D. Chapman, S. D. Richard, E. H. Leiter, and R. M. Tisch B lymphocytes are critical antigen-presenting cells for the initiation of T cell-mediated autoimmune diabetes in nonobese diabetic mice. J. Immunol. 161: Falcone, M., J. Lee, G. Patstone, B. Yeung, and N. Sarvetnick B lymphocytes are crucial antigen-presenting cells in the pathogenic autoimmune response to GAD65 antigen in nonobese diabetic mice. J. Immunol. 161: Noorchashm, H., Y. K. Lieu, N. Noorchashm, S. Y. Rostami, S. A. Greeley, A. Schlachterman, H. K. Song, L. E. Noto, A. M. Jevnikar, C. F. Barker, and A. Naji I-Ag7-mediated antigen presentation by B lymphocytes is critical in overcoming a checkpoint in T cell tolerance to islet cells of nonobese diabetic mice. J. Immunol. 163: Wang, J. K., S. T. Ju, and A. Marshak-Rothstein Protection of T cells from activation-induced cell death by Fas B cells. Eur. J. Immunol. 30: Hahne, M., T. Renno, M. Schroeter, M. Irmler, L. French, T. Bornard, H. R. MacDonald, and J. Tschopp Activated B cells express functional Fas ligand. Eur. J. Immunol. 26: Kaufman, D. L., M. Clare-Salzler, J. Tian, T. Forsthuber, G. S. Ting, P. Robinson, M. A. Atkinson, E. E. Sercarz, A. J. Tobin, and P. V. Lehmann Spontaneous loss of T-cell tolerance to glutamic acid decarboxylase in murine insulin-dependent diabetes. Nature 366: Elias, D., H. Marcus, T. Reshef, V. Ablamunits, and I. R. Cohen Induction of diabetes in standard mice by immunization with the p277 peptide of a 60-kDa heat shock protein. Eur. J. Immunol. 25: Deng, H., R. Apple, M. Clare-Salzler, S. Trembleau, D. Mathis, L. Adorini, and E. Sercarz Determinant capture as a possible mechanism of protection afforded by major histocompatibility complex class II molecules in autoimmune disease. J. Exp. Med. 178: Chen, Y., J. Inobe, R. Marks, P. Gonnella, V. K. Kuchroo, and H. L. Weiner Peripheral deletion of antigen-reactive T cells in oral tolerance. Nature 376: Tian, J., P. V. Lehmann, and D. L. Kaufman Determinant spreading of T helper cell 2 (Th2) responses to pancreatic islet autoantigens. J. Exp. Med. 186: Tian, J., M. Clare-Salzler, A. Herschenfeld, B. Middleton, D. Newman, R. Mueller, S. Arita, C. Evans, M. A. Atkinson, Y. Mullen, et al Modulating autoimmune responses to GAD inhibits disease progression and prolongs islet graft survival in diabetes-prone mice. Nat. Med. 2: Suss, G., and K. Shortman A subclass of dendritic cells kills CD4 T cells via Fas/Fas-ligand-induced apoptosis. J. Exp. Med. 183: Powrie, F., J. Carlino, M. W. Leach, S. Mauze, and R. L. Coffman A critical role for transforming growth factor- but not interleukin 4 in the suppression of T helper type 1-mediated colitis by CD45RB low CD4 T cells. J. Exp. Med. 183: King, C., J. Davies, R. Mueller, M. S. Lee, T. Krahl, B. Yeung, E. O Connor, and N. Sarvetnick TGF- 1 alters APC preference, polarizing islet antigen responses toward a Th2 phenotype. Immunity 8: Tisch, R., and H. O. McDevitt Antigen-specific immunotherapy: is it a real possibility to combat T-cell-mediated autoimmunity? Proc. Natl. Acad. Sci. USA 91: Daniel, D., and D. R. Wegmann Protection of nonobese diabetic mice from diabetes by intranasal or subcutaneous administration of insulin peptide B-(9 23). Proc. Natl. Acad. Sci. USA 93:956.