Little is known about the pathogenic role of B cell dysfunction in T cell-mediated autoimmune disease. We previously reported that B cell hyper-responsiveness, resistance to apoptosis, and accumulation in islets occur during the onset of insulitis, but not in type 1 diabetes (T1D), in NOD mice. In this study we extended these studies to further determine how islet-infiltrated B cells contribute to this inflammatory insulitis. We demonstrate the presence of an increased percentage of B7-1+ and a decreased percentage of B7-2+ B cells in the spleen of autoimmune disease-prone NOD and nonobese diabetes-resistant mice compared with the spleen of nonautoimmune disease-prone C57BL/6 and BALB/c mice. An age-dependent differential expression of B7-1 and B7-2 was associated with the development of insulitis and CD4+CD25+ T cell deficiency in autoimmune disease-prone mice. Whereas BCR and LPS stimulation increased B7-2 expression on B cells from autoimmune disease-prone and nonautoimmune disease-prone mice, LPS-induced B7-1 expression was higher on NOD than C57BL/6 B cells. Interestingly, increased expression of B7-1 and B7-2 was found on islet-infiltrated B cells, and this increase was associated with enhanced T cell costimulation. Islet-infiltrated B cells were shown to be a source of TNF-α production in islets. B7 blockade of BCR-stimulated NOD B cells by anti-B7-1 and anti-B7-2 mAbs during coadoptive transfer with diabetogenic T cells into NOD.scid mice protected these recipients from T1D. These results suggest that increased B7-1 and B7-2 expression on islet-infiltrated NOD B cells is associated with increased T cell costimulation and the development of inflammatory insulitis in NOD mice.
Autoimmune type 1 diabetes (T1D)3 is characterized by T cell-mediated destruction of insulin-producing β cells in the pancreatic islets (1). In the NOD mouse model of TID, not only CD4+ and CD8+ T cells, but also APCs, including B cells, macrophages (Mφ), and dendritic cells (DCs), infiltrate islets (2, 3, 4). APCs take up and present Ag to T cells in the context of MHC molecules (5). As APCs, B cells are highly efficient at Ag uptake due to their ability to bind and internalize specific Ags through cell surface-associated Igs (5).
Increased B7 expression on APCs is required for optimum Ag presentation to T cells. Resting spleen B cells from nonautoimmune disease-prone mouse strains express very low levels of the B7-1 and B7-2 costimulatory molecules (6). LPS- or BCR-mediated stimulation increases B7 expression on B cells (7, 8). Activated B cells constitute the major APC subpopulation in NOD islets before the onset of destructive insulitis (4). Previous reports demonstrate that B cells are required for optimum activation of CD4+ T cells in NOD mice (9, 10), and that NOD B cells are effective APCs (7, 8). Protection of B cell-depleted NOD mice from destructive insulitis and T1D further suggests a key role for B cells in the pathogenesis of T1D (9, 10).
B7 costimulation is also required for the generation of CD8+ CTL (11), which mediate islet β cell destruction in NOD mice (12, 13, 14). B7-1 provides a stronger costimulation than B7-2 in naive CD8+ T cells for proliferation and IL-2 production (15). B7-1 expression under the control of the rat insulin promoter (RIP-B7-1) in islets of nonautoimmune disease-prone mice does not initiate T1D onset (16); however, if RIP-B7-1 is coexpressed with a proinflammatory cytokine such as IL-2 (17) or TNF-α (16), a breakdown of self-tolerance ensues. In contrast, the expression of RIP-B7-1 alone on NOD islet β cells accelerates T1D onset in these mice (18), suggesting the existence of an inflammatory environment in NOD islets.
Several studies have investigated the role of B7 costimulation in the pathogenesis of autoimmune diseases (19, 20, 21, 22), including T1D (21, 23, 24, 25). In NOD mice, the role of B7 costimulation was investigated by disrupting B7/CD28 interactions by administration of either neutralizing anti-B7-1 and anti-B7-2 mAbs (25) or a CTLA-4Ig fusion protein (25) or by using B7- and CD28-deficient mice (23). NOD mice are protected from T1D upon receipt of anti-B7-2 or CTLA-4Ig treatment well before the onset of insulitis (25), but this therapy fails if started after the development of invasive insulitis. Similarly, anti-CD28 treatment, if administered before insulitis, protects NOD mice from T1D (26). In contrast, the onset of T1D is enhanced in B7- and CD28-deficient mice (23), possibly due to defective CD4+CD25+ T regulatory (Treg) cell homeostasis (23). Thus, CD28-B7-2 interactions appear to be deficient and mediate the pathogenesis of T1D in NOD mice.
Despite accumulating evidence on the role of B7 costimulation in the pathogenesis of T1D, information regarding the precise role of B7-1 and B7-2 expression by B cells in T1D is lacking. In this study we report that before the onset of destructive insulitis, a higher percentage of B7-1+ B cells and a decreased percentage of B7-2+ B cells are present in the spleens of autoimmune disease-prone, T1D-susceptible NOD and T1D-resistant nonobese diabetes-resistant (NOR) mice compared with nonautoimmune disease-prone C57BL/6 (B6) and BALB/c mice. Islet-infiltrated B cells produce TNF-α and show elevated levels of B7-1 and B7-2 expression. The latter increase in B7-1 and B7-2 was associated with enhanced T cell costimulation, as detected by the ability of B7 blockade of BCR-stimulated B cells to inhibit the transfer of T1D by diabetogenic T cells into NOD.scid mice. Our results suggest that the presence of an increased percentage of B7-1+ B cells in the spleen and of B7-1+ and B7-2+ B cells in the islets of NOD mice, particularly TNF-α-producing islet B cells, may be important in the development of inflammatory insulitis and the expansion of autoreactive T cells in the islets of autoimmune disease-prone mice.
Materials and Methods
NOD/Del, NOR/Lt, and NOD.scid mice were bred in a specific pathogen-free barrier facility at Robarts Research Institute. In our colony of female NOD mice, islet infiltration begins at 4–6 wk of age and progresses to destructive insulitis and overt diabetes by 4–6 mo of age. B6 and BALB/c mice were purchased from Charles River Laboratories. Female mice (4- to 30-wk-old) were used in this study.
Cell isolation and purification
Peritoneal cells were collected by peritoneal lavage with PBS. Islets were isolated by collagenase digestion of the pancreas (27), then were cultured overnight in complete RPMI 1640 to obtain islet-infiltrating cells as previously described (4). Cells that migrated out from the islets were harvested and used in cell proliferation assays and flow cytometric analyses. Spleen and pancreatic draining lymph node (PLN) cells were prepared as described previously (27). Spleen B cells were purified (purity, ≥98%) using a B cell enrichment mixture according to the manufacturer’s instructions (StemCell Technologies). T cells were purified (purity, ≥95%) using T cell columns (R&D Systems). Splenic DC were purified using CD11c microbeads (Miltenyi Biotec) following the manufacturer’s instructions. Splenic Mφ were purified by plastic adherence (28). DC and Mφ depletion from spleen, PLN, and islet-infiltrated cells was performed by removal of low density and plastic-adhering cells (28, 29). This procedure depletes ≥95% of Mφ and DC; after depletion, ≥98% of spleen, PLN, and islet-infiltrated cells were found to consist of T and B cells.
Spleen, PLN, and islet-infiltrating cells were reacted (15 min, 4°C) with Fc block (anti-CD16/32) to reduce the nonspecific binding of test Abs. The cells were then stained (45 min, 4°C) with CyChrome-, FITC-, and PE-conjugated anti-B220, anti-B7-1, anti-B7-2, anti-CD4, anti-CD5, anti-CD25, anti-CD11b, anti-CD11c, and anti-TNF-α mAbs. All mAbs were obtained from BD Biosciences. The cells were washed three times with PBS containing 0.1% sodium azide and 2% FCS. Intracellular TNF-α staining was performed using an intracellular staining kit (BD Biosciences) as follows. Briefly, after surface staining, cells were fixed for 20 min, washed twice with wash buffer, and permeabilized for 20 min before staining with anti-TNF-α. The cells were then washed and analyzed by flow cytometry using CellQuest software (BD Biosciences) as previously described (4).
Islet-infiltrated and spleen B cells (2 × 106/ml) were cultured for 0–48 h in the presence or the absence of ionomycin (500 ng/ml) plus PMA (5 ng/ml; Sigma-Aldrich) and anti-IgM F(ab′)2 Ab (5 μg/ml) plus anti-CD40 (10 μg/ml; BD Biosciences), respectively. Cell supernatants were collected and frozen at −70°C until use. The concentration of TNF-α secreted was determined using a mouse TNF-α ELISA kit (R&D Systems) according to the manufacturer’s instructions.
Mφ- and DC-depleted spleen, PLN, and islet-infiltrated cells, anti-IgM F(ab′)2 Ab (Jackson ImmunoResearch Laboratories)-stimulated B cells (105/well), LPS-stimulated Mφ, and DC plus purified T cells (105/well) were cocultured in complete RPMI 1640 supplemented with 10% heat-inactivated FCS, 10 mmol/L HEPES buffer, 1 mmol/L sodium pyruvate, 2 mmol/L l-glutamine, 100 U/ml penicillin, 0.1 mg/ml streptomycin, and 0.05 μmol/L 2-ME (Invitrogen Life Technologies) in 96-well tissue culture plates for 64 h at 37°C in 5% CO2. Spleen (105/well) and peritoneal (105/well) B cells were cultured under similar conditions. [3H]Thymidine (1 μCi/well) was added during the last 18 h of culture, and cells were harvested and assayed for [3H]thymidine incorporation (cpm).
Adoptive transfer of BCR-stimulated B cells with diabetogenic T cells into NOD.scid mice
NOD.scid mice (6- to 7-wk-old) were injected i.v. with 5 × 106 anti-IgM F(ab′)2-stimulated (48 h) B cells pretreated with neutralizing anti-B7-1 and anti-B7-2 mAbs together with equal numbers of diabetogenic spleen T cells from 13-wk-old prediabetic female NOD mice. To boost B cell costimulation, a second dose of BCR-stimulated B cells pretreated with anti-B7-1 and anti-B7-2 mAbs was injected 3 wk later. B7-1 and B7-2 blockade was performed by reacting (45 min, 4°C) activated B cells with purified anti-B7-1 (clone 16-10A1) and anti-B7-2 (clone GL1) mAbs (40 μg/ml). To determine whether the neutralizing anti-B7-1 and anti-B7-2 mAbs used are toxic to B cells, the anti-IgM F(ab′)2-induced proliferative responses of anti-B7 and isotype control IgG-treated B cells were compared. The blockade of B7-1 and B7-2 on stimulated B cells was confirmed by the lack of surface staining with PE-labeled anti-B7-1 and anti-B7-2 after treatment with the neutralizing anti-B7 mAbs. Activated B cells in a control group were incubated with isotype IgG (40 μg/ml), washed three times with ice-cold PBS, and injected into mice. T1D incidence was determined by measuring blood glucose levels (BGL) twice weekly starting at 4 wk post-transfer. A BGL >11 mM for two consecutive readings was considered positive for T1D.
Statistical analyses of the data were performed using Student’s t test, Mann-Whitney U test, and log-rank test where appropriate. Data are presented as the mean ± SD.
B7-1 and B7-2 are differentially expressed on B cells from autoimmune disease-prone and nonautoimmune disease-prone mice
We recently found that a significantly higher percentage of peripheral B cells in the spleen from autoimmune disease-prone, T1D-susceptible NOD and T1D-resistant NOR mice display an activated phenotype (CD69+) compared with spleen B cells from nonautoimmune disease-prone B6 and BALB/c mice (4). To understand the functional relevance of this increased activated B cell phenotype in autoimmune disease-prone mice, we took advantage of the fact that activated peripheral B cells display an increased surface expression of B7-1 and B7-2 (30, 31). Thus, we determined whether peripheral B cells from NOD and NOR mice differ in their constitutive and/or activated levels of expression of B7-1 and B7-2 relative to spleen B cells from B6 and BALB/c mice. FACS analyses of B7-1 and B7-2 expression on B220+ spleen B cells from mice of different ages showed that at 4–12 wk of age, the percentages of B7-1+ spleen B cells detected in NOD and NOR mice (∼17%) were ∼2- to 3-fold higher than those in age-matched B6 and BALB/c mice (∼6–7%) (Fig. 1,A). These differences were absent between 15 and 30 wk of age, because similar low percentages of spleen B cells (∼5%) were found in older NOD, NOR B6, and BALB/c mice. The age-dependent profile of B7-2 expression was similar to that of B7-1 expression, with the exception that a decreased percentage of B7-2+ spleen B cells was found during the first 4–6 wk of age in NOD and NOR mice (∼15%) compared with that in B6 and BALB/c mice (∼20%; Fig. 1 B). Thereafter, between 9 and 30 wk of age, no significant differences in the percentages of B7-2+ B cells were observed. These results demonstrate that spleen B cells from autoimmune disease-prone mice (NOD and NOR) possess a higher percentage of B7-1+, but not B7-2+, B cells than spleen B cells from nonautoimmune disease-prone (B6 and BALB/c) mice.
Next, we analyzed the levels of B7-1 and B7-2 expression on BCR- and LPS-stimulated spleen B cells from NOD and B6 mice. Significant increases in the percentage of B7-1+ B cells were observed in LPS-activated, but not in anti-IgM F(ab′)2-activated, NOD and B6 B cells (Fig. 1,C), with the highest increase evident in LPS-activated NOD B cells. In contrast, both anti-IgM F(ab′)2 and LPS induced a similar 3-fold increase in the number of NOD and B6 B7-2+ B cells (Fig. 1,D). It is known that anti-IgM F(ab′)2 stimulation through the BCR induces B7-2, but not B7-1, expression on B cells, whereas LPS stimulation via the TLR4 enhances both B7-1 and B7-2 expression on B cells (32). Thus, although NOD and B6 spleen B cells show similar levels of B7-2 expression after BCR and LPS stimulation, B cells from NOD and B6 mice differ in their constitutive (Fig. 1,A) and induced (Fig. 1 B) levels of expression of B7-1.
Decreased frequency of B7-2+ B cells correlates with a CD4+CD25+ T cell deficiency in NOD spleen
B7-2 provides a strong T cell costimulatory signal and is required to maintain the homeostasis of CD4+CD25+ T cells (23). NOD mice are deficient in the number of spleen B7-2+ B cells (Fig. 1,B) and CD4+CD25+ T cells (33). To determine whether there is a correlation between these deficiencies in NOD B7-2+ B cells and CD4+CD25+ T cells, the age-dependent variations in the number of CD4+CD25+ spleen T cells were examined in NOD, NOR, B6, and BALB/c mice. Interestingly, we found that CD4+CD25+ spleen T cells are deficient in autoimmune disease-prone NOD and NOR mice relative to those in nonautoimmune disease-prone B6 and BALB/c mice (Fig. 2). This T cell deficiency was greater at 4–6 wk of age (57 and 66% CD4+CD25+ T cells in NOD and NOR mice, respectively) than at 9–12 wk (41 and 46% CD4+CD25+ T cells in NOD and NOR mice), respectively (Fig. 2). These results demonstrate that a correlation exists between a decreased percentage of B7-2+ spleen B cells and an impaired homeostasis of CD4+CD25+ T cells in NOD mice.
Islet-infiltrated NOD B cells express high levels of B7 and produce TNF-α
We recently reported that in 10-wk-old NOD mice, ∼48% of the islet-infiltrating cells are B cells, and ∼25% of these B cells possess an activated phenotype (CD69+) (4). To determine whether this activated phenotype is also manifested by increased B7-1 and B7-2 expression, triple-color FACS analyses of islet-infiltrated B cells were conducted. Approximately 16 and 80% of islet-infiltrated B220+ B cells were found to be B7-1+ and B7-2+ B cells, respectively (Fig. 3,A). Interestingly, most B7-1+ B cells were also positive for B7-2 (Fig. 3 A). Thus, the vast majority of islet-infiltrated B cells express B7.
Transgenic expression of RIP-B7-1 in islets does not induce T1D in nonautoimmune disease-prone mice (16), but accelerates T1D in autoimmune disease-prone NOD mice (18). However, the coexpression of TNF-α with B7-1 (RIP-B7-1/TNF-α) on islet β cells renders nonautoimmune disease-prone mice hyperglycemic (20). In NOD mice, islet-infiltrated Mφ and DC produce TNF-α (3). We found that ∼18% of islet-infiltrated B cells also produce TNF-α, as revealed by intracellular TNF-α staining (Fig. 3,B) and TNF-α secretion by PMA- plus ionomycin-stimulated, islet-infiltrated and spleen B cells (Fig. 3,C). Islet-infiltrated B cells secreted significantly more TNF-α than spleen B cells (p < 0.05). Moreover, PMA and ionomycin or anti-IgM F(ab′)2 and anti-CD40 stimulated TNF-α secretion by NOD, NOR, and B6 B cells in a time-dependent manner (Fig. 3, D and E). Note that NOD B cells secreted significantly higher levels of TNF-α than either NOR or B6 B cells (p < 0.05; Fig. 3, D and E). These results suggest that islet-infiltrated, TNF-α-producing B cells may promote the establishment of an inflammatory environment that results in islet β cell destruction.
Islet-infiltrated B cells are not B1 B cells in NOD mice
More surface CD5+ B1 B cells are found in autoimmune disease-prone than nonautoimmune disease-prone mice (34, 35). CD5 negatively regulates signals delivered by the BCR and induces hyporesponsiveness to BCR stimulation (36). To determine whether an increased percentage of B1 B cells is present in an NOD islet infiltrate, B220+ B cells from islet infiltrates, spleen, PLN, and peritoneal cells were examined for their level of surface CD5 expression. Interestingly, the percentage of B220+CD5+ B1 cells was found to be similar (∼2–3%) in islet-infiltrated spleen and PLN cells (Fig. 4,A). However, as expected, a 10-fold increased percentage (23%) of B220+CD5+ B1 cells was observed in peritoneal cells (Fig. 4,A). To further confirm that the majority of islet-infiltrated B cells are not B1 cells, the responsiveness to BCR stimulation was compared between NOD islet-infiltrated B cells and peritoneal B cells. BCR stimulation elicited a 10-fold higher response in islet-infiltrated B cells than in peritoneal B cells (Fig. 4 B). This result also suggests that islet-infiltrated B cells are not B1 cells in NOD mice.
Islet-infiltrated NOD B cells provide costimulation to T cells
The presence of an increased percentage of B7-1+ and B7-2+ B cells in the islets of NOD mice suggests that B cells may be important in providing costimulation to T cells for their expansion. We tested this possibility by examining the autoproliferation in vitro of Mφ- and DC-depleted NOD spleen, islet-infiltrated, and PLN cells in the presence or the absence of anti-B7-1, anti-B7-2, or anti-B7-1 and anti-B7-2 mAbs. As expected, islet-infiltrated lymphocytes, which contain an increased percentage of B7-1+ and B7-2+ B cells, yielded a higher autoproliferative response than spleen cells and PLN cells (Fig. 5, A–C). Exogenously added anti-B7-1 (Fig. 5,A) or anti-B7-2 (Fig. 5,B) was equally effective at reducing the autoproliferation of islet-infiltrated lymphocytes, spleen cells, and PLN cells compared with the isotype control IgG (p < 0.001). An even lower autoproliferative response was observed upon addition of both anti-B7-1 and anti-B7-2 vs the isotype IgG control in all three cell populations tested (Fig. 5 C). The decreased proliferation obtained with PLN cells (consisting of ∼10% B cells) compared with spleen cells and islet-infiltrated cells (consisting of ∼35–48% B cells) further suggests a costimulatory role for B cells in the expansion of T cells.
Anti-IgM F(ab′)2 Ab-stimulated NOD B cells enhance lymphocyte autoproliferation
Anti-IgM F(ab′)2 stimulation preferentially enhances B7-2 expression on B cells (32) (Fig. 1, C and D). To determine whether anti-IgM F(ab′)2-stimulated purified spleen B cells can enhance lymphocyte autoproliferation if cocultured with spleen T cells, unstimulated or anti-IgM F(ab′)2-stimulated B cells and T cells were cocultured for 60 h. Significantly increased lymphocyte autoproliferation (p < 0.001) was detected when T cells were cocultured with anti-IgM F(ab′)2-stimulated B cells compared with T cells cocultured with unstimulated B cells (Fig. 5,D). Furthermore, increased lymphocyte autoproliferation was observed when purified B cells from 9- to 12-wk-old NOD mice were cocultured with syngeneic T cells compared with cocultures of syngeneic T cells with purified B6 or BALB/c B cells (Fig. 6,A). In addition to activated B cells, which comprise ∼48% of islet infiltrates (4), activated Mφ and DC (∼5% each) are also present in NOD islet infiltrates (Fig. 6,B). To elucidate the relative costimulatory capacities of different APC subsets that infiltrate islets, activated NOD B cells, Mφ, and DC were cocultured with NOD T cells at different APC:T cell ratios (1:1 to 1:50) for 64 h. Before coculture with T cells, Mφ and DC were also stimulated with LPS to increase B7 expression, as observed on islet-infiltrated Mφ and DC (Fig. 6,B). Although all three APC subsets increased lymphocyte autoproliferation at a 1:1 ratio, coculture of activated B cells plus T cells at a ratio of 1:10 yielded a significantly increased autoproliferation (p < 0.05) compared with cocultures of activated Mφ or DC with T cells at a 1:10 ratio. Note that the percentage of B cells in islet infiltrates is ∼10-fold higher than that in Mφ and DC (∼48 vs ∼5%). Thus, lymphocyte autoproliferation in B cell plus T cell cocultures at a 1:1 ratio is 4-fold greater than that achieved in cocultures of Mφ or DC with T cells at a 1:10 ratio (Fig. 6 C). Coupled with the results presented in the previous section, these findings indicate that activated B cells may provide sufficient costimulation to T cells for their expansion.
Blockade of B7-1 and B7-2 on activated NOD B cells reduces their ability to cotransfer T1D
The expression pattern of B7-1 and B7-2 on islet-infiltrated B cells resembles that on anti-IgM F(ab′)2-stimulated spleen B cells (Fig. 1, C and D, and Fig. 3,A). To circumvent the problem of obtaining sufficient numbers of B cells from NOD islet infiltrates for adoptive cotransfer experiments, diabetogenic T cells from 13-wk-old NOD mice were cotransferred into NOD.scid recipients with anti-IgM F(ab′)2-stimulated spleen B cells after B7-1 and B7-2 blockade by anti-B7-1 and anti-B7-2 mAbs. Evidence for B7-1 and B7-2 blockade was obtained by staining of the neutralizing anti-B7-1 and anti-B7-2 mAb-pretreated activated B cells with PE-labeled anti-B7-1 and anti-B7-2 mAbs. Although ∼13% B7-1+ and 79% B7-2+ BCR-stimulated B cells were detected before Ab blockade, no B7-1+- or B7-2+-activated B cells were detected after Ab blockade (Fig. 7,A). Anti-IgM F(ab′)2 stimulation induced similar levels of proliferation in both anti-B7 and control IgG-treated B cells (Fig. 7,B), suggesting that anti-B7 mAbs were not toxic to B cells. Furthermore, the transfer of diabetogenic T cells with anti-IgM F(ab′)2-stimulated B cells in the absence of B7 blockade yielded a 100% incidence of T1D at 8 wk post-transfer, whereas only 30% incidence of T1D was observed at 12 wk post-transfer upon transfer of T cells with the stimulated B cells after B7 blockade (Fig. 7 C; p < 0.05). Thus, B7 blockade significantly reduced the ability of these BCR-stimulated B cells to activate diabetogenic T cells and elicit the onset of T1D. These results suggest that B7-1- and B7-2-mediated B cell costimulation is required for the expansion of diabetogenic T cells and transfer of T1D into NOD.scid recipients.
This study investigates the role of B cell expression of B7-1 and B7-2 in the pathogenesis of T1D in NOD mice. Our results demonstrate the presence of an increased percentage of B7-1+ B cells in the spleen and islets of these mice. These findings agree with those of a previous report that B7-1 and B7-2 expression are increased on lymph node (LN)-derived NOD B220+CD69+ B cells (31). Although the functional significance of an increased percentage of B7-1+ B cells in the spleen (Fig. 1,A), LN (31), and islets (Fig. 3 A) of NOD mice is not known, the detection of these B7-1+ B cells may be associated with an ongoing inflammatory process. Note that the presence of an increased frequency of B7-1+ B cells in the spleen of proteolipid protein-induced relapsing experimental autoimmune encephalomyelitis in SJL/J mice (37) and in the peripheral blood of human multiple sclerosis patients (38) also supports the concept of B cell participation in an inflammatory response. Increased B7-1 expression on lymphocytes in proteolipid protein-induced experimental autoimmune encephalomyelitis in SJL/J mice (22) is also associated with epitope spreading and clinical relapses. B7-1 costimulation alone is not sufficient to induce T1D in nonautoimmune disease-prone mice (16), but its expression on APCs is important in both the T cell priming and effector phases of T1D (39).
There is mounting evidence that the presence of an increased percentage of B7-1+ B cells in the periphery is associated with inflammatory diseases (37, 38, 40). Increased B7-1 expression on LPS-stimulated B cells stimulates allogeneic T cell proliferation and can polarize toward a Th1-type immune response by enhancing IFN-γ, but not IL-4, production by T cells (41). The detection of an increased percentage of B7-1+ spleen B cells in NOD and NOR mice before the onset of destructive insulitis coupled with our recent report (4) that hyperactivated B cells accumulate in the islets of these mice suggest that B7-1+-activated B cells in the spleen (Fig. 1 A) promote the development of insulitis. In addition, protection from insulitis in NOD mice in which B cells were depleted from birth by injection of an anti-IgM Ab also supports the idea that B cells are required to initiate insulitis in NOD mice (9).
B7 costimulation is required for the generation of CTL responses in vivo (11). B7-1 provides a stronger costimulation than B7-2 for proliferation and IL-2 production by naive CD8+ T cells (15). Keeping in mind that CD8+ T cells elicit islet β cell destruction in NOD mice (12, 13, 14), it will prove interesting in future studies to determine whether the increased B7-1 expression on NOD B cells observed in this study has any impact on the generation of CD8+ CTL. RIP-B7-1 expression in islets of nonautoimmune disease-prone mice does not initiate T1D onset, but the coexpression of RIP-B7-1 with IL-2 (17) or TNF-α (16) in islets leads to the breakdown of self-tolerance. Importantly, transgenic expression of RIP-B7-1 alone on islet β cells accelerates T1D onset in NOD mice (18), which suggests the presence of an inflammatory environment in the islets of wild-type NOD mice. Moreover, the expression of RIP-B7-1 on islet β cells in B cell-deficient NOD mice (NOD.μMT−/−), which do not develop insulitis and T1D, induces T1D (18). The latter finding supports the idea that costimulation by islet-infiltrated B cells in NOD mice may be sufficient to induce T1D. Although Mφ and DC are the major source of TNF-α in NOD islets (3), our results identify activated B cells as an additional source of TNF-α production in islets (Fig. 3, B and C). Thus, TNF-α secretion by islet-infiltrated B7+ B cells may also contribute to the development of an inflammatory insulitis in NOD mice.
In contrast to a report by Chiu et al. (31), we detected a significant decrease, albeit small, in spleen B7-2+ B cells in 4- to 6-wk-old NOD and NOR mice relative to nonautoimmune disease-prone B6 and BALB/c mice. Note that Chiu et al. (31) examined B7-1 and B7-2 expression on B220+CD69+ activated B cells, whereas in our study B7-1 and B7-2 expression was determined on B220+ B cells, only 5–10% of which consist of CD69+ B cells. In contrast, we found an increased percentage of B7-2+ B cells in islets, where 20–25% of infiltrated B cells exhibit an activated phenotype (CD69+) (4). Our observation that BCR stimulation elicits higher proliferative responses by PLN-derived and islet-infiltrated B cells than spleen B cells (4) is consistent with reports that more activated B cells are present in the PLN and islets than in the spleen (4, 6). It is tempting to speculate that many of these B cells may localize to PLN and islets, because these sites may provide a source of islet autoantigens that stimulate the activation and expansion of these autoreactive B cells.
The significance of a slightly decreased percentage of B7-2+ B cells in the spleens of 4- to 6-wk-old autoimmune disease-prone NOD and NOR mice (Fig. 1,B) is not presently known. However, this observation may be related functionally to the CD4+CD25+ Treg cell deficiency (23, 33) and T cell hyporesponsiveness (26, 42) noted in these mice at this age. Decreased B7-2 expression occurs on spleen DC and Mφ from 8-wk-old NOD mice, and this reduced B7-2 expression on APCs is associated with the defective up-regulation of CTLA-4 on T cells and T cell hyporesponsiveness to anti-CD3 stimulation (42). The latter finding supports our results indicating decreased autoproliferation of Mφ- and DC-depleted splenocytes, which consist of fewer B7-2+ B cells (Fig. 1,B) than detected in Mφ- and DC-depleted islet-infiltrated cells (Fig. 3 A). Collectively, our results suggest that presence of a decreased percentage of B7-2+ B cells in NOD mice in conjunction with the reduced expression of B7-2 on other APCs (42) before the onset of invasive and destructive insulitis may be responsible for a CD4+CD25+ Treg cell deficiency (23, 33) and T cell hyporesponsiveness to TCR stimulation (26, 42).
In NOD mice, B cells comprise ∼48% of cells in an islet infiltrate (4), and ∼70–80% of these B cells express B7-2 (Fig. 3,A). Increased B7-2 expression on B cells after Ag stimulation is an indication of an activated phenotype (6, 31) that may enhance T cell expansion after binding to CD28 (43). The presence of a higher percentage of activated B cells in the PLN of NOD mice (31) also suggests that after Ag uptake, B cells may migrate to regional LN and stimulate the expansion of autoreactive T cells (43). The hyperproliferative responsiveness of islet-infiltrated and PLN B cells to BCR stimulation (4) indicates the presence of activated B cells in islets and PLN (31), which may provide costimulation for the expansion of autoreactive T cells (43, 44). This idea is further supported by our findings that Mφ- and DC-depleted islet-infiltrated cells undergo increased autoproliferation compared with Mφ- and DC-depleted splenocytes (Fig. 5). Accordingly, addition of anti-B7-1 and/or anti-B7-2 mAbs decreased the autoproliferation of Mφ- and DC-depleted islet-infiltrated, spleen and PLN cells, suggesting a role for B7 expressed on NOD B cells in lymphocyte autoproliferation.
Our in vivo studies demonstrated that the adoptive cotransfer of diabetogenic T cells with BCR-stimulated B cells after B7 blockade by anti-B7-1 and anti-B7-2 mAbs into NOD.scid mice decreased the incidence of T1D relative to that in NOD.scid recipients of diabetogenic T cells plus isotype control IgG-treated, BCR-stimulated B cells. These results appear to differ from a previous report that protection of NOD.scid mice from T1D occurs after adoptive cotransfer of LPS-activated NOD B cells with T cells from diabetic NOD mice (45). Note that B cells were activated by different stimuli in these two studies; an anti-IgM F(ab′)2 Ab was used in our study (Fig. 6 C), whereas LPS was used in the other study (45). Anti-IgM F(ab′)2 Ab signals through the BCR, whereas LPS signals through TLR-4 (46). Moreover, LPS stimulation induces the expression of both B7-1 and B7-2 on B cells (30, 31), whereas anti-IgM F(ab′)2 stimulation induces only B7-2 expression on B cells (32). Interaction of B7-1 and B7-2 with CD28 also induces distinct signaling pathways. For example, T cell stimulation by B7-1, but not B7-2, stimulates the tyrosine phosphorylation of CD28 (47). Because B cells can survive only for 2–3 wk after transfusion into a syngeneic host (7, 48), we repeated BCR-stimulated B cell transfusion 3 wk after the initial adoptive transfer to provide continuous costimulation. In contrast, Tian et al. (45) used only a single B cell transfusion. Furthermore, LPS stimulation induces increased expression of Fas ligand (FasL) by B cells (45, 49), whereas anti-IgM F(ab′)2 stimulation fails to induce FasL on B cells (4). Thus, LPS-stimulated FasL+ B cells may protect NOD.scid mice from T1D by deleting diabetogenic T cells through Fas:FasL interaction (45).
Phenotypic and functional properties of B cells, such as B7 expression (Fig. 1) and hyper-responsiveness to BCR stimulation (4), are similar in NOD and NOR mice, yet NOR mice are resistant to T1D. T1D is a multigenic disease, and NOD and NOR mice share ∼88.8% of the genome, including the MHC-linked I-Ag7 gene in the Idd1 susceptibility locus (50). The remaining 11.2% of the NOR genome is shared with nonautoimmune disease-prone B6 mice (50). The presence of peri-insulitis in NOR mice (2, 50) and the protection of B cell-deficient NOD mice from insulitis (9, 10) suggest that phenotypic and functional similarities in NOD and NOR B cells are linked to the development of insulitis (4). Thus, the factor(s) required to induce invasive insulitis and T1D onset may be encoded by the remaining 11.2% of the NOD genome.
In conclusion, we demonstrate for the first time that in T1D-prone, NOD mice, B7 expression is dysregulated on B cells. The presence of an increased percentage of B7-1+ B cells in the spleen of NOD mice may demarcate an ongoing inflammatory process and give rise to insulitis. The increased percentage of B7-1+ and B7-2+ B cells in NOD islets and the decreased incidence of T1D that results from the transfer of BCR-stimulated NOD B cells after B7 blockade highlight the importance of B7-mediated T cell costimulation in the pathogenesis of T1D.
We thank all members of our laboratory for their advice and encouragement.
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.
This work was supported by a grant from the Canadian Diabetes Association in honor of the late Olive I. Moore and a grant from the Ontario Research and Development Challenge Fund. S.H. is the recipient of a postdoctoral fellowship from the Canadian Diabetes Association in honor of the late Flora I. Nichol. T.L.D. is the Sheldon H. Weinstein Scientist in Diabetes at Robarts Research Institute and University of Western Ontario.
Abbreviations used in this paper: T1D, type 1 diabetes; BGL, blood glucose level; DC, dendritic cell; FasL, Fas ligand; LN, lymph node; Mφ, macrophage; NOR, nonobese diabetes-resistant; PLN, pancreatic draining lymph node; RIP, rat insulin promoter; Treg, T regulatory cell.