Abstract
The determinant spreading of T cell autoimmunity plays an important role in the pathogenesis of type 1 diabetes and in the protective mechanism of Ag-based immunotherapy in NOD mice. However, little is known about the role of APCs, particularly B cells, in the spreading of T cell autoimmunity. We studied determinant spreading in NOD/scid or Igμ−/− NOD mice reconstituted with NOD T and/or B cells and found that mice with mature B cells (TB NOD/scid and BMB Igμ−/− NOD), but not mice that lacked mature B cells (T NOD/scid and BM Igμ−/− NOD), spontaneously developed Th1 autoimmunity, which spread sequentially among different β cell Ags. Immunization of T NOD/scid and BM Igμ−/− NOD mice with a β cell Ag could prime Ag-specific Th1 or Th2 responses, but those T cell responses did not spread to other β cell Ags. In contrast, immunization of TB NOD/scid and BMB Igμ−/− NOD mice with a β cell Ag in IFA induced Th2 responses, which spread to other β cell Ags. Furthermore, we found that while macrophages and dendritic cells could evoke memory and effector T cell responses in vitro, B cells significantly enhanced the detection of spontaneously primed and induced Th1 responses to β cell Ags. Our data suggest that B cells, but not other APCs, mediate the spreading of T cell responses during the type 1 diabetes process and following Ag-based immunotherapy. Conceivably, the modulation of the capacity of B cells to present Ag may provide new interventions for enhancing Ag-based immunotherapy and controlling autoimmune diseases.
Type 1 diabetes (T1D)3 is thought to be a T cell-mediated autoimmune disease (reviewed in Refs.1, 2, 3, 4). Previous studies have suggested that the determinant spreading of T cell autoimmunity plays an important role in both the pathogenesis of T1D and in the protective mechanisms of Ag-based immunotherapy (5, 6, 7, 8, 9). Studies of the T1D process in NOD mice have shown that spontaneous proinflammatory Th1 responses sequentially spread among β cell Ags, such as glutamic acid decarboxylase (GAD), the dominant determinant of heat shock protein (HSP) 277, and insulin, creating a hierarchy of determinant spreading (5, 6, 10). Ag-based immunotherapies, by treatment with a β cell Ag to prime Th2 responses, result in the rapid spread of Th2 responses to other β cell Ags and the inhibition of disease progression in NOD mice (7, 8, 9, 11, 12). Presumably, these spreading processes are due to the cytokines produced by the first wave of autoreactive T cells, which create a microenvironment in the pancreatic islets and/or lymph nodes, leading to recruitment of second wave autoreactive T cells toward similar phenotype by positive feedback mechanisms (7, 13, 14). However, the role of APCs, particularly B cells, in the spreading of T cell autoimmunity in NOD mice has not been well-defined.
Notably, B cells appear to be necessary components contributing to the T1D process in NOD mice as few Igμ−/− NOD mice display spontaneous autoimmunity and develop diabetes (15, 16). Furthermore, B cells have been suggested to be the crucial APCs for the development of proinflammatory T cell responses to β cell Ags (17, 18, 19). Importantly, B cells have Ag-specific receptors, allowing them to concentrate Ag and present it much more efficiently to T cells than other APCs (20). Based on these properties, B cells may be the mediator of the determinant spreading of T cell autoimmunity to β cell Ags during the natural development of TID and following Ag-based immunotherapy. However, a recent report showed that dendritic cells educated by activated T cells could affect naive T cell responses to unrelated Ags (21). Apparently, educated dendritic cells may contribute to the determinant spreading of T cell immunity. Given that B cell-deficient NOD mice have dendritic cells, can they be experimentally primed to mount Th1 and Th2 responses to β cell Ags? Once primed experimentally, could dendritic cells as well as macrophages mediate the spreading of Th1 and Th2 responses to other β cell Ags in B cell-deficient NOD mice? As characterization of T cell responses in vitro relies on APC activity to present β cell Ags, how does APC activity, particularly that of B cells, affect the detection of spontaneously primed and experimentally induced T cell autoimmunity in vitro?
We studied the role of APCs in determinant spreading in two independent experimental models. In the first, we transfused NOD/scid mice with naive T and B cells isolated from 2- to 3-wk-old NOD mice to generate TB NOD/scid mice, or with naive T cells alone to generate B cell-deficient T NOD/scid mice. In the second, we reconstituted lethally irradiated Igμ−/− NOD mice with syngenic bone marrow cells and mature B cells isolated from 2- to 3-wk-old NOD mice to establish BMB Igμ−/− NOD mice, or with bone marrow cells alone to establish B cell-deficient BM Igμ−/− NOD mice. We then characterized their spontaneous and Ag-primed T cell immunity to a panel of Ags. We found that B cells were essential for the spreading of spontaneous Th1 autoimmunity as well as induced Th2 responses to β cell Ags. Furthermore, T NOD/scid and BM Igμ−/− NOD mice were fully competent to mount T cell responses to administered β cell Ags, but spreading to other β cell Ags was not detected, even in the presence of B cells in vitro. This suggests that other APCs in B cell-deficient NOD could not mediate determinant spreading of Th1 and Th2 responses to β cell Ags. Interestingly, we found that although macrophages and dendritic cells are capable of evoking β cell Ag-specific T cell responses, B cells can significantly enhance detection of T cell responses to β cell Ags in vitro. Our data from these two independent experimental models suggest that B cells, but not other APCs, mediate the determinant spreading of spontaneous Th1 responses during the T1D process in NOD mice and the induced Th2 spreading among β cell Ags following Ag-based immunotherapy.
Materials and Methods
Mice
NOD (H-2NOD), NOD/scid (Taconic Farms), NOR (The Jackson Laboratory), and Igμ−/− NOD mice (provided by Dr. D. V. Serreze, The Jackson Laboratory, Bar Harbor, ME) were bred under specific pathogen-free conditions. In our colonies, ∼85% of the female NOD mice, but not NOD/scid and Igμ−/− NOD mice, spontaneously develop T1D by 35–40 wk of age. Only female mice were used in our studies. All experimental procedures were approved by the University of California Los Angeles (UCLA) Chancellor’s Animal Research Committee.
Chemicals and Ags
Mouse GAD (65 kDa) and control β-galactosidase (β-gal) were prepared as previously described (5). HSP277 (11) was synthesized at >95% purity by Multiple Peptide Systems. Insulin B-chain was purchased from Sigma-Aldrich. The purified protein derivative (PPD) was prepared from Mycobacterium tuberculosis H37 Ra (Difco). The islet lysate was prepared by isolating the islets from NOD/scid mice as previously described (22) and homogenizing them in PBS.
Isolation of T and B lymphocytes
We isolated naive T cells from young NOD mice by negative selection. Splenic mononuclear cell suspensions were prepared from 2- to 3-wk-old NOD mice. After lysis of RBC with osmolar buffer, we purified naive T cells by eliminating other cell populations using two successive treatments with complement (Pel-Freez Biologicals) and anti-Br220, anti-Mac 1, and anti-CD11c Abs (BD Pharmingen). After washing twice with HL-1 medium, the remaining T cell-enriched populations were incubated with MACS microbeads (with anti-mouse CD25, CD69, and CD19; Miltenyi Biotec) and loaded on a MACS Separation LD column. To purify B cells, we depleted T cells, macrophages, and dendritic cells in a similar fashion using anti-CD3, anti-Mac 1, and anti-CD11c Abs for cytotoxicity and MACS microbeads (with anti-mouse CD-3, CD-11c, and MAC-1) followed by a magnetic separation column. The naive T or B cell populations in the flow-through were collected and characterized by FACS analysis using fluorescently labeled anti-CD3, anti-CD4, anti-CD8, anti-CD19, anti-Br220, anti-MAC-1, anti-MHC class II, and anti-CD11c. The purified naive T or B cells (>95% purity as determined by FACS analysis) were used for transfusion or as APCs (the purified B cells) in vitro.
Establishment of mouse models
NOD/scid mice at 3 wk of age were transfused with 50 million naive T cells or 50 million naive T cells and 30 million B cells, generating T NOD/scid or TB NOD/scid mice, respectively. As a second model, we isolated bone marrow cells from 3- to 4-wk-old Igμ−/− NOD mice and removed mature B and T cells by Ab/complement-mediated cytotoxicity. We then transfused 5 million bone marrow cells together with 30 million B cells (isolated from 2- to 3-wk-old NOD mice) into lethally irradiated (1000 rad) 4-wk-old Igμ−/− NOD mice to establish BMB Igμ−/− NOD mice. We also reconstituted lethally irradiated Igμ−/− NOD mice with syngenic bone marrow cells, but not mature NOD B cells, to generate BM Igμ−/− NOD mice.
Immunizations
To prime Th1 responses, groups of T NOD/scid or BM Igμ−/− NOD mice were immunized s.c. with 100 μg of GAD, HSP277, insulin B-chain, or β-gal in 50% CFA at the base of the tail 3 days posttransfusion or 6 wk postreconstitution, respectively. To induce Th2 responses, groups of TB or T NOD/scid mice were i.p. immunized with 100 μg of GAD, HSP277, insulin B-chain, or β-gal in 50% IFA 3 days posttransfusion. Similarly, groups of BM Igμ−/− or BMB Igμ−/− NOD mice were also immunized with one of those Ags in IFA 6 wk postreconstitution. Ten days later, the mice were boosted with the same Ag in IFA.
ELISPOT analysis
Spontaneously primed and experimentally induced T cell responses were characterized using the modified ELISPOT assays as previously described (7). Briefly, 106 splenic mononuclear cells were added per well (in duplicate) of an ELISPOT plate that had been coated with cytokine capture Abs and incubated with Ag (100 μg/ml for whole proteins, 7 μM for peptides) for 24 h for IFN-γ, 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 Cytomation) and 3-amino-9-ethylcarbazole. 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-γ, IL-4, and IL-5, respectively.
To determine whether the Ag-presenting activities of B cells affect the detection of effector and memory T cell responses in the ELISPOT assays, we mixed the purified T cells (4 × 105/well) from T or TB NOD/scid mice with 6 × 105/well T-depleted APCs (containing B cells) or T/B-depleted APCs (without B cells) from 2- to 3-wk-old NOD mice.
Assessment of diabetes
Groups of T or TB NOD/scid mice, BM Igμ−/−, or BMB Igμ−/− NOD mice were monitored twice weekly for the onset of diabetes by measuring their urine glucose levels by Clinistix-Strips (Bayer). After abnormal glucose levels in the urine were detected, the blood glucose levels were measured every other day using OneTouch Ultra strips (Lifescan). Two consecutive blood glucose levels ≥13 mM were considered as diabetes onset. T or TB NOD/scid mice were monitored up to 25 wk after transfusion. Igμ−/− BM or Igμ−/− BMB NOD mice were monitored up to 35 wk postreconstitution.
Statistical analysis
Group comparisons were analyzed by using the Student t test. Repeated measurements over time were assessed by Life Table analysis. Statistical significance was set at p < 0.05.
Results
B cells are crucial for the spreading of spontaneous Th1 autoimmunity among β cell Ags
B cells are professional APCs and can effectively present Ag to Ag-specific T cells. However, the role of B cells in the development of spontaneous T cell responses to β cell Ags remains in debate. Studies of Igμ−/− NOD mice have shown that they fail to develop spontaneous T cell autoimmunity (16, 17, 18). In contrast, a recent report showed that spontaneous T cell autoimmunity to some β cell Ags could be detected in NOD.μMT−/− mice, suggesting that B cells are not absolutely required for the development of spontaneous T cell autoimmunity (23). Furthermore, which APCs mediate the spreading of spontaneous Th1 responses among β cell Ags is poorly understood.
To examine the role of B cells in the determinant spreading of spontaneous T cell autoimmunity among β cell Ags, we established two independent experimental models. In the first model, we purified naive T and B cells from 2- to 3-wk-old NOD mice and transfused naive T cells with, or without, purified B cells to NOD/scid mice, generating TB NOD/scid mice or T NOD/scid mice, respectively. The second model was generated by reconstituting Igμ−/− NOD mice with syngenic bone marrow cells alone (BM Igμ−/− NOD mice) or with bone marrow cells and mature B cells isolated from 2- to 3-wk-old NOD mice (BMB Igμ−/− NOD mice). Following transfusion, we characterized splenic T cell autoimmunity to a panel of β cell Ags using ELISPOT assays. We found that 4 wk posttransfusion, TB NOD/scid mice developed weak Th1-biased responses to GAD, but not to HSP277 or insulin B-chain (Fig. 1⇓A). At 10 wk posttransfusion, we detected stronger Th1 responses to GAD, and found that spontaneous Th1 responses had spread to HSP277 and insulin B-chain (Fig. 1⇓B). Hence, following transfusion of naive T and B cells to NOD/scid mice, the spreading hierarchy of spontaneous Th1 autoimmunity is similar to that in unmanipulated NOD mice (7). In contrast, there was no detectable T cell response to any of the tested β cell Ags in T NOD/scid mice even at 10 wk posttransfusion.
TB NOD/scid and BMB Igμ−/−, but not B cell-deficient, NOD mice develop spontaneous Th1 autoimmunity. Splenic mononuclear cells were isolated from TB NOD/scid or T NOD/scid at 4 and 10 wk posttransfusion, or from BMB or BM Igμ−/− NOD mice at 6 or 12 wk postconstitution as well as unmanipulated Igμ−/− NOD mice at 6 or 12 wk of age. Their spontaneous T cell autoimmunity to a panel of β cell and control Ags were characterized by the ELISPOT assays at 4–6 wk (A) or 10–12 wk (B) postreconstitution to determine the frequency of β cell Ag-specific IFN-γ-, IL-4-, and IL-5-secreting T cells. Data are presented as the mean number of spot-forming colonies (SFC) per 106 splenic cells ± SEM. Mice from B cell-deficient or B cell-potent groups were tested simultaneously in two independent experiments (n = 4–6 for each group). All of the mice had undetectable levels of IL-4- and IL-5-secreting T cell responses to all of the tested Ags and had a similar frequency of T cells responding to anti-CD3 (data not shown). There were no detectable T cell responses to control Ag (β-gal, data not shown).
Similarly, we detected low frequency of GAD-reactive IFN-γ-secreting T cells in BMB Igμ−/− NOD mice, but not in BM Igμ−/− NOD mice, 6 wk after reconstitution (Fig. 1⇑A). At 12 wk postreconstitution, the frequency of GAD-reactive IFN-γ-secreting T cells increased in BMB Igμ−/− NOD mice (Fig. 1⇑B). We also detected HSP277- and insulin B-chain-reactive IFN-γ-secreting T cells, suggesting that the early spontaneous Th1 responses spread to other β cell Ags in vivo. However, we did not detect any Th1 responses to the tested β cell Ags in BM Igμ−/− NOD mice as well as unmanipulated Igμ−/− NOD mice, even though all strains of mice developed similar levels of Th1 responses after anti-CD3 stimulation (data not shown). Collectively, these data indicate that B cells are crucial for the determinant spreading of spontaneous Th1 autoimmunity among β cell Ags in NOD mice.
Immunization can prime a β cell Ag-specific Th1 response, which does not spread to other β cell Ags in B cell-deficient NOD mice
Next, we examined whether β cell Ag-specific Th1 responses could be induced in B cell-deficient NOD mice. Immunization with Ag in CFA is known to induce Th1-polarized responses to the injected Ag (24). T NOD/scid mice and BM Igμ−/− NOD mice were immunized with a β cell or control Ag in CFA 3 days posttransfusion or 6 wk postreconstitution, respectively. Their splenic T cell responses were characterized using ELISPOT assays 6 wk after immunization (Fig. 2⇓). As expected, we did not detect any IL-4- or IL-5-secreting Ag-specific T cells in either strains of B cell-deficient mice (data not shown), consistent with previous observations (24). Second, the frequency of IFN-γ-secreting PPD-reactive T cells was indistinguishable among the tested groups of B cell-deficient mice, and induced Th1 responses to β-gal were similar to that in immunocompetent mice (7), suggesting that B cells are not necessary for priming strong Th1 responses to a foreign Ag. Third, immunization with a β cell Ag (HSP277 or insulin B-chain) primed robust Th1 responses to the Ag, indicating that macrophages and dendritic cells are sufficient to mount Th1 autoimmunity to these Ags. Interestingly, immunization with GAD in CFA only induced weak Th1 responses, implying that B cells are important for the induction of strong Th1 responses to GAD.
Induced Th1 responses to β cell Ags in B cell-deficient NOD mice. Groups of T NOD/scid (3 days posttransfusion) and BM Igμ−/− NOD mice (6 wk postreconstitution) were immunized with the indicated Ag in CFA, and 6 wk later, their splenic T cell responses to a panel of Ags were characterized using the ELISPOT assays in T NOD/scid mice (A) or BM Igμ−/− NOD mice (B). Data are presented as the mean number of IFN-γ SFC per 106 splenic cells ± SEM. Mice from each group (n = 4–5) were tested simultaneously in two independent experiments. There were no detectable IL-4- and IL-5-secreting T cells to all of the tested Ags (data not shown).
We then evaluated whether experimentally primed Th1 responses to a β cell Ag could spread to other β cell Ags in B cell-deficient NOD mice. We found that the Th1 responses induced by immunization with a single β cell Ag in CFA failed to spread to other β cell Ags (Fig. 2⇑). For example, immunization with HSP277 primed Th1 responses only to HSP277, but not to GAD or insulin B-chain, while injection with insulin B-chain induced Th1 responses only to insulin B-chain in B cell-deficient mice. Thus, even after the induction of a vigorous β cell Ag-reactive Th1 response, proinflammatory T cell autoimmunity was not transmitted to other naive β cell Ag-reactive T cells and did not induce T1D in B cell-deficient NOD mice (data not shown). Apparently, cytokines produced by the primed first wave of T cells and T-dendritic/macrophage interactions are insufficient for mediating infectious β cell Ag-specific T cell autoimmunity in the absence of B cells. Interestingly, we found that B-potent NOR mice treated with single Ag in CFA developed unipolar Th1 responses to the injected Ag, but not to other Ags (data not shown), as observed in BALB/c mice (7). These data suggest that B cells within the stronger regulatory environment in NOR mice are incapable of mediating determinant spreading of T cell autoimmunity (25).
Macrophages and dendritic cells can evoke, and B cells can significantly enhance, detection of Th1 autoimmunity in vitro
As APC activity is critical for efficiently detecting T cell responses by T cell proliferation and ELISPOT assays, we tested how the Ag-presenting activity of APCs, particularly B cells, affects the detection of autoreactive T cells in vitro. We purified T cells from T NOD/scid mice or TB NOD/scid mice at 10 wk posttransfusion of T and B cells and mixed them with T-depleted APCs (containing B cells) or T/B-depleted APCs (without B cells) isolated from 2- to 3-wk-old NOD mice. We then characterized their spontaneous T cell autoimmunity to β cell Ags by ELISPOT assays (Fig. 3⇓A). In the absence of B cells in vitro, we failed to detect any IFN-γ-secreting Th1 cells responding to the tested β cell Ags in T NOD/scid mice. In the presence of B cells, we did detect low frequency of IFN-γ-secreting T cells (with tiny, but clear spots) from some T NOD/scid mice in response to the islet lysate, but not to the tested β cell autoantigens. This suggests that there is spontaneous priming of some islet-specific Th1 cells, but detection of these T cell responses by in vitro functional assays requires B cells. Furthermore, splenic T cells from TB NOD/scid mice responded to all of the tested β cell Ags in the presence of T/B-depleted APCs, although the frequency was significantly lower than that when cultured with T-depleted APCs (containing B cells). GAD-specific T cell responses, in particular, were enhanced by addition of B cells. These data indicate that determinant spreading of spontaneous Th1 autoimmunity had occurred in these mice. The appearance of detectable Th1 responses to the islet lysate in T NOD/scid mice and increased Th1 responses to β cell Ags in TB NOD mice in the culture with B cells suggest that APC function of the B cell is required for effectively recalling memory and effector Th1 responses to β cell Ags in vitro.
B cells significantly enhance detection of memory and effector Th1 responses to β cell Ags in vitro. A, Effects of B cells on detecting spontaneous Th1 responses. Splenic T cells (4 × 105/well) were purified from TB or T NOD/scid mice and mixed with T-depleted (+B) or T/B-depleted (−B) splenic APCs (6 × 105) in ELISPOT assays in the presence of the indicated Ag. B, Effects of B cells on detecting primed Th1 responses. T NOD/scid mice were immunized with the indicated β cell Ag in CFA, and 10 wk later, their splenic T cells were purified and mixed with T-depleted or T/B-depleted APCs in ELISPOT assays in the presence of the indicated Ag. Data are presented as the mean number of IFN-γ SFC per 4 × 105 splenic T cells ± SEM from three independent experiments. There were no detectable IL-4 and IL-5 responses, and all mice displayed similar levels of Th1 responses to anti-CD3 (data not shown).
Given the observed deficiencies in functional in vitro assays in detecting activated T cells in the absence of B cells, we reassessed whether addition of B cells in vitro could allow the detection of induced Th1 spreading to other β cell Ags. We immunized T NOD/scid with a β cell Ag in CFA. Ten weeks later, their splenic T cells were purified and their responses to a panel of β cell Ags were characterized using ELISPOT assays in the presence of T-depleted or T/B-depleted APCs (Fig. 3⇑B). We observed that T cell responses to PPD were similar in all groups, indicating that macrophages and dendritic cells are capable of evoking Th1 responses to this foreign Ag in vitro. In the presence of B cells, we detected significantly increased Th1 responses to the injected β cell Ag (p < 0.035 for GAD, p < 0.025 for both HSP277 and insulin B chain), but we failed to detect Th1 responses to other uninjected β cell Ags, indicating the lack of Th1 spreading in B cell-deficient mice. Together, these observations suggest that B cells, but not other professional APCs, mediate determinant spreading of Th1 responses in vivo.
B cells are important for the development of spontaneous T1D
To further study B cell involvement in T1D pathogenesis, we monitored diabetes onset in our B cell-potent or B cell-deficient NOD mouse models. Irradiated Igμ−/− NOD mice that had been reconstituted with syngenic bone marrow cells, or combined with mature B cells from young NOD mice, were monitored for diabetes development up to 35 wk postreconstitution. BM Igμ−/− NOD mice that did not receive B cells remained euglycemic, while BMB Igμ−/− NOD mice that received B cells developed diabetes with a rate similar to that of unmanipulated NOD mice (Fig. 4⇓). Following reconstitution, BMB Igμ−/− NOD mice began to develop diabetes ∼15 wk postreconstitution and ∼83% mice became diabetic by 28 wk postreconstitution. Furthermore, none of T NOD/scid mice developed diabetes over a 25 wk posttransfusion observation period. In contrast, TB NOD/scid mice developed diabetes in an accelerated fashion. They began to develop diabetes ∼11 wk posttransfusion and ∼92% of the mice were hyperglycemic by 20 wk of age. The association of B cells with diabetes development in our NOD models complements and extends previous studies (15, 16, 18), indicating that B cells are important for the T1D pathogenesis.
TB NOD/scid and BMB Igμ−/−, but not B cell-deficient, NOD mice spontaneously develop T1D. Female NOD/scid mice, which had been transfused with T and B cells or T cells alone, were monitored for the development of T1D up to 25 wk posttransfusion. Irradiated Igμ−/− NOD mice that had been reconstituted with syngenic bone marrow cells alone or bone marrow cells and mature B cells from young NOD mice. They were monitored for the T1D onset up to 35 wk postreconstitution. Two consecutive blood glucose levels of >13 mM were considered as disease onset (n = 16 for each group).
B cells are crucial for the spreading of inducible Th2 autoimmunity
Our previous studies, and those of others, have shown that following autoantigen-based immunotherapy, inducible Th2 responses can spread to other unrelated target-tissue autoantigens, which is associated with the inhibition of disease progression (7, 26, 27). A recent report showed that dendritic cells educated by Th2 cytokines could affect the naive T cell responses to unrelated Ags (21), which may contribute to infectious Th2 autoimmunity. To evaluate which population of APCs contributes to the determinant spreading of Th2 responses, we transfused NOD/scid mice with T, or T and B cells, and 3 days later, immunized them with a β cell Ag in IFA. Ten days later, the mice were reinjected with the same Ag, and 10 wk posttransfusion, their splenic T cell responses to a panel of β cell and control Ags were characterized by ELISPOT assays (Fig. 5⇓). Both B cell-deficient and B cell-potent mice developed similar levels of Th2 responses to β-gal, which failed to spread to any β cell Ags. Theses data suggest that macrophages and dendritic cells are sufficient to mount Th2 responses to this foreign Ag. Notably, T NOD/scid mice that had been immunized with a β cell Ag developed Th2 responses only to the injected Ag (Fig. 5⇓A). For example, HSP277-immunized mice displayed Th2 responses to HSP277, but not to insulin B-chain and vice versa. Apparently, educated dendritic cells and macrophages are incapable of mediating the spreading of Th2 autoimmunity in B cell-deficient NOD mice. In contrast, TB NOD/scid mice that had been immunized with a β cell Ag developed Th2 autoimmunity not only to the injected Ag, but also to other unrelated β cell Ags (Fig. 5⇓C). For example, the mice treated with GAD in IFA displayed vigorous Th2 responses to GAD as well as to HSP277 and insulin B-chain. Similarly, mice immunized with HSP277 displayed IL-5-secreting T cells responding to HSP277, GAD, and insulin B-chain. In parallel experiments, we observed that while Th2 responses induced by immunization with a β cell Ag were limited to the injected Ag in BM Igμ−/− mice, induced Th2 responses spread to other β cell Ags in BMB Igμ−/− mice (Fig. 5⇓, B and D). In addition, immunization of B-potent NOR mice with single Ag in IFA induced Th2 responses only to the injected Ag, which did not spread to other Ags (data not shown). Together, our data suggest that macrophages and dendritic cells are sufficient for priming Th2 responses to β cell Ags, and B cells in NOD mice are unique for the spreading of induced Th2 autoimmunity among β cell Ags.
Immunization with a β cell Ag promotes Th2 spreading in TB NOD/scid and BMB Igμ−/−, but not in B cell-deficient, NOD mice. Groups of T NOD/scid (A), TB NOD/scid (C), BM Igμ−/− (B), or BMB Igμ−/− NOD mice (D) were treated with a β cell or control Ag in IFA. Their splenic T cell responses to a panel of Ags were characterized using ELISPOT assays. Data shown are the mean number of IL-5 SFC per million splenic T cells ± SEM from four to six mice in two independent experiments. A similar pattern of IL-4 responses was observed (data not shown). All mice showed a similar level of T cell responses to anti-CD3 (data not shown).
Discussion
Previous studies have found that during T1D development in NOD mice, spontaneous splenic Th1 responses spread among β cell autoantigens, and are associated with disease progression (5, 6, 10). An understanding of this infectious T cell autoimmunity may be useful for the development of new approaches to control autoimmune diseases. In the present study, we systemically evaluated the role of B cells in spontaneous Th1 spreading among β cell Ags in two independent T1D mouse models. We found that TB NOD/scid and BMB Igμ−/− NOD mice, which have mature B cells, developed spontaneous Th1 responses to GAD followed by Th1 responses to HSP277 and insulin B-chain. These mice had a high incidence of diabetes, similar to that of NOD mice. In contrast, none of the B cell-deficient NOD mice (T NOD/scid and BM Igμ−/− mice as well as Igμ−/− NOD mice) displayed spontaneous Th1 responses to the tested β cell Ags nor became hyperglycemic even if they had a low level of infiltration in their islets (data not shown). The lack of spontaneous Th1 autoimmunity in B cell-deficient mice was not caused by deficiencies in β cell Ag-specific T cells as these mice could mount vigorous Ag-specific Th1 responses after immunization with a β cell Ag. However, the β cell Ag-primed Th1 responses never spread to unrelated β cell Ags and did not induce T1D in B cell-deficient NOD mice. These observations, together with the association of B cells with spontaneous spreading of Th1 autoimmunity among β cell Ags, strongly suggest that B cells mediate infectious Th1 autoimmunity, contributing to T1D development in NOD mice. Given that blocking determinant spreading inhibits disease progression in the experimental autoimmune encephalomyelitis model (28), the modulation of B cells in ways that reduce their capacity to promote determinant spreading may provide novel interventive therapies for T1D in humans. Notably, T cells from newly diabetic NOD mice and T cell clones recognizing β cell Ags can adoptively transfer T1D to NOD/scid mice (29, 30, 31). Accordingly, once the cascade of diabetogenic T cell autoimmunity has been established, B cells are not required for effector T cells to destroy β cells in the pancreatic islets.
Autoantigen-based immunotherapies which induce regulatory T cell responses can effectively inhibit organ-specific autoimmune diseases in animal models (22, 32, 33, 34, 35). We observed that autoantigen-primed Th2 responses rapidly spread to unrelated target-tissue Ags (9). Apparently, the first wave of Ag-primed Th2 responses promotes the development of Th2 responses to other β cell Ags, leading to infectious Th2 autoimmunity and inhibition of the pathogenic process (7, 24). This spreading of Th2 autoimmunity both increases bystander suppression and helps exhaust Ag-specific T cells that could give rise to pathogenic T cell responses (9). Recent studies have shown that activated T cells can educate dendritic cells which in turn influence the functional development of unrelated naive T cells toward a similar phenotype (21). Accordingly, educated dendritic cells may contribute to the process of Th2 spreading. In our studies of B cell-deficient NOD mice, we observed that although vigorous Th2 responses to a β cell Ag could be induced, the induced Th2 responses never spread to other β cell Ags. Hence, dendritic cells educated by β cell Ag-primed Th2 cells and macrophages were incapable of mediating the spreading of Th2 responses to other β cell Ags. In B cell-potent TB NOD mice and Igμ−/− BMB NOD mice, β cell Ag-primed Th2 responses spread to other uninjected β cell Ags. Thus, B cells are crucial for transmitting regulatory T cell responses following Ag-based immunotherapy. Accordingly, the manipulation of the APC capacity of B cells may enhance the efficacy of Ag-based immunotherapies in inhibiting organ-specific autoimmune diseases.
We examined how the Ag-presenting activities of B cells affect the detection of T cell responses in vitro. With T/B-depleted APCs, we detected spontaneous Th1 autoimmunity to multiple β cell Ags in B cell-potent NOD mice, as well as induced Th1 responses to β cell Ags in B cell-deficient NOD mice. Apparently, once T cell responses to β cell Ags have been induced in vivo, macrophages and dendritic cells can elicit memory and activated T cells responding to β cell Ags in vitro. Addition of B cells to the ELISPOT cultures significantly increased the detected frequency of β cell Ag-specific Th1 cells in B cell-potent NOD mice, and enhanced detection of T cell responses to an immunogen in B cell-deficient NOD mice, consistent with a recent report (36). Interestingly, in the presence of B cells in vitro, T cells from a few of B cell-deficient NOD mice displayed weak T cell responses to the islet lysate, which may reflect the ability of macrophages and dendritic cells to activate naive T cells (37). Thus, in the absence of mature B cells, dendritic cells as well as macrophages may initiate T cell autoimmunity to β cell Ags, while B cells appear to be important for expansion and diversification of the autoimmune cascade, mediating determinant spreading. It will be of interest to further determine the role each type of APC plays in β cell Ag-specific T cell autoimmunity and T1D development.
Homeostatic proliferation has been demonstrated to promote self-reactive T cell expansion, contributing to the initiation of autoimmunity (38, 39, 40, 41). We found that there is no significant difference in the total number of T cells, and their CD4 and CD8 population between B-potent and B-deficient NOD mice 4 wk after transferring T or T/B cells, or 6 wk after bone marrow constitution (data not shown). Importantly, the frequency of spontaneously primed Th1 responses in TB NOD/scid, BMB Igμ−/− NOD mice and induced Th1 responses in T NOD/scid mice and BM Igμ−/− NOD mice is proportionally similar to those NOD mice. These data suggest that naive T cells in NOD/scid mice underwent homeostatic proliferation, independent of B cell activity. Apparently, homeostatic proliferation of naive T cells did not preferentially increase (or diminish) the frequency of any particular β cell Ag-reactive T cell pool. We also noticed that following transfusion and reconstitution, proportional levels of B cells existed long-term in the recipients, suggesting that B cells also underwent homeostatic proliferation in the recipients. Indeed, IL-21R is expressed by B cells and IL-21 can potently stimulate B cell proliferation (42). Thus, the expansion and homeostatic proliferation of T and B cells may contribute to B cell-mediated determinant spreading of T cell autoimmunity.
The spreading of T cell autoimmunity has been thought to be due to cytokines, such as IFN-γ or IL-4, released by the first wave of autoreactive T cells, creating a microenvironment in the target tissue which educates APCs, and promotes subsequent waves of autoreactive T cells to development toward a similar phenotype (7, 13, 14, 43). We found that although vigorous Th1 or Th2 responses to β cell Ags could be induced in B cell-deficient NOD mice as well as in B-potent NOR mice, the spontaneous Th1 and induced Th2 responses could spread only in B cell-potent NOD mice. Apparently, B cells, but not other professional APCs or cytokines produced by the first wave of T cells, can effectively mediate the determinant spreading of T cell autoimmunity in NOD mice.
How do B cells mediate this spreading of T cell autoimmunity? Notably, B cells in NOD mice are deficient in tolerance to autoantigens, respond strongly to BCR stimulation by proliferation with relative resistance to activation-induced cell death, and have an unique pattern of B7 expression, which contribute to proinflammatory responses in the islets (44, 45). Furthermore, besides capturing and concentrating Ag for presentation, activated B cells express high levels of MHC II, B7, CD40, and other second signal-associated molecules, making them much more efficient presenters of Ag than other APCs (20). It is likely that the first wave of activated T cells promote B cell activation, through both secreted cytokines and T:B interaction in Ag-specific and/or bystander fashions, enhancing the APC function of the B cells. Subsequently, activated B cells, which should have diverse Ag receptors, effectively capture, and present different β cell Ags to T cells, mediating the spreading of T cell autoimmunity during the T1D pathogenesis, and following Ag-based immunotherapy.
In summary, we have shown that B cell-potent TB NOD/scid and BMB Igμ−/− NOD mice, but not B cell-deficient T NOD/scid and BM Igμ−/− NOD mice, developed spontaneous Th1 autoimmunity, which spread among β cell Ags, and were associated with a high incidence of T1D. Treatment of B cell-potent NOD mice with a β cell Ag promoted Th2 responses, which spread to unrelated β cell Ags. Although treatment with a β cell Ag could induce Th1 and Th2 responses to the injected Ag in B cell-deficient NOD mice, these T cell responses failed to spread to other β cell Ags. Finally, B cells as professional APCs in vitro can significantly enhance detection of Th1 responses that have been primed in vivo. Hence, our data suggest that B cells mediate the spreading of spontaneous Th1 autoimmunity during the T1D development and the spreading of induced Th2 autoimmunity following Ag-based immunotherapy in NOD mice. Our findings provide new insights into the mechanism(s) underlying the process of determinant spreading of T cell autoimmunity. Given that interference with determinant spreading has resulted in the inhibition of disease progression in the experimental autoimmune encephalomyelitis model (28), the modulation of the capacity of B cells to present Ag may provide new interventions for enhancing Ag-based immunotherapy and controlling autoimmune diseases.
Disclosures
The authors have no financial conflict of interest.
Footnotes
-
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 Institutes of Health, the Juvenile Diabetes Research Foundation International, and the Diabetes Action Research and Education Foundation.
-
↵2 Address correspondence and reprint requests to Dr. Jide Tian, Department of Molecular and Medical Pharmacology, David Geffen School of Medicine, University of California, Box 173517, 10833 Le Conte Avenue, Los Angeles, CA 90095-1735. E-mail address: jtian{at}mednet.ucla.edu
-
↵3 Abbreviations used in this paper: T1D, type 1 diabetes; GAD, glutamic acid decarboxylase; HSP, heat shock protein; β-gal, β-galactosidase; PPD, purified protein derivative; SFC, spot-forming cell.
- Received May 10, 2005.
- Accepted December 1, 2005.
- Copyright © 2006 by The American Association of Immunologists
References
In this issue
