Critical role for thymic CD19⁺CD5⁺CD1d^hiIL-10⁺ regulatory B cells in immune homeostasis

Ethics Committee approval

Care, use, and treatment of mice in this study were in strict agreement with international guidelines for the care and use of laboratory animals. This study was approved by the Animal Ethics Committee of the Beijing Institute of Basic Medical Sciences.

Mice

Seven- to 9-week-old C57BL/6, CD19-Cre mice and lupus-like NZB/NZW F1 mice (Chinese Academy of Medical Sciences, Beijing, China) were bred in our animal facilities under specific pathogen-free conditions.

Cytometric analysis and intracellular cytokine staining

All cell experiments were strictly prepared on ice, unless stated otherwise in other specific procedures. Cells (1 × 10⁶ cells/sample) were washed with FACS staining buffer (PBS, 2% FBS or 1% BSA, 0.1% sodium azide). All samples were incubated with anti-FcR antibody (clone 2.4G2; BD Biosciences, San Jose, CA, USA) before incubation with other antibodies diluted in FACS buffer, supplemented with 2% anti-FcR antibody. For intracellular cytokine staining, 50 ng/ml PMA and 1 mg/ml ionomycin (Sigma-Aldrich, St. Louis, MO, USA) were added, and then, 1 mg/ml brefeldin A and 2 mM monensin were added 3 h later. After 3 h, cells were collected and fixed for 20 min with 1 ml fixation buffer (Intracellular Fixation & Permeabilization Buffer Kit; eBioscience, San Diego, CA, USA). After washing, the fixed cells were stained. The samples were filtered immediately before analysis or cell sorting to remove any clumps. The following antibodies were purchased from eBioscience: anti-mouse CD3 (clone 145-2C11), CD4 (clone GK1.5), B220 (clone RA3-6B2), CD19 (clone MB19-1), CD5 (clone 53-7.3), CD1d (clone 1B1), IL-10 (clone JES5-16E3), and Foxp3 (clone NRRF-30). Anti-mouse CD72 antibody (clone K10.6) were purchased from BD PharMingen (San Diego, CA, USA). Data collection and analyses were performed on a FACSCalibur flow cytometer by use of CellQuest software (BD Biosciences).

Cell sorting

Approximately 6 × 10⁶ lymphocytes were separated by Ficoll solution from thymus of a 7- to 9-week-old female or male C57BL/6 mouse and used to sort 5 × 10⁴ CD3⁻CD4⁻B220⁺CD19⁺CD5⁺CD1d^hi B_regs and 2 × 10⁵ CD3⁻CD4⁻B220⁺CD19⁺ CD5⁻CD1d^lo B cells. Multicolor flow cytometry was performed by gating on CD3^-CD4⁻B220⁺CD19⁺ cells that were CD5⁺CD1d^hi or CD5⁻CD1d^lo. The lymphocytes from the spleen and mesenteric and inguinal LNs of 7-week-old female or male C57BL/6 mice were used to sort-purify CD4⁺CD25⁺ T_regs and CD4⁺CD25⁻ T cells. Multicolor flow cytometry was performed by gating on CD3⁺CD4⁺B220⁻ cells that were CD25⁺CD62L⁺ (CD4⁺CD25⁺ T_regs) or CD25⁻CD62L⁺ (CD4⁺CD25⁻ T cells). CD19⁺ B cells and CD4⁺ T cells were separated by CD19 microbeads and CD4 microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany), respectively, from the spleen and mesenteric and inguinal LNs of a 7- to 9-week-old female or male C57BL/6 mice. All flow cytometry data were acquired with FACSCanto, FACSCanto II, or FACSAria (BD Biosciences), gated on live lymphocyte-sized cells on the basis of forward- and side-scatter and analyzed by use of FlowJo software (Tree Star, Ashland, OR, USA).

In vitro B–T cell coculture

CD4⁺ T cells were labeled with CFSE (Molecular Probes, Eugene, OR, USA). CFSE was added to a final concentration of 5 μmol/L and the cells incubated for 10 min at 37°C. The T cells were washed 3 times through FCS and resuspended in complete medium. CD4⁺ T cells, at a concentration of 1 × 10⁶ cells/ml, were cocultured for 3 d in 96-well plates with 1 × 10⁶ cells/ml CD5⁻CD1d^lo or CD5⁺CD1d^hi cells from thymus in RPMI-1640 medium containing 10% FBS, 2 mM glutamine, penicillin (100 IU/ml), streptomycin (100 µg/ml), and 50 mM 2-ME. CD19⁺ B cells (1 × 10⁶ cells/ml) were cocultured for 3 d in 96-well plates with 1 × 10⁶ cells/ml CD4⁺CD25⁺ T_regs or CD4⁺CD25⁻ T cells in cultured in RPMI-1640 medium containing 10% FBS, 2 mM glutamine, penicillin (100 IU/ml), streptomycin (100 µg/ml), and 50 mM 2-ME with 10 µg/ml LPS (Sigma-Aldrich). In some conditions, cocultured cells were treated with 10 µg/ml isotype control antibody or neutralizing anti-mouse CD72 antibody (ab25320; Abcam, Cambridge, MA, USA).

Thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs were transferred into CD19-Cre or lupus-like mice

CD3⁻CD4⁻B220⁺CD19⁺CD5⁺CD1d^hiB (CD5⁺CD1d^hi) cells were sorted by FACS from the thymus of 7- to 9-week-old female or male C57BL/6 mice. Each age- and sex-matched CD19^Cre C57BL/6 mouse (6 mice/group) was injected in the tail veil (i.v.) with 1 × 10⁵ thymic CD5⁺CD1d^hi B cells. CD19⁺ B cells were sorted by CD19 microbeads (Miltenyi Biotec) from the thymus of 7- to 9-week-old female or male C57BL/6 mice. Each age- and sex-matched, lupus-like NZB/NZW F1 mouse (6 mice/group) was injected in the tail veil (i.v.) with 1 × 10⁵ CD5⁺CD1d^hiB_regs or CD19⁺ B cells.

Cytokine analysis by ELISA

Anti-dsDNA IgG and anti-dsDNA IgG1 antibody titers were measured by ELISA kits (Shibayaji, Gunma, Japan). In brief, diluted sera were added in triplicate to the plate (dsDNA precoated; Shibayaji) for 1 h at 37°C. To detect nonspecific IgG and IgG1 titers, diluted sera were coated to the plate (no antigen precoated) overnight at 4°C and then blocked with 4% BSA for 1 h at 37°C. Then, after washing, biotin rat anti-mouse IgG, IgG1 antibody (4 μg/ml) antibodies were added to the plate and were incubated for another hour at 37°C. Thereafter, unbinding antibodies were washed off, followed by addition of avidin-HRP (1/1000 diluted). Plates were incubated for 1 h at 37°C. Finally, the color was developed by incubation with o-phenylenediamine. The OD was read at 450 nm with an ELISA reader (Bio-Rad Laboratories, Hercules, CA, USA). Standard curves were established to quantitate the amounts of the respective cytokines.

Statistics

Statistics were generated by use of t-test in GraphPad Prism (version 5.0; GraphPad Software, La Jolla, CA, USA), and values are represented as mean ± sem. Results were considered statistically significant at P < 0.05.

B220⁺CD19⁺CD5⁺CD1d^hiIL-10⁺ B_regs are present in the thymus

Thymic B cells have been proposed to play a critical role in T cell-negative selection [22, 23]. To this end, we determined whether there exists a population of B_regs in murine thymus. Flow cytometric (FACS) analysis of thymocytes isolated from 7- to 9-week-old female or male C57BL/6 mice revealed that CD19⁺ B cells represent ∼0.11% of total thymocytes (Supplemental Fig. 1). To enrich CD19⁺ B cells, lymphocytes were purified further by Ficoll separation solution from thymocytes. In these purified thymic lymphocytes, ∼5% cells were CD3⁻CD4⁻CD19⁺B220⁺ B cells (Fig. 1A). Furthermore, ∼17% of thymic B220⁺CD19⁺ B cells could be induced to express IL-10, and many more B220⁺CD19⁺ B cells could be induced to express IL-10 than could CD19⁻B220⁺ B cells (Fig. 1B). Notably, the percentage of IL-10-expressing B_regs among thymic B cells is much higher than that among splenic B_regs, which are found at low frequencies (1–5%) in naïve mice [9]. We also found that only 2.57% of splenic CD19⁺ B cells could be induced to express IL-10 (Fig. 1C). In addition, thymic B220⁺CD19⁺IL-10⁺ B_regs express CD5 and CD1d^hi (Fig. 1B). We also found that the percentage of IL-10-expressing B cells was much higher in the thymic B220⁺CD19⁺CD1d^hiCD5⁺ B cells than that in B220⁺CD19⁺CD1d^loCD5⁻ B cells (Fig. 1D). Together, these results demonstrate that thymic IL-10-expressing B_regs are found predominantly within the CD1d^hiCD5⁺ B cell subpopulation and are phenotypically similar to splenic IL-10-producing B_regs [6, 8, 9]. In addition, there are ∼6 × 10⁶ lymphocytes, 3 × 10⁵ B cells, and 5 × 10⁴ CD5⁺CD1d^hiB cells in the thymi of 7- to 9-week-old female or male C57BL/6 mice (Fig. 1E). Furthermore, images of CD1d^lowCD5⁻ B cells and CD1d^hiCD5⁺ B_regs under a microscope were shown (Fig. 1F).

B220⁺CD19⁺CD5⁺CD1d^hi B_regs suppress anti-CD3/CD28 antibody-stimulated CD4⁺ T cell proliferation in vitro and maintained Foxp3⁺CD4⁺ T_regs

To detect the suppressive function of thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs on CD4⁺ T cell proliferation, coculture experiments were conducted by use of CD3⁻CD4⁻B220⁺CD19⁺CD5⁺CD1d^hi B_regs (CD5⁺CD1d^hi) and CD3⁻CD4⁻B220⁺CD19⁺CD5⁻CD1d^lo B cells (CD5⁻CD1d^lo), sorted from mouse thymus by FACS (purification >95%), and CD4⁺ T cells, sorted from mouse splenocytes by CD4 microbeads. CD5⁺CD1d^hi or CD5⁻CD1d^lo cells were cocultured for 3 d with CD4⁺ T cells in the presence of anti-CD3 antibody (precoated on the plate) and anti-CD28 antibody. The results suggest that thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs significantly suppress CD4⁺ T cell proliferation induced by anti-CD3 and anti-CD28 antibody stimulation (Fig. 2A and B).

Only transfer of splenic, IL-10-sufficient MZ and T2-MZP B cells, but not MZ or follicular B cells, can restore T_reg numbers through maintenance of Foxp3 expression in B cell-deficient (μMT) mice [12]. To examine the regulatory function of thymic CD19⁺B220⁺CD1d^hiCD5⁺ B cells in maintenance of CD4⁺Foxp3⁺ T cells, CD3⁻CD4⁻B220⁺CD19⁺CD5⁺CD1d^hi B_regs (CD5⁺CD1d^hi) and CD3⁻CD4⁻B220⁺CD19⁺ CD5⁻CD1d^lo B (CD5⁻CD1d^lo) cells were sort purified from the thymi of 7- to 9-week-old female or male C57BL/6 mice and then cocultured for 3 d with CD4⁺ T cells. Compared with freshly isolated Foxp3⁺CD4⁺ T cells, the percentage of Foxp3⁺CD4⁺ T cells decreased in a time-dependent manner during culture (data not shown). As compared with thymic CD5⁻CD1d^lo B cells, coculture of CD4⁺ T cells with CD5⁺CD1d^hi B cells maintained a greater percentage and number of CD4⁺Foxp3⁺ T_regs (Fig. 2C–E). A previous study demonstrated an IL-10-independent regulatory role for B cells in suppressing autoimmunity through maintenance of T_regs via glucocorticoid-induced TNFR-related protein ligand [28]. Our previous study demonstrated that interaction of CD5 and CD72 is involved in T_reg and B_reg homeostasis [29]. To determine whether the interaction between CD5 and CD72 was also involved in thymic B_reg-mediated maintenance of T_regs, neutralizing anti-mouse CD72 antibodies were used to block the CD5–CD72 interaction. First, we test the functional impact of blocking CD72 with neutralizing anti-mouse CD72 antibodies on B cell proliferation induced by LPS. The data demonstrated that LPS-induced proliferation of B cells was suppressed significantly by neutralizing anti-mouse CD72 antibodies (Supplemental Fig. 2). Furthermore, the blockage of the CD5–CD72 interaction with neutralizing anti-mouse CD72 antibodies during coculture of thymic B_regs and T_regs demonstrated that the CD5–CD72 interaction is involved in the maintenance of Foxp3⁺CD4⁺ T_regs induced by thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs (Fig. 2F–H).

Defective CD19⁺CD5⁺ B cells reduced Foxp3⁺CD4⁺ T_regs and up-regulated CD8⁻CD4⁺ and CD8⁺CD4⁻ T cells in Cd19^Cre mice

To examine the physiologic function of thymic B_regs, we used B cells isolated from homozygous Cd19^Cre-expressing C57BL/6 CD19^−/− [30]. Previous studies demonstrated that deletion of CD19 had no deleterious effects on the generation of B cells in the BM, but there was a significant reduction in the number of B cells in peripheral lymphoid tissues, such as blood and spleen, and in the peritoneal cavity [31, 32]. In addition, CD19^−/− mice are essentially devoid of IL-10-expressing B220⁺ B_regs, which leads to exacerbated inflammation and disease symptoms during contact hypersensitivity and in the experimental autoimmune encephalomyelitis model of multiple sclerosis [6, 33]. Furthermore, our data showed that deletion of CD19 resulted in a significant reduction in the number of mature B cells (Supplemental Fig. 1A–C) and IL-10-producing B_regs (Fig. 3A–C) in the thymus.

To examine the ability of B_regs to induce the conversion of CD4⁺ T cells into T_regs, we analyzed CD4⁺Foxp3⁺ T_regs by flow cytometry. We found that Foxp3⁺CD4⁺ T_regs are decreased in spleen, LN, and BM of CD19^−/− mice (Fig. 3D–F). Our data also demonstrated that the reduced frequency of Foxp3⁺CD4⁺ T_regs was associated with an increased frequency of total CD4⁺ T cells (Fig. 3D), consistent with studies in which T_regs were shown to suppress the proliferation of autologous CD4⁺CD25⁻ T cells [34]. By enumerating CD8⁻CD4⁻, CD8⁺CD4⁺, CD8⁻CD4⁺, and CD8⁺CD4⁻ T cells, we found that greater CD8⁻CD4⁺ and CD8⁺CD4⁻ T cell differentiation is associated with reduced thymic T_regs in CD19^−/− mice (Fig. 3G and H). Furthermore, we found that Foxp3⁺CD4⁺CD8⁻ T_regs are decreased in the thymus of CD19^−/− mice (Fig. 3F, I, and J). Our data also demonstrated that the reduced frequency of Foxp3⁺CD4⁺ T_regs is associated with an increased frequency of CD8^-CD4⁺ and CD8⁺CD4⁻ T cells (Fig. 3G–J). Taken together, these results suggest that by reducing the population of thymic IL-10-expressing CD19⁺CD5⁺ B_regs, Foxp3⁺CD4⁺ T_regs are also reduced, resulting in an expansion of CD8⁻CD4⁺ and CD8⁺CD4⁻ T cells in the thymus.

Transferred thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs promote conversion of Foxp3⁺CD4⁺ T_regs and suppress CD4⁺ and CD8⁺ T cell differentiation

To elucidate further the role of thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs in inducing T_regs and suppressing CD8⁻CD4⁺ and CD8⁺CD4⁻ T cell expansion/differentiation, CD3⁻CD4⁻B220⁺CD19⁺CD5⁺CD1d^hi B (CD5⁺CD1d^hi) and CD3⁻CD4⁻B220⁺CD19⁺ CD5⁻CD1d^lo B (CD5⁻CD1d^lo) cells were sort purified from the thymi of 7- to 9-week-old female or male C57BL/6 mice and transferred i.v. into CD19^−/− mice. The results demonstrate that compared with CD19⁺CD5⁻CD1d^lo B cells, transferred thymic CD5⁺CD1d^hi B_regs promote the development of Foxp3⁺CD4⁺ T_regs in the thymus, spleen, and LN but not in BM of recipient mice (Fig. 4A and B). We found reduced CD8⁻CD4⁺ and CD8⁺CD4⁻ T cell differentiation in CD19^−/− recipients receiving thymic CD5⁺CD1d^hi B_regs compared with CD5⁻CD1d^lo-transferred control mice (Fig. 4C and D). We also observed fewer CD4⁺ and CD8⁺ T cells in the spleens and LNs of CD19^−/− mice that received thymic CD5⁺ CD1d^hi B_regs compared with controls (Supplemental Fig. 2). These results are consistent with the idea that thymic CD19⁺ CD5⁺ B_regs suppress CD8⁻CD4⁺ and CD8⁺CD4⁻ T cell expansion/differentiation by up-regulating Foxp3⁺CD4⁺ T_regs in the thymus and concur with prior studies showing that thymic T_regs are actively involved in maintaining the peripheral tolerance of autoreactive T cells that have escaped thymic selection [35–38].

Transferred thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs suppress autoimmune disease by promoting Foxp3⁺CD4⁺ T_reg differentiation in lupus-like mice

To determine whether thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs suppress autoimmune disease progression, CD3⁻CD4⁻B220⁺CD19⁺CD5⁺CD1d^hi B (CD5⁺CD1d^hi) cells were sort purified, and CD19⁺ B cells were enriched by CD19 microbeads from the thymi of 7- to 9-week-old female or male C57BL/6 mice. Thymic B220 ⁺CD19⁺CD5⁺CD1d^hi B cells or CD19⁺ B cells were then transferred into lupus-prone NZB/NZW F1 mice. Flow cytometric analyses revealed that transfer of thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs resulted in an up-regulation of CD4⁺Foxp3⁺ T_reg differentiation (Fig. 5) as well as IL-10⁺ B_reg expansion (Fig. 6A–C) in recipient NZB/NZW F1 mice.

To determine whether IL-10⁺ B_regs were maintained/induced by CD4⁺Foxp3⁺ T_regs, CD4⁺CD25⁺ T_regs or CD4⁺CD25⁻ T cells, sort purified from the spleens and LNs of 7-week-old female or male C57BL/6 mice, were cocultured for 3 d with CD19⁺ B cells, followed by flow cytometric analysis. These experiments showed that B220⁺IL-10⁺ B_regs increase when cocultured with CD4⁺CD25⁺ T_regs (Fig. 6D and E). Furthermore, the CD5–CD72 interaction is involved in the expansion/maintaining of B220⁺IL-10⁺ B_regs induced by Foxp3⁺CD4⁺ T_regs, as demonstrated by experiments in which CD19⁺ B cells were cocultured with CD4⁺CD25⁺ T_regs in the presence of isotype control antibody or neutralizing anti-mouse CD72 antibody (Fig. 6F and G). These results suggest that T_regs induce/maintain IL-10⁺ B_regs via the CD5–CD72 interaction.

Among the murine lupus models, NZB/NZW F1 mice are the most similar to the human lupus phenotype, including autoantibody production (especially anti-dsDNA antibody) [39]. Therefore, we next assessed the effects B_regs on antibody production in this murine lupus model by examining antigen-specific or nonspecific antibody titers by use of ELISA. Compared with WT mice, unmanipulated, lupus-prone mice had higher levels of nonspecific IgG antibodies (Fig. 7A), nonspecific IgG1 antibodies (Fig. 7B), anti-dsDNA IgG antibodies (Fig. 7C), and anti-dsDNA IgG1 antibodies (Fig. 7D), as expected. Compared with untreated/control mice and mice receiving thymic CD19⁺ B cells, lupus-prone mice that received thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs had lower titers of nonspecific IgG antibodies (Fig. 7A), nonspecific IgG1 antibodies (Fig. 7B), anti-dsDNA IgG antibodies (Fig. 7C), and anti-dsDNA IgG1 antibodies (Fig. 7D). Histology images of untreated mice with lupus (not treated with CD19⁺ B cells or CD1d^hiCD5⁺ B_regs) or those of mice that received CD19⁺ B cells demonstrated shrinkage of the balloon cavity and appearance of excessive inflammatory cells in the surrounding interstitial vasculature. In contrast, CD1d^hiCD5⁺ B_reg-treated mice had fewer infiltrating inflammatory cells and an overall normal structure of the glomerular region (Fig. 7E). These results indicate that transferred thymic B220⁺CD19⁺CD5⁺CD1d^hi B_regs suppress autoimmune disease in recipient mice. Collectively, our results suggest that thymic B_regs may suppress autoimmunity via up-regulation of Foxp3⁺CD4⁺ T_regs and CD5⁺ CD1d^hi B_regs in lupus-prone mice.