Decrease in the proportion of CD24^hiCD38^hi B cells and impairment of their regulatory capacity in type 1 diabetes patients

Summary

B10 cells restore immune balance by producing interleukin (IL)‐10. Impaired B10 cell responses are related to numerous autoimmune diseases. However, the function of B10 cells in type 1 diabetes (T1D) patients is controversial. We hypothesized that there are numerical and functional defects of B10 cells in T1D. Sixty‐two patients with T1D and 74 healthy volunteers were included in our study. We showed that B10 cells in human peripheral blood belong to a CD24^hiCD38^hi B cell subpopulation. CD24^hiCD38^hi B cells from healthy individuals possessed regulatory capacity, suppressed interferon (IFN)‐γ, tumor necrosis factor (TNF)‐α and IL‐17A production and promoted IL‐4 production and forkhead box protein 3 (FoxP3) expression in CD4⁺ T cells through an IL‐10‐dependent mechanism. Compared to healthy controls, B10 cell percentages in T1D were significantly lower (5·6 ± 3·5 versus 6·9 ± 3·3%; P < 0·05), produced less IL‐10 (15·4 ± 4·3 versus 29·0 ± 4·5%; P < 0·001) and lacked regulatory capacity. In addition, Pearson’s correlation analysis showed that the frequency of circulating B10 cells was negatively correlated with the frequency of CD4⁺IFN‐γ⁺ and CD4⁺TNF‐α⁺ T cells (r = −0·248 and r = −0·283, P = 0·008 and P = 0·017, respectively), positively correlating with the frequency of CD4⁺CD25⁺FoxP3⁺ T cells (r = 0·247, P = 0·001). These data offer direct proof that there is a deficiency of circulating CD24^hiCD38^hi B cells in peripheral blood of patients with T1D, which participate in the T1D immune imbalance involved in the development of T1D.

Human B10 cells belong to a CD24hiCD38hi B cell subpopulation. CD24hiCD38hi B cells from healthy individuals possessed regulatory capacity through an IL‐10‐dependent mechanism. There are numerical and functional defects of CD24hiCD38hi B cells in T1D, which may participate in the T1D immune imbalance involved in the development of T1D.

Introduction

Type 1 diabetes (T1D) is a chronic autoimmune disease characterized by lack of insulin and irreversible destruction of pancreatic β cells. Environmental risk factors, together with genetic predisposition, contribute to T1D etiology. Autoimmune reaction is a key step during the progression of T1D, wherein T cells play a vital role 1. However, B cells also play important roles in the occurrence and development of T1D.

B cells have been characterized as positive regulators of humoral immune responses: (1) antigen‐presenting cells 2; (2) antibody secretion 3; (3) proinflammatory cytokine production 4; and (4) optimal T cell activation 5. Recently, regulatory B cells (B_regs) have been recognized as an important new component of the immune system. Specific B cell subsets were identified that exhibit negative regulatory function to suppress adaptive and innate immunity, inflammation and autoimmunity through the production of interleukin (IL)‐10 6.These IL‐10‐producing regulatory B cells have been labeled as B10 cells. B10 cells have been demonstrated in murine models of autoimmune disease, including systemic lupus erythematosus (SLE) 7, rheumatoid arthritis (RA) 8 and multiple sclerosis (MS) 7, 9. Murine B10 cells are reported to possess regulatory capacities in in‐vitro suppressive experiments and in‐vivo adoptive transfer assays in an IL‐10‐dependent mechanism.

Recently, human B10 cells were also found to regulate immune responses. Blair et al. demonstrated that CD24^hiCD38^hi B cells were potent IL‐10 producers upon CD40L stimulation. This cell subset could suppress differentiation of T helper type 1 (Th1) cells partly via IL‐10 10. However, Tedder et al. characterized human IL‐10‐producing B cells with the phenotype CD24^hiCD27⁺. They found that after stimulation with CD40L and cytosine–phosphate–guanosine (CpG), human CD24^hiCD27⁺ B cells suppressed the production of CD4⁺ tumor necrosis factor (TNF)‐α⁺T cells in an IL‐10‐independent mechanism 11. Consequently, there are controversies in the characterization of the phenotype and immunoregulatory mechanisms of human B10 cells. Moreover, the role of B10 cells in the pathogenesis of T1D and their relevance to the modulation of immune tolerance are relatively unknown.

In this study, we aimed to investigate the role of B10 cells in the pathogenesis of T1D. The phenotype of B_regs that produces the immunoregulatory cytokine IL‐10 was identified. We found that B10 cells in human peripheral blood belonged to a CD24^hiCD38^hi B cell subset. CD24^hiCD38^hi B cells from healthy individuals possessed negative regulatory function, including suppression of interferon (IFN)‐γ, TNF‐α and IL‐17A production and promotion of IL‐4 production and forkhead box protein 3 (FoxP3) expression in CD4⁺ T cells in an IL‐10‐dependent mechanism. There are numerical and functional deficiencies of circulating CD24^hiCD38^hi B cells in the peripheral blood of T1D patients.

Research design and methods

Patients and controls

Sixty‐two patients with T1D (29 females and 33 males; average age = 23·76 ± 5·89 years; age range = 7–29 years) and 74 healthy volunteers (39 females and 35 males; average age = 24·91 ± 2·92 years; age range = 20–30 years) were included in our study. To be enrolled into this study, an individual had to have a clinical diagnosis of T1D confirmed by criteria of the T1D Exchange Clinical Registry 12. The clinical manifestations of T1D patients and healthy volunteers are summarized in Supporting information, Table S1. This study received approval from the First Affiliated Medical Institution with the Nanjing Medical University ethical committee. All patients and healthy volunteers provided written informed consent to take part in this research.

Immunofluorescence analysis

For B cell staining, the following monoclonal antibodies (mAbs) were used from BioLegend (San Diego, CA, USA): CD19‐fluorescein isothiocyanate (FITC), CD24‐peridinin chlorophyll (PerCP)‐Cy5·5, CD38‐phycoerythrin (PE) and CD27‐allophycocyanin (APC). For identification of T cells, the following mAbs were used from BD Biosciences (San Diego, CA, USA): CD4‐PerCP and CD25‐APC. For intracellular staining, such as IFN‐γ‐Alexa Fluor (AF) 647, TNF‐α‐APC, IL‐4‐APC, IL‐17A‐APC, FoxP3‐FITC and IL‐10‐APC or PE mAbs (BioLegend), cells were first stained with mixtures of CD4‐PerCP, CD19‐FITC, CD24‐PerCP‐Cy5·5 and CD38‐PE mAbs and further fixed and permeabilized for intracellular staining. Data were obtained with a fluorescence activated cell sorter (FACS)Calibur (Becton Dickinson, Franklin Lakes, NJ, USA) flow cytometer and analyzed using FlowJo Software (TreeStar, Ashland, OR, USA).

Cell isolation and culture

Sodium‐heparinized venous blood was obtained from fasting T1D and control individuals. Peripheral blood mononuclear cells (PBMCs) were isolated through Ficoll‐Paque Plus (GE Healthcare, Chicago, IL, USA) gradient centrifugation within 2 h of sampling. B cellular subpopulations, CD4⁺CD25⁻ T cells and PBMCs depleted of B cellular subpopulations were sorted by FACSAria (Becton Dickinson).

Purified B and CD4⁺CD25⁻ T cells were cocultured in 96‐well U‐bottomed plates in RPMI‐1640 medium (gibco, Carlsbad, CA, USA) containing 10% fetal calf serum (FCS) (gibco), penicillin (200 µg/ml; Sigma, St Louis, MO, USA), streptomycin (200 U/ml, Sigma) and 4 mM L‐glutamine. For cytokine detection, CD4⁺CD25⁻ T cells in co‐cultures with B cells were stimulated with purified plate‐bound CD3 mAb (0·5 µg/ml, BD) for 72 h. Brefeldin A (BFA) (10 µg/ml; Biolegend) was added for the last 5 h together with phorbol 12‐myristate 13‐acetate (PMA) (50 ng/ml; Sigma‐Aldrich, St Louis, MO, USA) and ionomycin (1 µg/ml, Sigma‐Aldrich). Alternatively, anti‐IL‐10 (5 µg/ml) (JES3‐9D7), anti‐IL‐10 receptor (3F9), anti‐CD80 (10 µg/ml) (2D10.4) or anti‐CD86 (10 µg/ml) (IT2.2) or with their matching irrelevant isotype control was added to the co‐culture of CD3 mAb‐stimulated B and T cells. For experiments with T1D T cells, CD19⁺CD24^hiCD38^hi B cells were sorted by flow cytometry from PBMCs of healthy controls and cultured 1 : 1 with either autologous flow cytometric (FCM)‐sorted CD4⁺CD25⁻ T cells and autologous mitomycin cl (MMC) pre‐treated (25 µg/ml, 37℃, 30 min) accessory cells (PBMCs depleted of CD4⁺ T cells by flow cytometry) or with T1D FCM‐sorted CD4⁺CD25⁻ T cells and MMC pre‐treated accessory cells (1 : 10 MMC pre‐treated PBMCs: CD4⁺CD25⁻ T cells). T/B cells co‐culture were then stimulated with plate‐bound CD3 mAb for 72 h with phorbol myristate, ionomycin and brefeldin A (PIB) added for the last 5 h of culture. For analysis of IL‐10 secretion, FCM‐sorted B cells were cultured with CD40L (1 µg/ml; R&D Systems, Minneapolis, MN, USA) and CpG ODN 2006 (10 µg/ml; InvivoGen, San Diego, CA, USA) for 72 h. PMA + ionomycin + BFA were added for the last 5 h of culture. CD24^hiCD38^hi B cells were depleted from PBMCs of healthy donors and T1D patients by FCM sorting. CD24^hiCD38^hi B cell‐depleted PBMCs and non‐depleted PBMCs were then stimulated with plate‐bound CD3 mAb for 72 h, with PIB added for the last 5 h of culture.

Statistical analysis

Data were analyzed using unpaired t‐tests or paired t‐tests for comparisons between two groups or two‐way analysis of variance (anova) for comparison among three or more groups, as appropriate. Spearman’s rho was performed to calculate correlation coefficients and their significance. All statistical analyses were performed using GraphPad Prism. P‐values < 0·05 were considered statistically significant.

Results

Human peripheral blood B10 cells are CD24^hiCD38^hi

As noted earlier, the phenotypes of B10 cell subsets remain disputed. In humans, both CD24^hiCD38^hi 10 and CD24^hiCD27⁺ 11 B cells have been identified. Because B cell cytoplasmic IL‐10 was visualized after stimulation with CpG and PIB for 5 h, cell permeabilization was determined with respect to phenotype [immunoglobulin (Ig)M, IgD, CD1d, CD5, CD20, CD24, CD27, CD38, CD40]. We found that IL‐10⁺ B cells exhibited high CD24 and CD38, in contrast to IL‐10⁻ B cells (Fig. 1a). The median fluorescence intensity (MFI) of CD24 and CD38 was higher in IL‐10⁺ B cells than in IL‐10⁻ B cells (data not shown). As shown in Fig. 1b, a large mass of B10 cells was gathered within the CD24^hiCD38^hi gating after stimulation. Moreover, CD24^hiCD38^hi B cells define the brightest IL‐10 staining cells (Fig. 1c). To identify the phenotype of B10 cells, CD24^hiCD38^hi, CD24^hiCD38⁻ and CD24^intCD38^int B cells were sorted by FCM from healthy donors, cultured with CD40L and stimulated with CpG and PIB for 5 h. Intracellular staining for IL‐10 suggested that a drastically higher percentage of CD19⁺IL‐10⁺ B cells was in CD24^hiCD38^hi subset than in the CD24^hiCD38⁻ and CD24^intCD38^int subsets (Fig. 1d).

Figure 1.

Phenotypes of human blood B10 cells. (a) Representative cell surface molecule expression by interleukin (IL)‐10⁺ (solid line) and IL‐10⁻ (dashed line) CD19⁺ B cells from healthy individuals. Shaded histograms represent isotype‐matched control monoclonal antibody (mAb) staining (n = 13). (b) Representative dot‐plot showing CD19⁺IL‐10⁺ B cells (red dots) overlaid on a plot depicting CD24 and CD38 expression by total CD19⁺ B cells (black dots) (n = 13). (c) Representative dot‐plots showing the CD24 and CD38 expression by IL‐10⁻, IL‐10⁺ and IL‐10^hi B cells (n = 13). (d) Intracellular staining for IL‐10 in different B cell subpopulations are shown (n = 13). Bars represent the mean ± standard error (s.e.). ***P < 0·001.

CD24^hiCD38^hi B cells from healthy individuals possess regulatory capacity via an IL‐10‐dependent mechanism

To demonstrate the regulatory ability of CD24^hiCD38^hi B cells, sorted CD19⁺CD24^hiCD38^hi, CD19⁺CD24^hiCD38⁻ or CD19⁺CD24^intCD38^int B cells were co‐cultured with CD4⁺CD25⁻ T cells (1 : 1), which were stimulated with plate‐bound CD3 monoclonal antibody (B cell subpopulations and T cells were sorted with purities > 90%, as shown in Supporting information, Fig. S1). B cells producing IL‐10, CD4⁺IFN‐γ⁺, CD4⁺TNF‐α⁺, CD4⁺IL‐17A⁺, CD4⁺IL‐4⁺ and CD4⁺FoxP3⁺ T cells were evaluated by FCM. CD4⁺CD25⁻ T cells cultured with either CD24^hiCD38⁻ or CD24^intCD38^int B cells produced equivalent quantities of cytokines and transcription factors compared to cultures containing CD4⁺CD25⁻ T cells alone. CD24^hiCD38^hi B cells from healthy groups suppressed IFN‐γ, TNF‐α and IL‐17A production while promoting IL‐4 production and FoxP3 expression in CD4⁺ T cells (Fig. 2 and Supporting information, Fig. S2a).

Figure 2.

CD24^hiCD38^hi B cells from healthy individuals possess regulatory capacity. Flow cytometric (FCM)‐sorted CD19⁺CD24^hiCD38^hi, CD19⁺CD24^hiCD38⁻ or CD19⁺CD24^intCD38^int B cells from healthy individuals were cultured 1 : 1 with sorted CD4⁺CD25⁻ T cells. Cells were then stimulated with plate‐bound CD3 monoclonal antibody (mAb) for 72 h, with phorbol myristate, ionomycin and brefeldin A (PIB) added for the last 5 h of culture. T/B cells co‐culture were stained with CD19, CD4 and interferon (IFN)‐γ, tumor necrosis factor (TNF)‐α, interleukin (IL)‐4, IL‐17 or forkhead box P3 (FoxP3). Representative histograms showing the frequencies of CD4⁺IFN‐γ⁺ (a), CD4⁺TNF‐α⁺ (b), CD4⁺IL‐4⁺ (c), CD4⁺IL‐17A⁺ (d) and CD4⁺FoxP3⁺ (e) T cells (n = 4). Bar chart shows cumulative data. Bars represent the mean ± standard error (s.e.). *P < 0·05, **P < 0·01.

If CD24^hiCD38^hi B cells possess regulatory capacity, what are the potential effector mechanism(s)? To address this, we cultured CD4⁺CD25⁻ T cells either alone or with CD24^hiCD38^hi B cells (1 : 1), and neutralizing IL‐10/IL‐10R, CD80 and CD86 antibodies were added separately or together to the co‐culture. The addition of blocking antibodies against IL‐10 and IL‐10R, CD80 or CD86 individually and a combination of the three mAbs appeared to reduce the ability of CD24^hiCD38^hi B cells to inhibit the productions of IFN‐γ and TNF‐α by CD4⁺ T cells (Supporting information, Fig. S3a,b). Although blockade of IL‐10 and IL‐10R appeared to release the ability of CD24^hiCD38^hi B cells to inhibit the production of IL‐17A and promote the production of IL‐4 and FoxP3, the addition of blocking antibodies against CD80 or CD86 was not sufficient to release the suppression of IL‐17A and the promotion of IL‐4 and FoxP3. However, a combination of all three mAbs reversed the suppression of IL‐17A (Supporting information, Fig. S3d) and the promotion of IL‐4 and FoxP3 expression by CD4⁺ T cells (Supporting information, Fig. S3c,e). Thus, CD24^hiCD38^hi B cells from healthy individuals possessed the regulatory capacity, suppressed IFN‐γ, TNF‐α and IL‐17A production and promoted IL‐4 production and FoxP3 expression in CD4⁺ T cells through an IL‐10‐dependent mechanism.

CD24^hiCD38^hi B cells are decreased and imbalanced in Th cells in T1D patients

T1D is an autoimmune disorder characterized by the T cell‐mediated destruction of insulin‐producing pancreatic β cells. B cells also play important roles in the occurrence and development of T1D. First, we observed a numerical deficit in CD24^hiCD38^hi B cells in the peripheral blood of T1D patients compared to healthy controls (5·6 ± 3·5 versus 6·9 ± 3·3%, P < 0·05) (Fig. 3a). To further demonstrate the immune imbalance in T1D, the proportions of circulating T cell subsets (CD3⁺, CD4⁺, CD8⁺, CD4⁺IFN‐γ⁺, CD4⁺TNF‐α⁺, CD4⁺IL‐4⁺, CD4⁺IL‐17A⁺ and CD4⁺CD25⁺FoxP3⁺ T cells) were detected by FCM. Differences in the percentage of CD3⁺ (69·30 ± 7·11 versus 65·96 ± 9·59%, P = 0·025), CD4⁺ (52·61 ± 8·74 versus 48·55 ± 8·14%, P = 0·006), CD4⁺IFN‐γ⁺ (16·59 ± 5·52 versus 14·27 ± 3·28%, P = 0·024), CD4⁺TNF‐α⁺ (44·62 ± 8·21 versus 40·50 ± 8·11%, P = 0·038), CD4⁺IL‐4⁺ (1·65 ± 0·85 versus 3·18 ± 0·93%, P<0·001) and CD4⁺IL‐17A⁺ (3·29 ± 1·18 versus 2·87 ± 0·88, P = 0·046) T cells between patients with T1D and healthy controls were statistically significant. However, the percentages of CD8⁺ (35.36 ± 7·87 versus 36·42 ± 8·35%, P = 0·442) and CD4⁺CD25⁺FoxP3⁺ (5·46 ± 1·48 versus 5·29 ± 1·35%, P = 0·493) T cells in PBMCs of T1D patients versus controls were not statistically significant (Supporting information, Fig. S4). Pearson’s correlation analysis was performed to explore the correlations. Of note, the frequency of circulating CD24^hiCD38^hi B cells was negatively correlated with the frequency of CD4⁺IFN‐γ⁺ and CD4⁺TNF‐α⁺ T cells (r = −0·248 and r = −0·283, P = 0·008 and P = 0·017, respectively) and positively correlated with the frequency of CD4⁺CD25⁺FoxP3⁺ T cells (r = 0·247, P = 0·001) (Fig. 3b–i). There was also a negative correlation between the frequency of circulating CD24^hiCD38^hi B cells and HbA1c（r = −0·382, P = 0·015) (Fig. 3j–m) in T1D patients.

Figure 3.

Type 1 diabetes (T1D) patients exhibit reduced numbers of CD24^hiCD38^hi B cells compared to healthy individuals. Peripheral blood mononuclear cells (PBMCs) isolated from patients with T1D and healthy controls were stained with CD19, CD24 and CD38. Representative flow cytometry plots of B cell subset gating for a T1D sample and a healthy sample showing CD24^hiCD38^hiB cells (a). Cumulative results are expressed as the mean ± standard error (s.e.). *P < 0·05. Graphs showing the correlation between the frequency of CD24^hiCD38^hiB cells and the frequencies of CD3⁺ T cells (b), CD3⁺CD8⁺ T cells in CD3⁺ T cells (c), CD3⁺CD4⁺ T cells in CD3⁺ T cells (d), CD4⁺interferon (IFN)‐γ⁺ (e), CD4⁺tumor necrosis factor (TNF)‐α⁺ (f), CD4⁺interleukin (IL‐4⁺ (g), CD4⁺IL‐17A⁺ (h), CD4⁺forkhead box protein 3 (FoxP3)⁺ (i) T cells in CD4⁺ T cells, age (j), age of onset (k), T1D duration (l), or HbA1c (m) from 74 healthy individuals and 62 T1D patients.

CD24^hiCD38^hi B cells lack regulatory ability in T1D patients

We next evaluated whether the regulatory capacity of CD24^hiCD38^hi B cells should be changed in T1D. Because it was difficult to obtain sufficient numbers of CD24^hiCD38^hi B cells from peripheral blood samples of patients with T1D, we were unable to evaluate whether CD24^hiCD38^hi B cells could regulate T cell responses directly. Alternatively, we attempted to overcome this problem by depleting the CD24^hiCD38^hi B cells from healthy volunteers or T1D patient PBMCs and investigating the changes of cytokine production or transcription factor expression using an indirect method. We found a large increase in the frequency of CD4⁺IFN‐γ⁺, CD4⁺TNF‐α⁺, CD4⁺IL‐17A⁺ T cells and a substantial decrease in the frequency of CD4⁺IL‐4⁺, CD4⁺FoxP3⁺ T cells in CD24^hiCD38^hi‐depleted B cells compared to non‐depleted PBMCs from healthy volunteers. In comparison to healthy volunteers, depletion of CD24^hiCD38^hi B cells from the PBMCs of patients with T1D did not lead to any dramatic increases in the proportion of CD4⁺IFN‐γ⁺, CD4⁺TNF‐α⁺ or CD4⁺IL‐17A⁺ T cells and any dramatic decreases in the percentages of CD4⁺FoxP3⁺ T cells (Fig. 4a). In addition, the increasing in Th1/Th2 and Th17/regulatory T cell (T_reg) ratio was observed in CD24^hiCD38^hi‐depleted PBMCs compared to non‐depleted PBMCs from healthy controls. No significant changes were seen in T1D patients (data not shown). These data suggested that CD24^hiCD38^hi B cells in T1D patients had abnormal function.

Figure 4.

CD24^hiCD38^hi B cells from type 1 diabetes (T1D) patients fail to regulate T cell responses and express reduced interleukin (IL)‐10 compared to healthy individuals. (a) CD19⁺CD24^hiCD38^hi B cells were depleted by flow cytometry sorting, and the remaining peripheral blood mononuclear cells (PBMCs) were collected. Non‐depleted PBMCs were also collected as a control. Cells were then stimulated with plate‐bound CD3 monoclonal antibody (mAb) for 72 h, with phorbol myristate, ionomycin and brefeldin A (PIB) added for the last 5 h of culture. Depleted and non‐depleted PBMCs were surface‐stained with CD4 mAbs, permeabilized and stained with interferon (IFN)‐γ, tumor necrosis factor (TNF)‐α, IL‐4, IL‐17A and forkhead box protein 3 (FoxP3) mAbs. Representative histograms showing the frequencies of CD4⁺IFN‐γ⁺ (a), CD4⁺TNF‐α⁺ (b), CD4⁺IL‐4⁺ (c), CD4⁺IL‐17A⁺ and CD4⁺FoxP3⁺ T cells in CD4⁺ T cells. Graphs on the right showing differences in the frequency of cytokine production between depleted and non‐depleted PBMCs from the same individual. The results from at least four healthy individuals and three T1D patients are shown. P‐values comparing depleted and non‐depleted PBMCs were calculated by two‐tailed t‐test. *P < 0·05, **P < 0·01. (b) Healthy CD19⁺CD24^hiCD38^hi B cells regulate CD4⁺CD25⁻ T cells from T1D patients. Flow cytometric (FCM)‐sorted CD19⁺CD24^hiCD38^hi B cells from healthy individuals cultured 1 : 1 with either autologous FCM‐sorted CD4⁺CD25⁻ T cells or with T1D FCM‐sorted CD4⁺CD25⁻ T cells, then stimulated with plate‐bound CD3 mAb for 72 h. PIB added for the last 5 h of culture. Cells were surface‐stained with CD4 mAbs, permeabilized and stained with IFN‐γ, TNF‐α, IL‐4, IL‐17A and FoxP3 mAbs. Bar chart shows mean ± standard error (s.e.) percentage regulation of IFN‐γ, TNF‐α, IL‐4, IL‐17A and FoxP3 expression by CD4⁺CD25⁻ T cells after culture with CD19⁺CD24^hiCD38^hi B cells compared to T1D CD4⁺CD25⁻ T cells cultured alone. (c) PBMCs isolated from 13 patients with T1D and 13 healthy controls were surface‐stained for expression of CD19, CD24 and CD38 and intracellularly stained for IL‐10. Bar chart shows cumulative data. Bars represent the mean ± s.e. * P < 0·05, **P < 0·01, *** P < 0·001.

It is possible that effector T cells are refractory to the regulatory effect sustained by CD24^hiCD38^hi B cells or numerical deficiency of the circulating CD24^hiCD38^hi B cells in T1D (hence depleting the CD24^hiCD38^hi B cells in T1D would have minimal effect), or a functional defect of the immature B cells to regulate. To address these possibilities, we co‐cultured healthy CD24^hiCD38^hi B cells in the presence of either autologous CD4⁺CD25⁻ T cells or with CD4⁺CD25⁻ T cells isolated from T1D patients. Our results show that CD24^hiCD38^hi B cells isolated from healthy controls were able to regulate IFN‐γ, TNF‐α, IL‐4, IL‐17A and FoxP3 expression by T1D CD4⁺CD25⁻ T cell almost as efficiently, as they were able to regulate cytokine and transcription factors expression by autologous healthy CD4⁺CD25⁻ T cells (Fig. 4b). Therefore, our data demonstrated that effector T cells from T1D patients are not refractory to regulation.

IL‐10 production of CD24^hiCD38^hi B cells is decreased in T1D patients

IL‐10 is considered a hallmark of B10 cell regulation. We and others have shown that compared to other B cell subpopulations, CD24^hiCD38^hi B cells produced higher levels of IL‐10. We assessed whether IL‐10 production in CD24^hiCD38^hi B cells was also decreased in T1D patients. After PBMCs from healthy donors or T1D patients were cultured with CD40L and CpG for 72 h with PIB added for the last 5 h, IL‐10 production of CD19⁺ B cells and CD24^hiCD38^hi B cells was dramatically lower in patients with T1D compared to healthy controls (Fig. 4c). Therefore, there is a functional deficit of CD24^hiCD38^hi B cells in T1D patients.

Discussion

T1D is characterized by the autoimmune destruction of pancreatic β cells, mainly by CD4⁺ and CD8⁺ T lymphocytes, resulting in patients having to rely upon exogenous insulin. In the progression of T1D, genetic predisposition combines with environmental aspects to contribute to expression of pancreatic β cell autoantigens, which are required for development of autoreactive T lymphocytes that infiltrate pancreatic islets 13, 14. At present, there is a broad consensus that Th1 and Th17 cells induce β cell apoptosis by secreting cytokines; meanwhile, Th2 and T_reg cells play a protective function. We and others have confirmed an imbalance of circulating T lymphocytes in peripheral blood of T1D patients 15, 16, 17.

Although T1D has been classically described as a CD4⁺ T cell‐mediated disease, B cells also play an important function in the autoimmune destruction of pancreatic islets. B cells participate in the progression of T1D by presenting islet‐derived peptides to autoreactive T cells, producing autoantibodies to the islet cell, insulin, glutamic acid decarboxylase 65 (GAD 65), protein tyrosine phosphatases‐2 (IA‐2) and the more recently described zinc transporter 8 (ZnT 8)3, 18, 19. B cell depletion leads to reduced insulitis and delays the onset of diabetes in non‐obese diabetic (NOD) mice 20, 21. Similarly, rituximab (anti‐CD20) has been proved to delay the fall in C‐peptide levels and is associated with lower HbA1c and decreased requirement for insulin therapy in recent‐onset T1D patients 22, 23. Moreover, studies suggest that the reconstituted B cell pool expresses an immunosuppressive phenotype, which may directly contribute to the clinical effects of B cell depletion therapies 24, 25.

In recent years, an increasing body of work has been issued that supports the existence of B_regs that are capable of regulating T cell‐dependent autoimmune responses 6. Numerical and functional defects in B_regs have been described in several autoimmune diseases, including SLE 10, 26, RA 27 and MS 11. Our study aimed to identify the role of human B_regs in T1D. First, we identified the phenotypical characteristics of B_regs. In human, reported phenotypes of B_reg subsets include CD19⁺CD24^hiCD38^hi 10, 26, 27, 28, 29, 30, CD19⁺CD27^intCD38⁺ 31 and CD19⁺CD27⁺CD24^hi 11, 32, 33. These differences are due most probably to the use of a different panel of markers to characterize B cell subsets, different disease models and organic sources and the use of different induction methods. Like Blair et al., our combined results describe B10 cells in human peripheral blood as being predominantly found within the CD24^hiCD38^hi B cell subpopulation upon CD40 activation 6, 34.

The regulatory mechanism of human B10 cells remains to be fully elucidated. IL‐10 production is central to their negative regulation of immune responses. We noted that CD24^hiCD38^hi B cells from healthy volunteers possessed regulatory capacity that suppressed IFN‐γ, TNF‐α and IL‐17A production. In addition, activated B10 promoted IL‐4 production and induced FoxP3 expression in CD4⁺ T cells, which exhibited broader suppressive functions. Furthermore, immunomodulation by B10 was IL‐10‐dependent. Our findings imply a dual ability of CD24^hiCD38^hi B cells to maintain immune tolerance. On one hand, they exert their regulation effects through the production of IL‐10 by preventing commitment of CD4⁺ T cells into Th1 and Th17 cells. On the other hand, they may convert effective T cells into Th2 cells and T_regs. These suggest that B10 cells could have an effect on the regulation of pathology.

The role of CD24^hiCD38^hi B cells in T1D patients and their relevance to the maintenance of peripheral tolerance remains elusive. Habib et al. found increased CD24^hiCD38^hi B cells in T1D patients 35. However, Thompson et al. found no diversity in the frequency of CD24^hiCD38^hi B cell subset in patients with T1D compared to healthy controls 36. Interestingly, the large mass of CD19⁺CD1^hiCD5⁺ B cells, previously reported to be regulatory in mice, was contained within the CD24^hiCD38^hi B cell subpopulation 27. Deng et al. reported a decrease in CD1^hiCD5⁺ B cells in T1D patients 37. T1D is a heterogeneity of disease which might explain these discrepant results. Kleffel et al. suggested that antigen‐activated B10 cells may maintain tolerance to islet autoantigens by inhibiting the production of IFN‐γ⁺ T cells and hyperglycemic (Hglc) mice and T1D patients lack this population of regulatory B cells 25.

Our study demonstrated that patients with T1D are characterized by a numerical and functional deficiency in circulating CD24^hiCD38^hi B cells. In contrast to healthy controls, the proportion of CD24^hiCD38^hi B cells was significantly reduced in T1D patients and was negatively correlated with the HbA1c level. However, we did not observe the same age‐associated changes in B cell population reported by Thompson et al. These discrepancies might be a result of Chinese patients (29 female and 33 male; age range = 7–29 years) in our study versus Caucasian (21 female and 24 male; age range = 9–42 years) in their study. Furthermore, CD24^hiCD38^hi B cells isolated from patients with T1D displayed damaged IL‐10‐producing ability upon CD40 activation, were unable to inhibit Th1 and Th17 responses, failed to convert CD4⁺ CD25⁻ T cells into T_regs and lacked the ability to sustain Th1/Th2 and Th17/T_reg cell balance. However, our data showing that CD24^hiCD38^hi B cells from healthy volunteers were able to regulate cytokine and transcription factors expression by T1D CD4⁺CD25⁻ T cell indicates that T1D effector T cells are not refractory to the regulatory effect sustained by CD24^hiCD38^hi B cells. Therefore, our data suggest that a numerical and functional deficiency in circulating CD24^hiCD38^hi B cells was present in T1D patients.

Our study is subject to several limitations. In this study, we compare numbers and function of CD24^hiCD38^hi B cells from individuals with overt T1D and healthy controls (HC). However, from a disease‐predictive viewpoint it would be of value to determine if diminished CD24^hiCD38^hi B cell numbers and/or activity also distinguish individuals that are autoantibody‐positive, but have not yet progressed to overt T1D to HC. Next we will undertake these studies in at‐risk individuals. It is noted that in matching the patients and controls for age (23·76 ± 5·89 versus 24·91 ± 2·92, P = 0·17), 13 of the individuals with type 1 diabetes are aged 18 years or less and that these are not matched with healthy control individuals. In addition, we found no difference in the frequency of circulating CD24^hiCD38^hi B cells with age, age of onset or sex. However, we could not obtain sufficient CD24^hiCD38^hi B cells to investigate the association between the function of CD24^hiCD38^hi B cells with clinical factors.

Various immunotherapies have attempted to prevent T1D or to ameliorate the course of the disease after clinical diagnosis. Rituximab, which binds CD20 and depletes B cells, has been proved to preserve β cell function in patients with T1D. Declining C‐peptide responses and the use of exogenous insulin were both reduced. Moreover, hemoglobin A1c levels were lower in T1D patients treated with rituximab 22, 23. However, the mechanism of rituximab therapy in T1D remains poorly understood. For rituximab therapy in pemphigus patients, Colliou et al. found that patients in complete remission after therapy with rituximab exhibited dramatically higher numbers of IL‐10‐producing CD24^hiCD38^hi B cells than patients with active lesions at baseline or those in incomplete remission 38. We hypothesized that rituximab is a B cell‐depleting antibody with the capability to reshape the B cell repertoire during reconstitution. Anti‐B cell treatment with anti‐CD20 mAb depletes B cells, and whether rituximab‐mediated anti‐B cell treatment alters the frequencies of IL‐10‐producing CD24^hiCD38^hi B cells in T1D needs to be further investigated.

Cellular immunotherapy is becoming increasingly popular; for example, T_regs have already demonstrated high efficacy in the therapy of T1D. B_regs are also significant modulators of the immune response and boost immunological tolerance 39, 40. Our group has recently demonstrated that adoptive transfer of FACS‐sorted B10 cells delays graft survival in a mouse model of experimental islet transplantation 41. However, further investigations on the plasticity and functional stability of B_regs are necessary to understand how to maintain a prolonged B_reg phenotype in vivo.

In summary, we provide direct evidence that there is a deficiency of circulating CD24^hiCD38^hi B cells in patients with T1D, which might participate in the T1D immune imbalance that is involved in the development of T1D. The data presented herein provide new insights into immune therapy in patients with T1D. New treatments targeting B_reg activity could be a promising approach to treat T1D.

Disclosures

The authors have no financial interests or any conflicts of interest to disclose.

Supporting information

Fig. S1. Example of human B and T cell subset purity after purification

Fig. S2. Flow cytometry gating strategies for analysis of intracellular cytokine and transcription factors in the co‐culture system.

Fig. S3. CD24hiCD38hi B Cells possess regulatory capacity via an IL‐10‐dependent pathway

Fig. S4. Imbalance of circulating T cell subsets in peripheral blood of patients with T1D.

Table S1. The clinical characteristics of participants