TGF-β–Induced Regulatory T Cells Directly Suppress B Cell Responses through a Noncytotoxic Mechanism

Substantial evidence has revealed that the thymus-derived naturally occurring CD4⁺Foxp3⁺ regulatory T cells (nTreg) are essential for immune homeostasis (1). Lack or dysfunction of nTreg in mice and human was associated with many autoimmune diseases (1–3). In addition, adoptive transfer of nTreg prevents or alters the course of autoimmune diseases in several animal models (4, 5).

It is notable that the therapeutic effect of nTreg on established autoimmune diseases is fairly unsatisfactory (5, 6). Several reasons may account for this inability. First, the low frequency makes sufficient numbers for cell immunotherapy a significant challenge. Although expansion in vitro can overcome this problem, it has been observed that repeated stimulation leads to diminished Foxp3 expression and suppressive activity (7). In addition, T effector cells contaminating in the nTreg may preferentially be expanded following this protocol, and therefore, this possibly limits their use in autoimmune disease treatment. Third, it is very likely that when Ag-specific effector cells are induced, nTreg cannot counteract the effector T cell responses and function. In addition, when inflammation is established, a large number of proinflammatory cytokines, such as IL-6, IL-1, and TNF-α are released. It has been reported that increased levels of IL-6 released by activated dendritic cells abolished the suppressive ability of nTreg (8). TNF-α possibly has a similar ability to affect the suppressive activity of nTreg (9). Furthermore, T cells from autoimmune conditions may be resistant to nTreg even though the properties of the nTreg are normal (10). In addition, a final possible reason could be dynamic alteration and plasticity of nTreg in the inflammatory milieu.

Considerable studies have also documented that induced Treg (iTreg) induced ex vivo with IL-2 and TGF-β share similar phenotypic and functional characteristics with nTreg (11). iTreg can suppress T cell activation, proliferation, cytokine production, and autoimmunity diseases in animal models (12). Moreover, iTreg but not nTreg are resistant to IL-6–driven Th17 cell conversion (13). Although previous studies by others have addressed the possibility of direct suppression on B cell by nTreg subset (14–16), it is less clear whether the iTreg subset has the similar suppressive capacity on B cells. In this study, we demonstrate that both nTreg and iTreg directly suppress B cells, but unlike nTreg, the suppression of B cells by iTreg does not involve cytotoxicity. We further show that TGF-β signaling is important for the effects of iTreg on B cells. Moreover, iTreg directly suppress autoreactive B cells in vitro and in vivo in lupus mice and display a superior effect on lupus B cell responses compared with nTreg subset.

Female C57BL/6 (B6), female granzyme B–deficient, female perforin-deficient mice, female NZB/NZW F₁ (NZBWF₁/J, BWF₁), B6 Rag1^−/− mice were purchased from The Jackson Laboratory (Bar Harbor, ME). C57BL/6 Foxp3-GFP knockin mice were provided by Dr. T. Chatilla (University of California, Los Angeles, CA). Female New Zealand mixed (NZM) 2328 mice were gifted by C. O. Jacob (University of Southern California, Los Angeles, CA). NZM2328 Foxp3-GFP mice were generated by breeding NZM2328 and C57BL/6 Foxp3-GFP together for 12 generations on NZM2328 background. Lymphocyte-specific TGF-β type II receptor (TβRII) conditional knockout (CKO) mice on a C57BL/6 background were provided by Dr. W. Shi (University of Southern California). All animals were treated according to National Institutes of Health guidelines for the use of experimental animal with the approval of the Penn State University and Sun Yat-sen University Committee for the Use and Care of Animals.

The following fluorescence conjugated mouse Abs were used for flow cytometric analysis from BioLegend (San Diego, CA): FITC-CD69 (H1.2F3), FITC-CD86 (PO3), allophycocyanin-CD80 (16-10A1), allophycocyanin-CD138(281-2), PE-B220 (RA3-6B2), Alexa Fluor 647-granzyme B (GB11), Alexa Fluor 488-Foxp3 (150D), PE-IL-10 (JES5-16E3), PE-TNF-α (MP6-XT22), PerCP/cy5.5-CD5 (53-7.3), PE-MHCII-I-A^b (AF6-120.1), allophycocyanin-CD24 (30-F1), PE-CD23 (B3B4), and PE-Ki-67 (16A8); eBioscience (San Diego, CA): FITC-CD21/CD35 (eBio4E3 (4E3)), PE-perforin (eBioOMAK-D), and PE-CD357(GITR) (DTA-1); and Santa Cruz Biotechnology (Santa Cruz, CA): PE-granzyme A (3G8.5). Cell subset was stained with mAbs and isotype control indicated above and analyzed on a FACSCalibur flow cytometer using Cell Quest Software (BD Biosciences). For intracellular staining, such as Foxp3, granzyme A, granzyme B, and perforin, cells were first stained with surface marker CD4 and further fixed and permeabilized for intracellular staining. Plot figures were prepared using FlowJo Software (Tree Star, Ashland, OR).

Naive CD4⁺CD62L⁺CD25⁻CD44^low T cells were isolated from spleen cells of C57BL/6, TβRII CKO, granzyme B KO, perforin KO, BWF₁, or NZM2328 mice using naive CD4⁺ T cell isolation kit (Miltenyi Biotec, Auburn, CA). Cells were cultured in 48-well plates and stimulated with anti-CD3/CD28 microbeads (one bead per five cells; Invitrogen) in the presence of IL-2 (50 U/ml; R&D Systems) with (iTreg) or without (CD4_med) TGF-β (2 ng/ml; R&D Systems) for 3 d. RPMI 1640 medium supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin, 10 mM HEPES (Invitrogen), and 10% heat-inactivated FCS (HyClone Laboratories) was used for all cultures. Foxp3 expression was determined by flow cytometry.

CD4⁺CD25⁺ cells sorted from the thymus in C57BL/6 mice were expanded with anti-CD3/CD28–coated beads (one bead per three cells; Invitrogen) and IL-2 (300 U/ml) for 3 d. A total of 300 U/ml IL-2 was renewed at day 2. In one experiment (Fig. 6), CD4⁺CD25⁺ cells were isolated from spleen cells. After cultures, cells were harvested and beads were removed. The percentage of Foxp3⁺ cells was examined by flow cytometry before and after 3 d expansion.

To examine the suppressive activity of Treg on B cell in vitro, B cells from C57BL/6, TβRII CKO, or BWF₁ mice were isolated by B220 positive sorting. Then, B cells were stimulated with or without LPS (2 μg/ml, Escherichia coli serotype 0111: B4; Sigma-Aldrich) in the presence or absence of nTreg, iTreg, or CD4_med cells. The graded concentration of Treg to B cell ratios was 1:2–1:16. Cells were cultured for 3 d, and 1 μCi/well [³H]thymidine was added for the last 16 h of culture. The proliferation was analyzed by the value of thymidine incorporation. Fresh B cells labeled with CFSE were cultured with nTreg, iTreg, or CD4_med cells (the ratio of T cells:B cells was 1:2–1:8) in the presence of LPS (2 μg/ml), B cell proliferation was determined by the CFSE dilution rates after 2 d of culture. In addition, CFSE-labeled B cells were preincubated with LPS (2 μg/ml) for 24 h, and then, cells were extensively washed to remove LPS. These B cells were recultured with Treg or control cells for an additional 2 d. The proliferative levels of B cells were judged by the rates and intensity of CFSE dilution with flow cytometry.

To determine the mechanism of Treg inhibition, different Abs, including anti–TGF-β (10 μg ml⁻¹; R&D Systems), TβRI (ALK5) inhibitor (10 μg ml⁻¹; Sigma-Aldrich), anti–IL-10 (10 μg ml⁻¹; R&D Systems), anti–IL-10R (10 μg ml⁻¹; BioLegend), anti–PD-1 (10 μg ml⁻¹; BioLegend), anti–programmed death ligand 1 (PD-L1) (10 μg ml⁻¹; BioLegend), anti-CTLA4 (10 μg ml⁻¹; BioLegend), anti-GITR (10 μg ml⁻¹; BioLegend), adenosine 5′-(α,β-methylene) diphosphate (100 μM; Sigma-Aldrich), sodium polyoxotungstate (100 μM; Tocris Bioscience), and their correlated isotype controls were added to the culture system described as above.

iTreg, nTreg, or CD4_med cells were cocultured with B cells in the presence or absence of LPS (2 μg/ml) for 16 h. The cells were then collected and stained with Annexin V and 7-aminoactinomycin D (7-AAD) using an Annexin V apoptosis detection kit (BD Biosciences) following the manufacturer’s instructions. Both Annexin V and 7-AAD expression were measured by FACSCalibur flow cytometer (BD Biosciences) gated on B220⁺ cells.

To compare the IgG Ab production from B cells in vitro, the supernatants were collected from different cultural systems after 3 d culture. For the in vivo autoantibody detection, mice were bled at the indicated time points, and sera were collected. IgG, IgM, and anti-dsDNA autoantibodies were measured by an ELISA as described previously (17). All samples tested for anti-DNA were performed at the same time. Sera were diluted 1/400 or 1/800 for anti-DNA and 1/40,000 for measuring IgG.

Total RNA was extracted from cells by using TRIzol reagent and used to determine the expression and relative level of the TβRI in Treg subsets. The mRNA levels of targeted genes were measured by quantitative RT-PCR (Applied Biosystems) by using the SYBR Green RT-PCR Reagents Kit (Life Technologies). The relative expression of TβRI on T cells before and after LPS stimulation was determined by normalizing expression of each target to hypoxanthine guanine phosphoribosyl transferase (HPRT): TβRI, 5′-TGT GCA CCA TCT TCA AAA ACA-3′ (forward) and 5′-ACC AAG GCC AGC TGA CTG-3′ (reverse); and HPRT, 5′-TGA AGA GCTACT GTA ATG ATC AGT CAA C-3′ (forward) and 5′-AGC AAG CTT GCA ACC TTA ACC A-3′ (reverse).

BWF₁ mice were monitored regularly. The sera IgG and anti-DNA autoantibodies were measured by ELISA. Proteinuria was assessed using Albustix reagent strips (Bayer, Elkart, IN). When the proteinuria developed, BWF₁ mice were received with a single dose of 300 μg anti-mouse CD3 Ab (17A2; BioLegend) to deplete the endogenous CD3⁺ T cells (18) or the isotype control. Seven days later, CD4_med or iTreg were adoptively transferred to the mice with Ab treatment. The percentages of plasma were detected and sera were collected 0, 1, and 2 wk after cell transfer for IgG measurement. Although homogeneous iTreg (derived from 10- to 12-wk age NZM2328) and nTreg (sorted from the thymus in 10- to 12-wk age NZM2328) were adoptively transferred to another lupus-prone mouse, NZM2328 (>4 mo age), the sera were also collected. Proteinuria and anti-DNA autoantibodies were measured. Sera IgG and IgM autoantibodies cluster were measured by Autoantigen Microarrays.

Fresh B cells were obtained from NZM2328 (16 wk age) and stimulated with LPS (2 μg/ml) for 24 h and then washed to remove LPS. These B cells were cotransferred with nTreg, iTreg, or CD4_med cells (the ratio of Treg cells:B cells was 1:2, B cells were 8 × 10⁶ for each mouse) into B6 Rag1^−/− mice. In some groups, TβRI (ALK5) inhibitor (1 mg/kg), anti–IL-10R (1 mg/kg), or IgG (control for anti–IL-10R) or DMSO (control for ALK5i) were given to mice by i.p. injection at days 0 and 2. The mice were sacrificed, and splenic B cells were harvested and prepared to test Ki-67 expression by flow cytometry.

Autoantibody activities against a penal of 123 autoantigen specificities were measured using an Autoantigen Microarray platform developed by University of Texas Southwestern Medical Center (19, 20). Briefly, sera samples were pretreated with DNAse-I and then diluted 1:50 in PBST buffer for autoantibody profiling. The autoantigen array bearing autoantigens and control proteins were printed in duplicates onto Nitrocellulose film slides (Grace Bio-Labs). The diluted sera samples were incubated with the autoantigen arrays and autoantibodies were detected with cy3-labeled anti-mouse IgG and cy5-labeled anti-mouse IgM using a Genepix 4200A scanner (Molecular Device) with laser wavelength of 532 and 635 nm. The resulting images were analyzed using Genepix Pro 6.0 software (Molecular Devices). The median of the signal intensity for each spot was calculated and subtracted the local background, and data obtained from duplicate spots were averaged. Finally, the net fluorescence intensity (NFI) of the Ab against each Ag was used to generate heat maps using Cluster and Treeview software (http://rana.bl.gov/EisenSoftware.htm). Each row in the heat map represents an autoantibody, and each column represents a sample. The red color represents the signal intensity higher than the mean value of the raw, and the green color means signal intensity is lower than the mean value of the raw. The black color represents the signal closing or equal to the mean value of the raw, and the gray color indicates value is 0 or missing. Microarray data have been deposited in the ArrayExpress public database (https://www.ebi.ac.uk/arrayexpress/browse.html; accession number E-MTAB-4441).

Data were expressed as mean ± SEM unless otherwise indicated. Data were analyzed using the unpaired t tests (Mann–Whitney) or paired t tests for comparison between two groups or ANOVA for comparison among multiple groups as appropriate. Differences were considered statistically significant when p < 0.05.

To determine whether iTreg can suppress B cell responses, the effects of both nTreg and iTreg cells on the activation, proliferation, and autoantibody production of stimulated B cells were compared. We detected the expression of CD69, CD80, CD86, and CD138 on B cells in the Treg-B cell coculture system. Unlike the control CD4_med (CD4⁺ T cells stimulated with TCR and IL-2 without TGF-β), both iTreg and nTreg similarly suppressed the expression of CD69, CD80, CD86, and CD138 on B cells (Fig. 1A), indicating that iTreg also suppress B cell activation and differentiation. It was noted that CD4_med also slightly suppressed B cell differentiation (Fig. 1A); possible explanations include 1) CD4_med may contain minor Foxp3⁺ cells and 2) CD4_med express high levels of CD25 that consume IL-2 and then indirectly reduce the B cell differentiation.

Using CFSE labeling, it is clearly observed that both Treg subsets significantly suppressed B cell proliferation (1:2 ratio; CD4_med versus Treg subsets; p < 0.001) (Fig. 1B). Interestingly, these Treg subsets even inhibited the proliferation of B cells that had been pretreated with LPS (Fig. 1B, lower panel). ³H incorporation assay further validated that both Treg subsets suppressed B cell proliferation (data not shown). Moreover, both nTreg and iTreg subsets also significantly suppressed TNF-α production by B cells (Supplemental Fig. 1A), but only iTreg cells upregulated IL-10 production by B cells (Supplemental Fig. 1A). However, both Treg subsets did not upregulate or downregulate MHC class II, CD21/35, CD23, and CD24 related to Ag-presenting ability and regulatory B cell (Breg) phenotypes (Supplemental Fig. 1B). Further work is needed to determine whether iTreg can induce Breg cell development.

The in vitro suppression by the iTreg subset on B cell proliferation was equivalent to the nTreg subset at the ratios of 1:2–1:8 (Fig. 1B) and exhibited a dose-dependent effect (data not shown). To determine whether the iTreg subset suppresses B cell function, supernatants from a Treg–B cell coculture system were tested, and IgG production by activated B cells was markedly reduced in cultures containing either nTreg or iTreg cells than that in cultures containing CD4_med or no T cells (Fig. 1C). The suppressive effect of both nTreg and iTreg on IgG levels produced by B cells was similar (Fig. 1C). Furthermore, both nTreg and iTreg subsets, but not CD4_med, markedly suppressed IgM production by B cells (Fig. 1D). The IgM levels were significantly lower in cultures containing iTreg than those containing nTreg cells (Fig. 1D), implicating the iTreg subset as having a stronger effect on B cell–mediated IgM production.

Given previous studies attribute the suppressive ability of nTreg subset on B cell responses to cytotoxic mechanisms (14, 18), we next tested whether the iTreg subset also has a similar target cell killing effect on B cells. When B cell and CD4_med were cocultured, the ongoing apoptosis levels of B cells were substantially elevated, and coculturing with nTreg subset further increased the B cell apoptosis, confirming the previous reports that nTreg suppress B cells through their killing (14). Nonetheless, the addition of iTreg subset did not change the levels of B cell apoptosis (Fig. 2A, 2B). To learn whether this functional difference is related to the expression levels of cytotoxic molecules on different Treg subsets, we detected the expression of granzyme A, granzyme B, and perforin because they are expressed with some extents on nTreg subsets (14). Although the nTreg subset expressed granzyme A and granzyme B (8–12%) as well as perforin (2%), the expression of these cytotoxic molecules in iTreg subset was almost undetectable (Fig. 2C, 2D). To exclude the possibility that the cell culture may change the expression of these molecules on Treg subsets, we also analyzed the expression of granzyme A, granzyme B, and [erforin before and after cocultures. These data revealed that granzyme A and granzyme B (Supplemental Fig. 2) and perforin (data not shown) were only expressed on the nTreg subset and cell culture did not change their levels, whereas the iTreg subset did not express these molecules before or after the cocultures (Supplemental Fig. 2).

We then studied the functional link of these molecules between Treg subset and B cell. While the nTreg subset isolated from wild-type (WT) mice suppressed B cell proliferation and induced B cell apoptosis, these cells failed to suppress B cell proliferation and apoptosis when isolated from granzyme B or perforin KO mice (Fig. 2E, 2F). Conversely, the suppression of iTreg subsets generated from WT, granzyme B, or perforin KO mice on B cell responses were similar and sustained (Fig. 2E). Thus, the TGF-β–induced iTreg subset suppresses B immune responses independent of the cytotoxicity mechanism that characterizes the nTreg subset.

Several studies have demonstrated that cell contact is required for suppression by both nTreg and iTreg subsets. To determine the importance of cell contact in Treg-mediated suppression of B cells, Treg subsets were separated from target B cells by Transwell inserts. Under these conditions, little suppression was observed in either nTreg or iTreg coculture system (Fig. 3A), suggesting that cell contact is needed for suppressive function of both Treg subsets on B cells. To further explore the mechanisms responsible for iTreg-mediated suppression on B cells, we analyzed several potential candidates, including anti–IL-10, anti–IL-10R, anti–TGF-β, anti–PD-1, anti–PD-L1, anti-CTLA4, anti-GITR, adenosine 5′-(α,β-methylene) diphosphate (CD73 inhibitor), or sodium polyoxotungstate (CD39 inhibitor), and no effect was observed on Treg-mediated B cell suppression (data not shown). However, using the TβRI antagonist-ALK5 inhibitor (Fig. 3B) or B cells isolated from TβRII KO mice (Fig. 3C), we found the suppressive effect of iTreg subset on B cells disappeared. This result is not surprising because we previously have reported that the iTreg subset expressed surface TGF-β and secreted activated TGF-β (21). We further demonstrated that iTreg also express increased levels of TβRI, but LPS stimulation did not change the TβRI expression on Treg subsets (Fig. 3D). That is reasonable because LPS mainly stimulates B cells but not T cells. This finding demonstrates that iTreg suppression is different from that of the nTreg, whose suppression is independent upon TGF-β signal (22).

An in vivo experiment further validates that the iTreg suppression does need TGF-β signal. B cells pretreated with LPS were cotransferred with CD4_med, iTreg, or nTreg (2:1 ratio) into Rag1^−/− mice; in some groups, ALK5i, anti–IL-10R Ab, control IgG, or DMSO (for Alk5i control) were i.p. injected (Fig. 4A). Ki-67 expression by LPS-pretreated B cells was dramatically increased 3 d after cell transfer; compared with cotransfer of CD4_med, cotransfer of iTreg or nTreg significantly prevented the upregulation of Ki-67 on B cells in vivo (Fig. 4B, 4C). iTreg effect was also superior to nTreg against B cell proliferation in vivo (Fig. 4B, 4C). iTreg suppression was mostly dependent on TGF-β signal because blockade of TβRI almost completely abolished the suppression of iTreg on B cell proliferation in vivo. However, IL-10 seems to also play a partial role in iTreg-mediated B cell suppression in vivo (Fig. 4B, 4D).

Although some groups have a concern that iTreg are unstable (23, 24), we and other have demonstrated that iTreg are stable under the right protocol to generate iTreg (13, 25, 26). New data in this study demonstrated that iTreg were fairly stable in vivo 3 wk after cell transfer (Supplemental Fig. 3). iTreg were completely resistant to Th17 cell conversion, although few of them had converted into Th1 cells (Supplemental Fig. 3). A previous study has revealed that iTreg subset maintains their functional feature even they began to express IFN-γ (27).

Previous study has documented that nTreg have a mildly suppressive effect on lupus (5). To determine whether the iTreg subset displays suppressive effects on lupus B cells, we first approached this using in vitro coculture experiments as described above, and showed that iTreg but not CD4_med control cells generated from young or old BWF₁ mice significantly suppressed IgG production by lupus B cells (Fig. 5A).

To further determine whether iTreg directly suppress B cells in lupus mice in vivo, iTreg or CD4_med control cells were adoptively transferred to old lupus mice that had been previously depleted of endogenous T cells. Treatment with iTreg but not CD4_med cells significantly decreased IgG Ab and anti-dsDNA titers in lupus mice (Fig. 5B–D). Moreover, adoptive transfer of iTreg to 16-wk-old lupus mice that had developed some levels of proteinuria almost completely prevented the increase in proteinuria after 12 wk of cell transfer, whereas CD4_med did not inhibit the development of proteinuria in lupus mice (Fig. 5E). We also found that the percentages of blood CD138⁺ plasma cells were significantly reduced in iTreg treatment group compared with CD4_med cell treatment group (Supplemental Fig. 4A, 4B). Although iTreg also slightly reduced the frequency of CD138⁺ plasma cells in spleen and bone marrow, the effect was not significantly different from CD4_med treatment (Supplemental Fig. 4C–E).

We demonstrated that TGF-β similarly induced CD4⁺CD25⁺Foxp3⁺GITR⁺ cells from either young (10 wk) or old (16 wk) NZM2328 mice (Fig. 6A). Moreover, the suppressive activity exerted by iTreg induced from either young or old mice was comparable as well as similar to splenic nTreg subsets from same aged mice (Fig. 6B). We choose mice with age of 16 wk because of the fact that NZM2328 mice at 12–16 wk have developed renal pathology (28, 29). This result suggests one may induce iTreg from the established autoimmune diseases, having an important clinical implication.

We directly compared the effects of nTreg and iTreg subsets on lupus B cells in NZM2328 lupus mice. Although the titers of almost all of Abs (both IgG and IgM) as indicated by NFI are elevated in lupus mice from days 0 to 32, infusion of either nTreg or iTreg not only prevented the continuous elevation of autoantibodies but also reduced these Abs, particularly at 32 d after cell transfer (Fig. 7). The mircoassays have included almost all of autoantibodies related to lupus. Except for anti-KU(P70/P80), –proliferating cell nuclear Ag, and –tissue transglutaminase IgG, the levels of all autoantibodies (IgG) were significantly decreased in lupus mice received iTreg than in lupus mice received nTreg subset (Fig. 7A). Regarding autoantibodies (IgM), iTreg subset treatment displayed a superior suppressive effect on production of autoantibodies compared with nTreg subset treatment (Fig. 7B).

We also measured the dynamic changes of total IgG and IgM autoantibodies. As shown in Fig. 8A and 8B, infusion of either nTreg or iTreg reduced IgG levels on 2 wk after cells transfer. This effect was further enhanced on 4 wk after cell transfer. Nonetheless, iTreg infusion resulted in greater suppression against autoantibodies production, particularly on 4 wk after cell transfer. iTreg but not nTreg subset controlled IgM autoantibodies production on 2 wk after cell transfer, although nTreg restored their suppression on week 4, the effect of iTreg subset on IgM autoantibodies production suppression was significantly greater than nTreg subset at that time point (Fig. 8A, 8B).

We simultaneously observed the changes of several important autoantibodies related to lupus development such as subsets of autoantibodies of IgG and IgM: anti-dsDNA, anti-Matrigel, anti-thyroglobulin, anti-histone, and anti-myosin. nTreg subset infusion significantly suppressed IgG anti-thyroglobulin, IgM anti-thyroglobulin, IgM anti-Matrigel, and IgG anti-histone at 4 wk after cell transfer; however, they failed to suppress IgG anti-dsDNA, IgM anti-dsDNA, IgG anti-Matrigel, IgM anti-histone, IgG anti-mysin, and IgM anti-myosin at 4 wk after cell transfer. Conversely, iTreg subset infusion significantly suppressed all of IgG and IgM Abs mentioned above at 4 wk after cell transfer (Fig. 8C–L). These results strongly imply that iTreg subset may have more potent suppressive effects on Ab production by autoreactive B cells in lupus mice.

Treg modulate immune tolerance through immune regulation mechanisms. Treg suppress T and other cells through complicated but as yet poorly understood mechanisms. Although T cells are considered to be the main target of Treg, recent studies have also demonstrated that Treg act on other immune cells as well as nonimmune cells (17, 30–33), implicating Treg in regulating tolerance balance through a wide range of influence.

Although pathogenic T cells are a key culprit, B cells are also required for the pathogenesis of autoimmune diseases. B cells are clearly needed for development of collagen-induced arthritis in mice and mice develop an arthritis that can be driven almost entirely by Igs (34–36). B cells can contribute to autoimmune diseases by the secretion of autoantibodies, presentation of autoantigen, secretion of inflammatory cytokines, modulation of Ag processing and presentation, and generation of ectopic germinal center (36).

Thus, the suppression of B cell function provides a new approach to combat autoimmune and inflammatory disease. B cell depletion using rituximab has been used for the treatment of a number of autoimmune and chronic inflammatory diseases (37, 38). Rituximab treatment results in nearly undetectable circulating B cell levels 1 mo after therapy, and B cell counts remain low for 6–12 mo (39). However, its role is limited by the fact that long-lived plasma cells and bone marrow stem cells are not directly depleted (36, 40), and B cells located in the peritoneal cavity are also strongly resistant to depletion therapy (41).

Treg therapy may provide an innovative strategy for controlling B cell–mediated diseases. It has been reported that nTreg suppress B cell responses as T cells (14, 18); nonetheless, the mechanisms underlying are completely different. Although iTreg cells suppress T cell by cell contact or suppressive cytokines, the role played by nTreg on B cells is related to direct B cell killing (14). This raises a concern as to whether nTreg are real suppressor cells and whether nTreg therapy is appropriate to treat B cell–mediated diseases.

iTreg are another Treg subset that has displayed similar phenotypes and suppressive activity in some autoimmune diseases (42–44). In this study, we demonstrated that iTreg also directly suppress B cells but their mechanism of action is completely different from nTreg and is almost completely independent of cell killing. This difference could be explained by different expression of perforin and granzyme A/B on both Treg subsets because nTreg express much higher levels of these molecules compared with iTreg subset. This observation was further confirmed by the effect of Treg subsets using perforin and granzyme KO mice.

We further revealed that the iTreg subset suppresses B cell responses mainly through the TGF-β receptor signaling pathway. This was shown by either adding TβRI antagonist or by using B response cells derived from TβRII KO mice. It is consistent with previous studies showing that an iTreg subset expresses surface TGF-β and secreted active TGF-β (21) and suppresses T and dendritic cells through TGF-β signaling (45).

The current study demonstrates that iTreg subsets also suppressed B cell differentiation to plasma cell in circulating blood, this is likely that iTreg eventually reduce autoantibodies production in lupus. Although the effect of iTreg on plasma cell development in spleen and bone marrow was minimal, given that most short- and long-lived plasma cells locate on spleen and bone marrow, it is unclearly whether the long-term effect of iTreg subsets on autoantibodies in lupus will be compromised.

We also studied the phenotype and cytokine profiles of B cells. iTreg significantly suppressed production of inflammatory cytokines but increased IL-10 by B cells. Although the B cells do not display the phenotypic characteristic of Breg after iTreg treatment, the increase of IL-10 production by B cells is likely significant because IL-10⁺ B cells have been considered as Breg that also play an important role in controlling inflammatory diseases. Because IL-10 is an intracellular cytokine, we are currently unable to sort IL-10⁺ B cells to analyze their functional activity. We are developing IL-10 reporter and Foxp3/IL-10 double reporter mice that will enable us to determine whether iTreg can induce Breg and IL-10–producing Tr1 cells eventually. Although all of these Treg subsets have a capacity to regulate immune responses, their effects are different. For example, Foxp3⁺ cells are important for initial immune suppression, whereas Tr1 cells are more important for long-term immune tolerance in a transplantation model (46). We previously have demonstrated that iTreg can induce the formation of tolerogenic dendritic cells or of new generation of Treg that builds the foundation of infectious tolerance and contributes to the long-term therapeutic role in lupus and other autoimmune diseases (17, 47). It has been noted that iTreg subset did not suppress MHC class II expression on B cells, and it warrants a further study to determine whether the Ag-presenting ability of B cells keeps normal after iTreg treatment.

One of the interesting observations is that the iTreg subset has an advantage in controlling B cell–mediated immune response. Although nTreg subset mildly suppressed IgG and IgM autoantibodies, iTreg subset markedly suppressed both IgG and IgM autoantibodies in lupus mice. Using Ab microassays, we have analyzed most autoantibodies related to lupus. These observations have strongly suggested that iTreg subset has a superior effect on the production of IgG, particularly of IgM Abs compared with nTreg in lupus mice. It is possible that iTreg are stable in the inflammatory condition (13, 27). Although others reported that iTreg subset is unstable (24), using the comparison study, we recently demonstrated iTreg are stable and functional as long as the appropriate protocol for the differentiation of iTreg is used (48). Thus, the current findings may provide a new strategy of using iTreg for treating lupus and other B cells–mediated autoimmune and inflammation diseases without the side effect of B cell killing/depletion. The clinical trial on iTreg therapy will determine this feasibility. Nonetheless, this observation may at least have translational applications.

The authors have no financial conflicts of interest.