We have previously reported that IL-10+ regulatory B cells, known to play an important role in controlling autoimmunity and inflammatory disorders, are contained within the transitional 2 immature (T2) B cell pool (T2 Bregs). Therapeutic strategies facilitating their enrichment or enhancing their suppressive activity are highly attractive. In this study, we report that agonistic anti-CD40 specifically targets T2 B cells and enriches Bregs upon short-term in vitro culture. Although transfer of unmanipulated T2 B cells, isolated from mice with established lupus, failed to confer protection to diseased mice, transfer of in vitro anti-CD40-generated T2 B cells (T2-like-Bregs) significantly improved renal disease and survival by an IL-10-dependent mechanism. T2-like-Bregs readily accumulated in the spleen after transfer, suppressed Th1 responses, induced the differentiation of IL-10+CD4+T cells, and conveyed a regulatory effect to CD4+T cells. In addition, in vivo administration of agonistic anti-CD40, currently on trial for the treatment of cancer, halted and reversed established lupus. Taken together, our results suggest a novel cellular approach for the amelioration of experimental lupus.
Regulatory B cells (Bregs)4 are present in several murine models of chronic inflammation, including collagen-induced arthritis (CIA), inflammatory bowel disease, and experimental autoimmune encephalomyelitis (1, 2, 3). This regulatory function appears to be directly mediated by the production of IL-10 and by the ability of B cells to interact with pathogenic T cells to dampen harmful immune responses (4). Although the most widely used markers for Bregs are the expression of IL-10 and CD19, combinations of other surface molecules including CD1d/CD21 (2), CD5/CD1d (5), or CD1d, CD21, CD23, and CD24 (transitional 2 B cells: (T2)) (6) have emerged as additional markers for the identification of this subset of B cells.
It has previously been shown that, in contrast to immunized wild-type mice, chimeric mice lacking IL-10 or CD40 exclusively on their B cells fail to recover from experimental autoimmune encephalomyelitis (3). In addition, transfer of B cells from TCRα−/− mice to B cell-deficient TCRα−/− mice with established colitis markedly decreased the number of pathogenic colonic CD4+TCRα−ß+ T cells in recipient mice while B cells isolated from CD40KO mice failed to confer similar protection (7). We have also shown that stimulation of arthritogenic splenocytes with anti-CD40 stimulates B cells to produce IL-10 (1). Even though activation of B cells via CD40 appears to be a requisite for the production of IL-10, it remains to be established whether all B cells, or a discrete subset, are the direct targets of CD40-induced IL-10 production. Similarly to regulatory T cells (Tregs), the low number of Bregs (and the unknown phenotype) limits the possibilities of using them as cellular therapy. Thus, the discovery that a specific B cell subset is the target of anti-CD40 stimulation would lead to an optimization of Breg expansion, a key factor for cellular therapy.
Patients with systemic lupus erythematosus and MRL/lpr mice spontaneously develop a severe autoimmune disease characterized by hypergammaglobulinemia, immune complex-associated end organ disease of the kidney, and production of anti-dsDNA, anti-small nuclear ribonucleoprotein and rheumatoid factor specificity autoantibodies (8, 9). In this study, report an efficacious therapeutic strategy, which polyclonally enriches a defined set of Bregs, contained within the T2 B cell subset, previously reported to control experimental arthritis (6). Upon adoptive transfer, these newly generated T2-like Bregs reversed autoimmunity in MRL/lpr mice, suppressed Th1 responses, and induced the differentiation of IL-10+CD4+T cells able to convey suppressive effect to other T cells. This protective and immunosuppressive effect was abrogated by in vivo neutralization of IL-10. Finally, in vivo administration of anti-CD40 reversed nephritis and enhanced survival in mice with new onset disease by increasing the number of T2 B cells, which we show to be numerically impaired in diseased mice.
Materials and Methods
MRL/Mp-Tnfrsf6lpr (MRL/lpr), C57/BL6, and MRL/Mp mice were purchased from Harlan and housed in a specific pathogen-free animal facility at the University College of London. All animal studies were conducted in accordance with protocols approved by the Home Office (United Kingdom). hCD20-transgenic (Tg) mice were previously made by using bacterial artificial chromosomes incorporating the hCD20 locus. To generate hCD20 Tg MRL/lpr mice, we backcrossed the founder line to MRL/lpr mice for over 15 generations (10).
The treatment Abs used were: rat IgG agonistic mAb reactive with CD40 (FGK45, provided by Prof. D. Gray, Edinburgh, U.K.), rat IgG, blocking mAb reactive with mouse IL-10R (1B1.2, purchased from American Type Culture Collection), rat IgG1 anti-IL-10 (JES5-2A5; American Type Culture Collection), anti-mouse IgM (F(ab′)2) R6-60.2, BD Biosciences), AFRC-Mac-1 (isotype control) rat IgG anti-dog chlamydomonas cell wall glycoprotein (European Collection of Animal Cell culture, Salisbury, U.K.), and anti-TGF-β (clones 1D11 was purchased from R&D Systems). The mAbs were purified from culture supernatants by affinity chromatography using a staphylococcal protein G column (Bioprocessing) and filter sterilized. All Abs used throughout the experiments were below the limit of detection for LPS levels using the Limulus amebocyte lysate test (BioWhittaker). The following anti-mouse Abs were purchased from BD Biosciences: anti-CD19- PE-Cy7 (1D3), anti-CD21-FITC and 7G6, anti-CD23-PE (B3B4), anti-IgM-allophycocyanin (II/41), anti-CD24-biotin (M1/69), anti-CD4-FITC (H129.19), anti-CD25-PE (7D4), anti-CD1d-PE (1B1), anti-AA4-PE, anti-IL10-allophycocyanin (JES5-16E3), anti-TNF-α-allophycocyanin (MP6-XT22), and anti-IFN-γ- allophycocyanin (XMG1.2). Goat anti-IgG-FITC, anti-mouse IgG-alkaline phosphatase (AP), IgG1-AP, and IgG2a-AP were from Southern Biotechnology Associates. Purified anti-mouse CD3ε (145-2C11; BD Biosciences) and F(ab′)2 goat anti-mouse IgM (Jackson ImmunoResearch Laboratories) were used for in vitro assays.
Murine cells were cultured in RPMI 1640 containing l-glutamine and NaHCO3 (R8758; Sigma-Aldrich) supplemented with 100 U/μg/ml penicillin/streptomycin (Life Technologies) and 10% FCS (BioWhittaker) in 96-well U-bottom plates (Nunc).
Murine lupus assessment
Survival was assessed in MRL/lpr mice that died of disease spontaneously and those sacrificed due to general debility. Urinary protein levels were assessed weekly semiquantitatively using reagent strips for urinalysis (Albustix; Bayer). Histopathology was evaluated on paraffin-embedded, formalin-fixed tissue sections by routine H&E staining. IgG immune deposits were determined by direct immunofluorescence on 5 μM OCT-embedded frozen kidney sections using FITC anti-mouse IgG. Sections were analyzed with a Bio-Rad MRC 1024 confocal system equipped with an argon and helium/neon laser for excitation at 522 and 585 nm. Images were acquired with the LaserSharp 3.2 software (Bio-Rad) and analyzed using Confocal Assistant 4 (Bio-Rad). Histopathological assessments were determined in a blinded fashion by one of us (A.K.C.). Kidneys were graded for glomerular inflammation, proliferation, crescent formation, and necrosis. Interstitial changes and vasculitis were also noted. Scores from 0 to 3 were assigned for each of these features and then added together to yield a final renal score. For example, glomerular inflammation was graded: 0, normal; 1, few inflammatory cells; 2, moderate inflammation; and 3, severe inflammation. Serum titers of IgG anti-dsDNA Abs were measured as previously described using AP-conjugated goat anti-mouse IgG (BD Biosciences). The assay were calibrated using pooled serum from five MRL/lpr mice (19 wk old), which was arbitrarily assigned a value of 1500 U/ml at a 1/100 dilution; all readings were related to that using appropriate 1/2 serial dilutions starting from 1/100. For analysis of anti-dsDNA IgG isotypes, anti-mouse IgG1-AP and IgG2a-AP were used.
Assessment of skin disease
All mice were free from skin disease at the start of the experiments. Skin disease was scored as follows: 0, no rash; 1, beginning of rash but little or no hair loss; 2, discrete hairless rash less than ∼10 mm in length; 3, extensive rash more than ∼20 mm in length; and 4, extensive rash that includes the ears.
Flow cytometric analyses and intracellular staining
All staining profiles were based on live, gated cells, as determined by 4′,6-diamidino-2-phenylindole dilactate (DAPI; Sigma-Aldrich).
Mouse cytokine detection
Staining for flow cytometric analyses were performed on single-cell suspensions in 96-well U-bottom plates. Cells (5 × 105) were stimulated for 48 h with combinations of agonistic anti-CD40 Ab (5 μg/ml), anti-CD40 isotype control (5 μg/ml), and anti-IgM (F(ab′)2; 10 μg/ml), CpG (oligodeoxynucleotide 1826, at 25 μg/ml), LPS (Escherichia coli; 10 μg/ml) for B cells, or 1 μg/ml anti-CD3 for T cells/B-T cocultures. GolgiStop (BD Biosciences), PMA (50 ng/ml; Sigma-Aldrich), and ionomycin (250 ng/ml; Sigma-Aldrich) were added for the last 6 h of culture. To detect surface Ags, cells were washed with PBS/FCS/azide and then stained with combinations of CD19-PE-Cy7, CD4-FITC, CD21-FITC, CD21-biotin, CD23-PE, and CD24-biotin. Cells were fixed and permeabilized by incubation in Cytofix™ (BD Biosciences) for 20 min at 4°C, followed by washing in 1× Perm Wash (BD Biosciences) at 4°C. Permeabilized cells were incubated in 1× Perm Wash with anti-mouse IFN-γ-allophycocyanin, TNF-α-allophycocyanin, IL-10-allophycocyanin, or appropriate allophycocyanin-conjugated isotype controls (BD Biosciences). To demonstrate the specificity of staining, fixed/permeabilized cells were incubated with the excess of unlabeled anti-IFN-γ or anti-IL-10 (BD Biosciences) before incubation with allophycocyanin anti-IFN-γ or allophycocyanin anti-IL 10 or allophycocyanin anti-TNF-α. The cells were acquired by FACS scan LSR (BD Biosciences) and analyzed using FlowJo version 5.7.1 software (Tree Star). Alternatively, supernatants were collected before the addition of GolgiPlug and IL-10, IFN-γ, and TNF-α levels were measured by ELISA (BD Biosciences), according to the manufacturer’s instructions.
Cell sorting and transfer experiments
For the purification of mouse B cells, single-cell suspensions from spleens isolated from 8- to 10-wk-old MRL/lpr mice were incubated for 15 min at 4°C with 10 μl of anti-CD43 magnetic beads/107 cells for negative selection or CD19+ magnetic beads for positive selection (Miltenyi Biotec) per 90 μl of cell suspension. The cells were washed twice to remove unbound beads resuspended in 500 μl of MACS buffer (PBS, 0.5% FCS, and 2 mM EDTA) and purified using a MACS system. FACS staining for the B cell marker CD19 was used to control cell purity. This procedure normally yielded B cell preparations which were >95% CD19+. Purified B cells were cultured in complete RPMI 1640 medium (11) (Sigma-Aldrich) for 48 h with combinations of agonistic anti-CD40 Ab (5 μg/ml), anti-CD40/isotype control (5 μg/ml), and anti-IgM (F(ab′)2) (10 μg/ml). After 48 h of incubation, B cells were stained with anti-mouse anti-CD19-Cy7, anti-CD21-FITC, anti-CD23-PE, anti-CD24-biotin, and DAPI and sorted by a MoFlo cell sorter (DakoCytomation) or a BD Biosciences FACSAria. Cells (5 × 105) of a given sorted subset were injected into the tail veins of 9- to 10-wk-old MRL/lpr once a week for a total of 3 wk. Alternatively, T2 or marginal zone (MZ) B cell subsets were first sorted by a MoFlo cell sorter (DakoCytomation) or a BD Biosciences FACSAria and then stimulated with anti-CD40 (5 μg/ml) or with an isotype control (5 μg/ml) for 48 h. For the purification of mouse CD4+CD25−T cells, splenic CD4+ T cells were isolated from 8- to 10-w- old MRL/lpr mice by negative selection using a CD4 T cell isolation kit and passing through a MACS LS column (Miltenyi Biotec) as previously described (1). The purity was checked after purification and reported to be 98% (data not shown).
In vivo treatment of MRL/lpr mice with anti-CD40
Nine- to 10-wk-old MRL/lpr mice were given a daily i.p. injection of ∼300 μg/day anti-CD40 (FGK-45) in PBS for 2 wk. Following a 2-wk break, the treatment was repeated for another 2 wk. A control group of nine 10-wk-old MRL/lpr mice received isotype control over the same time periods. Disease progression was monitored until the majority of control mice had succumbed to disease.
Five × 105 cells/well were incubated in triplicate in 96-well plates coated with 2.5 μg/ml anti-mouse anti-CD3. After 72 h in culture, cells were pulsed with 1μCi of [3H]thymidine for the remaining 18 h of culture. Proliferation was measured using a liquid scintillation counter.
For the statistical analysis of the data, the Mann-Whitney U test and the Fisher exact test were applied to analyze clinical results. Unpaired t tests were applied on cytokine quantification experiments. A value of p < 0.05 was considered significantly different.
Numerical and functional impairments of Bregs in MRL/lpr mice
The low frequency of IL-10-producing T2 B cells represents the major obstacle to empowering this subset for the treatment of autoimmunity. The numbers of T2 and MZ B cells (gated as previously shown (6, 12)) were measured in the spleens of MRL/lpr mice at different ages. The results in Fig. 1,A show a significant reduction in the absolute numbers of T2 B cells with the progressing age but an increase in MZ B cells. The significant inverse correlation between proteinuria and the absolute numbers of T2 B cells further supports a relationship between disease progression and reduction of T2 B cell numbers (Fig. 1,B). MRL/Mp mice, which also develop lupus-like disease (although less rapidly than MRL/lpr mice), display an increased ratio of MZ:T2 B cells compared with non-prone C57BL/6 mice (Fig. 1, C and D). However, the numbers and percentages of MZ and T2 B cells did not change with age in C57BL/6 and MRL/Mp mice, suggesting that a decrease in T2 B cell numbers is a feature exclusive to MRL/lpr mice.
We next assessed the suppressive capacity that T2 B cells isolated from MRL/lpr mice have in an adoptive transfer system. Purified T2 and MZ (as control) B cells were transferred to syngeneic mice. Proteinuria levels and the survival analysis showed no significant differences in mice treated with either B cell subset compared with the control group (Fig. 1, E and F, and supplemental Fig. S1A5 for example of purity postsort). Therefore, these results suggest that T2 Bregs in MRL/lpr mice are functionally impaired.
Anti-CD40 stimulation induces IL-10-producing T2-like B cells
We addressed whether splenic B cells, isolated from MRL/lpr mice, cultured with agonistic anti-CD40 with or without anti-IgM (F(ab′)2) (as a surrogate autoantigen) would expand IL-10-producing B cells. Supernatants collected from B cells stimulated with anti-CD40 contained significantly higher levels of IL-10 compared with supernatants from isotype- and IgM-treated B cells (Fig. 2,A). Simultaneous engagement of BCR and anti-CD40 lead to a significant reduction of IL-10 production compared with B cells stimulated with anti-CD40 alone. Furthermore, intracellular staining revealed that the majority of the IL-10-producing B cells were contained within the T2 B cell subset (Fig. 2,B and supplemental Fig. S1B). Further analysis of ex vivo CD40-stimulated T2 B cells demonstrated that they express high levels of CD1d, previously identified as a Breg marker (2, 5), and CD93, which in the spleen is expressed exclusively on immature B cells (Fig. 2 C). However, since stimulation with anti-CD40 is known to modulate the expression of CD23 on different B cell subsets (Ref. 13 and supplemental data Fig. S2A and B), hereafter T2 generated after stimulation with anti-CD40 will be referred to as T2-like to distinguished them from unstimulated immature T2 B cells.
No differences in the IL-10 production were found if T2 B cells were sorted first and directly stimulated with anti-CD40, (as opposed to prestimulation of all B cells followed by sorting) (Fig. 2,D). Interestingly, CpG and LPS, which have been previously shown to up-regulates IL-10-producing B cells (14), do not have the same selective effect as anti-CD40 on purified T2, but induce the differentiation of equal frequencies of IL-10-producing MZ and T2 B cells (Fig. 2 E). Hence, anti-CD40 stimulation induces T2-like B cells that make IL-10. Of interest, in contrast to BCR stimulation, anti-CD40 stimulation prevented the spontaneous differentiation of sorted T2 B cells into a “mature” B cells (supplemental Fig. S3). Since anti-CD40 stimulation of purified B cells yielded higher numbers of viable T2 B cells (compared with anti-CD40 stimulation of purified T2-like B cells), we have used this protocol for the remainder of the experiments.
T2-like B cells suppress Th1 responses and convey suppressive capacity to CD4+CD25− T cells
To assess whether T2-like B cells have acquired suppressive function, T2, MZ, or follicular (FO)-like B cells were cultured with anti-CD3-stimulated CD4+CD25− T cells. Only T2-like, but not FO- or MZ-like, or isotype-treated B cells significantly inhibited the production of the IFN-γ and TNF-α by CD4+CD25−T cells (Fig. 3, A and B, and supplemental Fig. S4 for dose response). No IL-4 was detectable in any of the tested conditions (data not shown). The suppressive capacity of T2-like B cells was comparable to the inhibitory capacity exerted by Tregs (Fig. 3, C and D). In addition, only T2-like B cells induced the differentiation of IL-10-producing T cells (Fig. 3,E). The majority of the IL-10- producing T cells following coculture with T2-like B cells were FoxP3− (∼90% of the IL-10+CD4+T cells are FoxP3− while ∼10% are FoxP3+), and there was no overall increase in FoxP3 expression under the same conditions (Fig. 3, F and G). Despite their capacity to suppress the release of Th1-like cytokine, no inhibition of T cell proliferation was detected (data not shown).
The generation of newly formed IL-10-producing T cells could be a mechanism by which T2-like B cells might control inflammation in vivo. Anti-CD3-stimulated CD4+CD25− T cells were cultured 1:1 with T2-like B cells for an initial 72 h (first stimulation). These CD4+T cells were then purified, stained with CFSE (CFSE+CD4+), and cultured 1:1 with freshly isolated syngeneic CD4+CD25− T cells (CD4+ control (CTRL)) for an additional 2 days (second stimulation). The result in Fig. 3 H showed that only CFSE+CD4+ (CD4+ T cells which have been in culture with T2-like B cells) have acquired suppressive function and inhibit the production of TNF-α by “fresh” CTRL CD4+ T cells in the second round of stimulation. Neither CD4+ T cells derived from CD4+ T:FO cell cultures, CD4+ T:MZ cell cultures, nor T cells purified from isotype control-treated T2 B cells acquired suppressive activity (data not shown). The lack of expression of FoxP3, the capacity to produce IL-10, and the ability to suppress CD4+CD25− T cells are hallmarks of Tr1 cells (15), thus demonstrating that T2-like B cells suppress Th1 differentiation and induce the differentiation of Tr1 cells.
T2-like B cells control the progression of lupus, suppress T cell proliferation and differentiation into Th1 cells
Our findings raised the possibility that ex vivo expansion with anti-CD40 might empower T2-like B cells with regulatory capacity and allow them to suppress disease. Five × 105 T2, MZ, and FO-like B cells were purified by flow cytometry (purity in supplemental Fig. S1) and transferred to 9-to 10-wk-old MRL/lpr mice. Only transfer of T2-like B cells significantly decreased the mortality of recipient mice. At 25 wk of age, 100% of the control and >70% of MZ-like and FO-like B cell-treated mice had succumbed, while only 33% of T2-like B cell-treated mice had died (Fig. 4,A). This decrease in mortality was mirrored by lower levels of proteinuria. At 3 mo of age, 71% of the PBS-treated mice or mice transferred with MZ-like B cells had proteinuria levels higher than 300 mg/dl compared with only 25% in the T2-like B cell-treated group (Fig. 4,B). Histological examination of the kidneys from 23-wk-old mice receiving T2-like B cells revealed reduced levels of Ig deposition, minimal glomerular hypercellularity, and interstitial infiltration and showed a relative preservation of structure. In contrast, kidneys from control or from mice treated with MZ or FO-like B cells demonstrated the classical severe histological picture that characterizes this strain of mice (Fig. 4, C–E). IgG Ab levels against dsDNA (Fig. 4,F) and against anti-Sm (data not shown) were also lower in T2-like-treated mice. Interestingly, in the T2-like B cell-treated group, the isotype of anti-dsDNA switched from IgG2a to IgG1, suggesting a skew from a pathogenic Th1 response to a more favorable Th2 response (Fig. 4 F). A similar suppressive capacity was observed if sorted T2 B cells were stimulated directly with anti-CD40 and then transferred to syngeneic mice, whereas no suppression was observed by anti-CD40-stimulated MZ B cells (supplemental Fig S5).
The total numbers of CD4+, CD8+ T, and double-negative T cells were significantly reduced in mice given T2-like B cells compared with the other groups (supplemental Fig. S6). Splenocytes from the T2-like B cell-treated group also showed a weaker response to anti-CD3 stimulation than other groups. This defective responsiveness, along with a significant reduction in IFN-γ production, was mirrored by an increase in IL-10 production (Fig. 4 G) by CD4+ T cells (data not shown). Thus, transfer of T2-like B cells inhibits T cell proliferation and Th1 responses in favor of an anti-inflammatory IL-10 response.
The protective effect of T2-like B cell transfer is IL-10 dependent
To assess the requirement of IL-10 for the protective effect observed in vivo, T2-like B cells were transferred to syngeneic mice that were treated with isotype control or with anti-IL10R/anti-IL10 mAbs (16). A group of mice was treated with anti-IL10R/anti-IL10 mAbs alone. One hundred percent of mice treated with T2-like B cells were still alive at 20 wk of age, compared with only 20% in the group treated with T2-like B cells and anti-IL-10R/anti-IL-10 or in the control group (Fig. 5, A and B). Whereas T2-like-treated mice produced significantly less total IgG and IgG2a anti-dsDNA Ab (Fig. 5, C and D), neutralization of IL-10 restored the levels of total anti-dsDNA and the IgG2a Abs to the same level as the control group. Interestingly, previous work in the CIA model has shown that blockade of IL-10/IL-10R at an early stage of disease development does not affect disease course (17). Similarly, in MRL/lpr mice, the short period of blockade of IL-10, while sufficient to inhibit the immune-regulatory cascade initiated by Bregs, did not exacerbate the disease.
Unlike neutralization of IL-10, blocking TGF-β did not inhibit the suppressive capacity of T2-like B cells as shown by the unchanged levels proteinuria levels or mortality rate among the different groups (Fig. 5, E and F).
In vivo tracking of the developmental kinetics of T2-like B cells
To ascertain where T2-like Bregs exert their suppressive activity, anti-CD40-stimulated splenic hCD20 Tg B cells (10) were transferred to wild-type MRL/lpr mice. Analysis of the B cell subsets at day 1 after transfer showed that hCD20 Tg B cells homed to the spleen and maintained a T2:MZ:FO B cell ratio similar to the endogenous ratios measured in the wild-type mice (Fig. 6,A). Seven days after transfer, FO B cells had migrated from the spleen into the lymph node (LN; Fig. 6 B). In agreement with the nonrecirculating nature of “conventional” T2 and MZ B cells, hCD20+ Tg T2 and MZ-like B cells were recovered prevalently from the spleen, suggesting that stimulation via CD40 does not alter their migratory capacity. No hCD20+ Tg B cells were found in the kidneys of the recipient mice (data not shown).
Our in vitro experiments suggest that CD40 engagement on T2 B cells alters their maturation. To assess T2-like maturation in vivo, we transferred sorted hCD20 T2-like or isotype-treated T2 B cells to wild-type mice. Seven days after transfer, spleens and draining LN were harvested. Virtually all recovered hCD20+ T2 B cells stimulated in vitro with isotype control had differentiated into FO or MZ B cells (Fig. 6,C). In contrast, only 40% of transferred T2-like B cells had differentiated into mature FO B cells with ∼40% of the T2-like B cells retaining their original “immature-like” phenotype (Fig. 6,D). This was highly consistent over different experiments. Due to their very low frequency, we have not been able to reisolate undifferentiated transferred T2-like B cells and compare their suppressive function with those that have differentiate into mature B cells. Since the transfer of mature FO B cells failed to protect recipient mice from disease and, in vitro, FO B cells did not inhibit T cell inflammatory cytokine production (Fig. 3, A and B), it is plausible to hypothesize that disease suppression is mediated by T2-like Bregs retaining their undifferentiated phenotype.
In vivo administration of anti-CD40 ameliorates lupus activity in MRL/lpr mice and increases T2 B cell numbers
Agonistic mAbs to CD40 (CD40 mAbs) have a puzzling dual therapeutic effect in experimental animal models. CD40 mAbs induce tumor regression by potentiating antitumoral T cell responses (18), yet we have previously shown that anti-CD40 exert immunosuppressive activity in chronic autoimmune inflammatory processes (19). Our results showing that the transfer of adaptive T2-like Bregs protect MRL/lpr mice from disease suggested that T2 Bregs could be more sensitive to agonistic CD40 stimulation in vivo than other B cell subsets.
MRL/lpr mice were treated for 4 wk with 300 μg/day anti-CD40 i.p. or with isotype control. Anti-CD40-treated mice survive significantly better than control-treated mice (p = 0.0461; Fig. 7,A), displayed significantly lower proteinuria than controls (p = 0.0145; Fig. 7,B), suffered less severe skin disease (p < 0.05; Fig. 7,C), and displayed a significantly reduced amount of anti-dsDNA IgG (Fig. 7,D). Importantly, analysis of B cell subsets 2 wk after the beginning of the treatment showed a significant increase in T2 B cell numbers (Fig. 7, E and F). Identical suppressive results were also obtained using a different agonistic anti-CD40 (3/23; data not shown), suggesting that the suppressive effect is not due to epitope specificity of the FGK45 Ab. To ascertain the effect of anti-CD40 treatment on different cell populations in the spleens of treated mice compared with the control group, mice were treated for 2 wk and splenocytes were collected at different time to establish the kinetics of cytokine production. The results in Fig. 7,G show a significant increase in the frequency of CD19+IL-10+ cells over the treatment period compared with the control-treated group. In contrast, CD40+CD19− cells appeared to remain unaffected by the treatment and produced an equivalent amount of proinflammatory cytokines at each time point (Fig. 7 H). These results provide additional evidence that anti-CD40 mAb therapy targets T2 Bregs and provides initial support for the notion that anti-CD40 treatment, similarly to rituximab, could be used in treatment for autoimmune diseases.
Adoptive cellular therapy for the control of autoimmune disorders has seen a surge of interest due to the feasibility of manipulating Tregs in vitro (20). Even though it is now well established that Bregs, like Tregs, are directly involved in the maintenance of tolerance, no strategies for expanding them in vitro have been identified. In this study, we demonstrate that in vitro Bregs suppress Th1 differentiation, convey suppressive capacity to CD4+T cells, and induce the differentiation of CD4+FoxP3−IL-10+ T cells. In vivo, anti-CD40-generated T2-like Bregs reverse lupus-like disease and induce long-term tolerance via the inhibition of Th1 responses and T cell proliferation. It is interesting to observe that suppression of T cell proliferative responses was only observed after Breg transfer in vivo. These could be due to the initiation of a complex suppressive cascade reaction in vivo, which is unlikely to take place in vitro. The in vivo neutralization of IL-10 supports the hypothesis that IL-10 produced by the T2-like Bregs is responsible for the regulation of lupus-like disease in MRL/lpr mice. The role of Th17 in lupus, unlike in other autoimmune diseases, is not well understood (21). Since no IL-17-producing T cells were detected in the spleens of MRL/lpr mice, in the same experimental condition found to be optimal for T2-like Breg-mediated Th1 suppression, it is at the moment difficult to conclusively surmise whether T2-like Bregs suppress Th17 as well as Th1 differentiation. Our results are interesting in the context of the recent finding showing that the in vivo production of IL-10 induced by dendritic cells can lead to the differentiation of Tr1 (22). Our data, along with those of others (17), suggest that Bregs should also be included in the pool of cells responsible for the differentiation of Tr1. It will be interesting in the future to assess whether chimeric mice lacking IL-10 exclusively on B cells, previously shown to lack Bregs (3), develop a normal pool of suppressive Tr1 cells. Unfortunately, the previously reported IL-10-deficient MRL/lpr mice known to develop an exacerbated disease compared with wild-type mice (23) are no longer available.
CD40-CD154 signaling is important for the differentiation of plasma cells and the production of autoantibodies; however, the same costimulatory pathway, depending on the density of CD154 expression, has also been implicated in preventing Ab production (24) and in inhibiting the development of chronic arthritis (19). In this study, we demonstrate that CD40 ligation halts the maturation of T2 B cells (supplemental Fig. S3 and Fig. 6) and induces the differentiation of IL-10+ T2-like B cells. Our data (supplemental Fig. S7) are also in agreement with previous studies showing that CD40 ligation rescues B cells and transitional B cell subsets from apoptosis (25, 26) and prevents their further differentiation into mature FO B cells (27). From our data, it appears that anti-CD40 specifically targets T2 B cells and induces the differentiation of IL-10-producing T2 Bregs, whereas anti-CD40/anti-IgM (F(ab′)2) stimulation induces the differentiation of TNF-α- and IFN-γ-producing mature B cells (data not shown), that may contribute to the pathogenesis of lupus disease. We speculate that the balance between Ag exposure (i.e., BCR cross-linking) and the strength of CD40 engagement might not only influence B cell maturation (e.g., FO vs MZ differentiation) but could also affect the outcome of the immune response (activation vs regulation) (4).
Other B cell subsets have been associated with IL-10 production and immunoregulation (14, 28, 29). Neonatal CD5+ B cells, responsible for the lack of inflammatory responses in neonates, have been shown to secrete IL-10 upon TLR9 receptor stimulation and to control the priming capacity of neonatal dendritic cells. Moreover, adult CD5+CD1dhigh B cells have been shown to possess regulatory function in contact hypersensitivity responses (5, 29). Gray et al. (14) have shown that MZ B cells produce higher levels of IL-10 than FO B cells, but made no comparison between T2 and MZ B cells and their capacities to produce IL-10 and suppressive capacity (14). We have analyzed the IL-10 production by purified B cells in response to CpG, LPS, and anti-CD40 and show that whereas MZ produce more IL-10 in response to CpG than FO B cells and equal amounts to T2-like B cells, the highest amount of IL-10 was observed when T2 B cells were stimulated with anti-CD40 (Fig. 2 E). In addition, even though MZ B cells produce some IL-10 in response to anti-CD40 stimulation, given that in our model MZ-like B cells failed to confer protection, we have excluded them as possible Breg candidates.
T2-like B cells retain all of the archetypical characteristics of the T2 Bregs previously identified in the CIA model (6), including a CD23highCD21highCD24highCD93CD1dhigh phenotype (12, 30, 31, 32, 33). Nevertheless, markers like CD19, CD1d, CD21, and CD5 are consistently expressed by Bregs in different models (4, 5), suggesting the possible existence of a unique Breg lineage common to the different experimental models. It is feasible that the modulation of other markers expressed by Bregs (i.e., CD5, CD23) may be up- or down-regulated according to environmental control (i.e., cytokines or costimulatory signals provided by activated CD4+T cells). In this context, our data have shown that stimulation with anti-CD40 up-regulates the expression of CD23 on MZ B cells (supplemental Fig. S2B). Therefore, Bregs generated after anti-CD40 stimulation, despite being phenotypically indistinguishable from conventional unmanipulated T2 B cells, might contain B cells originating from either MZ of FO B cells. Nevertheless, the data showing that the progression of spontaneous lupus-like disease in MRL/lpr mice correlates with a decrease in the absolute numbers and percentages of T2 and an increase in MZ B cell numbers, further support the existence of Bregs within the natural T2 B cells in autoimmunity.
Taken together with the data showing that immature B cells are the repopulating subset after depleting anti-CD20 treatment both in rheumatoid arthritis patients (34, 35) and in the NOD mouse model (36), our observation that in vivo treatment with anti-CD40 increases IL-10-producing B cells, but has a negligible effect on other CD40+cells, and expands T2 Bregs has important implications for clinical therapy. We could envisage a scenario where human immature B cells, also enriched of Bregs (P. A. Blair and C. Mauri, unpublished results) could be isolated from rituximab-treated systemic lupus erythematosus patients and expanded ex vivo with anti-CD40 and readministered at the time of reflaring as an alternative to further B cell depletion. This could be coupled with lower amounts of immunosuppressant. These strategies could control the pathogenic response while reestablishing an enduring homeostatic balance.
We thank the staff of the Biological Service Facility for animal husbandry.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the Wellcome Trust (Grant 068629 to C.M.), the University College of London Hospital charities Clinical Research Development Committee (Grant G140 to C.M.) and equipment Arthritis Research Campaign Grant ID 17746. P.A.B. is supported by the Oliver Bird Rheumatism Programme (Grant RHE/001124/G to D.A.I.). K.A.C.R. is supported by the Instituto Mexicano Del Seguro Social. M.J.S. was supported by National Institutes of Health Grant R01AR044077.
Abbreviations used in this paper: Breg, regulatory B cell; CIA, collagen-induced arthritis; Treg, regulatory T cell; hCD20, human CD20; AP, alkaline phosphatase; DAPI, 4′,6-diamidino-2-phenylindole dihydrochloride; MZ, marginal zone; FO, follicular; LN, lymph node.
The online version of this article contains supplemental material.