Interleukin-10 From Marginal Zone Precursor B-Cell Subset Is Required for Costimulatory Blockade-Induced Transplantation Tolerance

Abstract

Background 

Blocking CD40-CD40L costimulatory signals induces transplantation tolerance. Although B-cell depletion prevents alloantibody formation, nonhumoral functions of B cells in tolerance have not been well characterized. We investigated whether specific subsets of B cell or B cell–derived interleukin (IL)-10 contribute to tolerance.

Methods 

Wild type C57BL/6, or B cell–specific interleukin (IL)-10^−/− (CD19-Cre^+/−::IL-10^fl/fl) mice, received vascularized BALB/c cardiac allografts. BALB/c donor-specific splenocyte transfusion and anti-CD40L monoclonal antibody were used as tolerogen. B cells were depleted with antimouse CD20 monoclonal antibody. Various B-cell subsets were purified and characterized by flow cytometry, reverse transcription polymerase chain reaction, and adoptive transfer.

Results 

B-cell depletion prevented costimulatory blockade-induced allogeneic tolerance. Costimulatory blockade increased IL-10 in marginal zone precursor (MZP) B cells, but not other subsets. In particular, costimulatory blockade did not change other previously defined regulatory B-cell subsets (Breg), including CD5⁺CD1d^hi Breg or expression of TIM1 or TIM4 on these Breg or other Breg cell subsets. Costimulatory blockade also induced IL-21R expression in MZP B cells, and IL-21R⁺ MZP B cells expressed even more IL-10. B-cell depletion or IL-10 deficiency in B cells prevented tolerance in a cardiac allograft model, resulting in rapid acute cardiac allograft rejection. Adoptive transfer of wild type MZP B cells but not other subsets to B cell–specific IL-10 deficient mice prevented graft rejection.

Conclusions 

CD40 costimulatory blockade induces MZP B cell IL-10 which is necessary for tolerance. These observations have implications for understanding tolerance induction and how B cell depletion may prevent tolerance.

In Brief

Anti-CD40L mAb blockade can induce tolerance, the mechanism of which has been traditionally related to effector T cell inhibition and Treg induction. The current study shows that CD40L blockade also induces a critical subset of marginal zone precursor B cells which produce IL-10 required for tolerization.

Many T-cell costimulatory receptor-ligand interactions have been identified (CD28-CD80, CD28-CD86, CTLA-4-ICOS, CD27-CD70, CD134-OX40L and CD40L-CD40), and costimulatory blockade has been used to induce tolerance in murine as well as in nonhuman primate models.¹ In particular, blockade of CD40-CD40L suppresses alloimmunity and induces long-term tolerance to skin, islet, bone marrow, heart, kidney, myoblast, and limb allografts.¹ CD40 is expressed on B cells, dendritic cell, macrophages, epithelial cells, hematopoietic progenitors, and activated T cells; whereas CD40L (CD154) is expressed on activated T cells, activated B cells, and activated platelets.² During inflammation, peripheral blood monocytes, human vascular endothelial cells, smooth muscle cells, and mononuclear phagocytes may also express CD40L.² Costimulatory blockade-induced tolerance can be potentiated through administration of alloantigen, such as donor-specific splenocyte transfusion (DST), to induce peripheral tolerance to alloantigen.³ It has been proposed that CD40-CD40L blockade induces peripheral tolerance by inhibiting antigen presenting cells (APC) maturation, T-cell activation, and alloantibody and autoantibody production while promoting the generation of regulatory T cells.¹ Based on these observations, some investigators have shown that B-cell depletion also partially inhibits alloantigen presentation and alloantibody production, thereby promoting graft survival.^4,5 In contrast, others have found evidence that B cells may promote graft survival or tolerance.^6-8 The role of B cells in costimulatory blockade-induced transplantation tolerance is not fully understood.

B-cell functions include antibody production, antigen presentation to T cells, secretion of proinflammatory and anti-inflammatory cytokines, help for T-cell repertoire development and maintenance, and lymphoid organogenesis. Alloantibodies produced by B cells are clearly involved in the pathogenesis of graft rejection, and depletion of B cells has been suggested as a therapeutic approach to prevent or treat rejection.⁹ However, there are additional ways B cells may influence tolerance. (i) B cells can tolerize antigen-specific CD8⁺ T cells directly via CD95-mediated activation induced deletion.¹⁰ (ii) Activated B cells presenting antigen via MHC class I can induce anergy in CD8⁺ T cells.¹¹ (iii) B cells help in the induction of Foxp3⁺ regulatory T cells (Treg).¹² (iv) Activated B cells with increased surface expression of B7-2 inhibit proliferation of self-reactive CD4⁺ T cells in a CD40-CD40L–dependent manner.¹³ (v) B cells control the antigen presenting function of DCs in a cytokine-dependent manner, increasing tolerogenic responses.¹⁴ (vi) B cell secretion of immunoglobulin (Ig)G linked with latent transforming growth factor-β (IgG-transforming growth factor-β) inhibits CTL function in an antigen nonspecific manner.¹⁵ (vii) A subset of interleukin (IL)-10–producing CD1d^hiCD5⁺ B cells in mice^16,17 and CD19⁺CD24⁺CD38⁺ B cells in humans¹⁸ has protective function in autoimmune diseases.¹⁹ However, how the nonhumoral functions of B cells contribute to the generation of costimulatory blockade-induced alloantigen-specific tolerance is not known.

In the present study, we showed that depletion of B cells inhibited the development of costimulatory blockade-induced transplantation tolerance, leading to acute cellular rejection of allogeneic cardiac allografts. Costimulatory blockade specifically induced greater numbers of interleukin (IL)-10⁺ marginal zone precursor (MZP) cells, but not other putatively tolerogenic regulatory B cell (Breg) subsets, and IL-21R⁺ MZP B cells produced even more IL-10. The MZP B cells were required for tolerance, and transfer of this subset specifically restored graft survival.

MATERIALS AND METHODS

Mice

BALB/c (H-2^d), C57BL/6 (H-2^b), and CD19-cre (H-2^b) mice 8 to 10 weeks old were purchased from The Jackson Laboratory. Human CD20 transgenic (hCD20Tg; H-2^d) mice were from Dr. Mark I Shlomchik (Yale University, New Haven, CT). The IL-10^fl/fl (H-2^b) mice were from Dr. Christopher Karp (Cincinnati Children's Hospital Research Foundation, Cincinnati, OH). CD19Cre^+/−:IL-10^fl/fl (B-IL-10^−/−) mice were generated by breeding CD19-Cre^+/+ with IL-10^fl/fl mice. All mice were housed in a specific pathogen-free facility in microisolator cages. All experiments used age- and sex-matched mice in accordance with protocols approved by the Institutional Animal Care and Utilization Committee.

Antibodies and Reagents

Antimouse CD4 (GK1.5), antimouse CD8 (53.67), antimouse B220 (RA3-6B2), PE-antimouse CD19 (MB19-1), APC antimouse B220 (RA3-6B2), APC-Cy7 antimouse CD11c (N418), APC-Cy7 antimouse CD1d (1B1), PE-Cy7 antimouse CD5 (53–7.3), PE antimouse GR-1 (RB6-8C5), Pacific Blue antimouse CD45 (30 F-11), PE-antimouse IgM (II/41), PE antimouse CD11b (M1/70), PE antimouse CD8 (53–6.7), APC antimouse CD4 (GK1.5), FITC antimouse CD25 (PC61.5), biotin antimouse TIM-1 (RMT1-4), PE-antimouse TIM-4 (RMT4-54), PE antimouse/rat Foxp3 (FKJ-16s) antibodies, and isotype control antibodies were purchased from eBioscience (San Diego, CA). FITC antimouse CD93 (AA4.1), PE/Cy7 antimouse CD21/CD35 (7E9), PE-antimouse CD23 (B3-B4), Pacific Blue antimouse IgD (11-26c 2A), BV421 antimouse IL-10 (JES5-16E3), and biotin antimouse IgM (MRM-47) were purchased from Biolegend, (San Diego, CA). Rabbit antimouse Foxp3 polyclonal antibody and MOMA-1 were purchased from Abcam (San Francisco, CA). Antimouse CD20 monoclonal antibody (mAb) (clone 5D2, mouse IgG2a) was received from Genentech, Inc. (San Francisco, CA). Cy3-donkey antirat, FITC-donkey antirat, Cy5-donkey antirat, Cy3-goat antihamster, Cy5-antihamster, Cy5-goat antirabbit antibodies and fluorochrome conjugated streptavidin were purchased from Jackson Immunoresearch Laboratory, Inc. (West Grove, PA). Antimouse CD40L (MR-1) and control isotype antibodies were purchased from BioXcell (West Lebanon, NH). 4′,6-diamidino-2-phenylindole (DAPI) were purchased from Invitrogen (Carlsbad, CA). Antimouse CD19 (clone 1D3; American Type Culture Collection, Manassas, VA) and mouse antihuman CD20 mAb (clone 2H7; from Dr. Mark I Schlomchik, Yale University) were purified in our laboratory.

Cardiac Transplantation and Treatment Protocols

Transplantation tolerance was induced by administering 1 × 10⁷ BALB/c DST on day −7 before transplantation and anti-CD40L mAb (days −7, −4, 0 and +4; 250 μg/injection intravenously [IV]) to C57BL/6 recipients of BALB/c donor vascularized cardiac allografts.³ B cells were depleted by 1 IV injection (d +1; 100 μg/mouse) of antimouse CD20 mAb (5D2), antimouse CD19 mAb (1D3), or isotype control mAb in C57BL/6 mice or antihuman CD20 mAb (2H7) in hCD20Tg mice. Graft function was monitored every other day by abdominal palpation.

Isolation of Cells from Lymph Node, Spleen, and Grafts

Cardiac grafts were perfused with phosphate buffered saline (PBS), minced and digested with 1 mg/mL collagenase D (Roche, Indianapolis, IN) in Roswell Park Memorial Institute medium for 45 minutes at 37°C. Single cell suspensions from spleen, lymph nodes (LNs), and graft were made, and RBC were lysed using ammonium-chloride potassium lysis buffer (Lonza, Walkerville, MD).

Cell Staining and Flow Cytometric Analysis

Staining of cells was performed with the indicated antibodies with 1 μg/10⁶ cells at 4°C for 30 minutes. We purified MZP B cells using a 100-μm nozzle at an event rate of 10,000 to 12,000 cells/second using a FACS ARIA II sorter (BD Bioscience, Mountainview, CA). Trypan Blue staining of cells showed more than 98% viable cells, and purity was found to be more than 98% for each subset. Data were acquired using the FACSCantoII, LSR Fortessa, or LSRII flow cytometer (BD Biosciences) and analyzed using FlowJo software (Tree Star Inc., Ashland, OR).

Real-Time Quantitative Reverse Transcription Polymerase Chain Reaction

Cells were lysed in TRIzol reagent (Invitrogen), total RNA was purified, and cDNA was made using oligo d(T)_12–14 primer and Omniscript RT kit (Invitrogen) as per manufacturer's guidelines. Messenger RNA (mRNA) expression was quantified by CFX96 thermal cycler (Bio-Rad, Hercules, CA) or 7900 fast real-time polymerase chain reaction (PCR) (Applied Biosystem, Carlsbad, CA) using SYBR Green PCR kit (Qiagen, Valencia, CA). Polymerase chain reaction consisted of a 15-minute at 95°C denaturation step, followed by 40 cycles of 15 seconds at 94°C, 20 seconds at 56°C, and 20 seconds at 72°C. Relative mRNA expression of specific gene was calculated as: 2^{(Ct of cyclophilin-A - Ct of IL-10)}. The primers used for real-time PCR: mouse IL-10, forward 5′-GGGTTGCCAAGCCTTATCGGAAAT-3′ and reverse 5′-CCTTGATTTCTGGGCCATGCTTCT-3′; mouse cyclophilin-A, forward 5′-AGGGTGGTGACTTTACACGC-3′, and reverse 5′-ATCCAGCCATTCAGTCTTGG-3′.

Histopathological Analysis

Cardiac grafts, spleens, and LN were harvested, frozen directly in optimal cutting temperature compound (Sacura Finetek, Torrance, CA), and stored at −80°C. Eight-micron sections were cut with a Leica 1900CM cryomicrotome, fixed with chilled acetone, blocked with 2.5% normal horse serum (Vector Laboratories, Burlingame, CA), stained with the indicated primary antibodies for 30 minutes, stained with conjugated secondary reagents, and mounted with VectaShield mounting solution (Vector Laboratories) with or without DAPI. Images were acquired with a Leica fluorescence microscope (Leica Mikrosysteme, Vertrieb, Germany) and a digital Hamamatsu CCD camera (Hamamatsu Corporation, Bridgewater, NJ). Separate images were collected on the CY3, GFP, and DAPI channels, overlaid, and analyzed with Openlab software (Improvision, Lexington, MA) and Leica MMF software. Quantification of graft infiltrating cells was performed by counting 4 to 5 fields per tissue section and 3 to 4 sections per graft.

For hematoxylin-eosin (H&E) staining, PBS-buffered formalin fixed tissues were embedded in paraffin, sectioned at 5 μm and H&E staining performed. Parenchymal rejection (PR) score of graft was performed as described earlier.²⁰

Based on modified protocol published from International Society for Heart and Lung Transplantation, severity of histopathology of allograft can be accessed with H&E.^20-22 The PR scoring was performed as: 0, no rejection; 1, focal mononuclear cell infiltration without necrosis; 2, focal mononuclear cell infiltrates with necrosis; 3, multifocal infiltrates with necrosis; 4, widespread infiltrates with hemorrhage and/or vasculitis.

Luminex ELISA

Sera were collected and stored at −80°C. Cytokines were quantified using the Bio-Plex Pro Mouse Cytokine Th1/Th2 assay kit (Bio-Rad). Assays were performed according to the manufacturer's instructions and analyzed on a Bio-Plex 200 System (Bio-Rad).

Adoptive Transfer of MZP B Cells

Different subsets of B cells were purified from wild-type C57BL/6 or B-IL-10^−/− spleens using flow cytometry and gating on specific populations as shown in Figure S1A. Purity of B-cell subset was typically 90% to 99%. Sorted cells were used either for adoptive transfer or analysis of mRNA expression.

Intracellular Cytokine Staining

Single-cell suspension from spleen was prepared, RBCs removed by ammonium-chloride potassium lysis buffer treatment. Splenocytes (2 × 10⁶ cells/well) were stimulated with phorbol 12-myristate 13-acetate (PMA; 50 ng/mL), ionomycin (500 ng/mL) and monensin (2 μM) for 5 hours in 24-well plates in complete Roswell Park Memorial Institute medium containing 10% fetal bovine serum at 37°C in 5% CO₂ incubator. Cells were stained for surface markers, fixed with fixation buffer (Biolegend), and permeabilized with 1× permeabilization buffer (Biolegend). Intracellular IL-10 staining was performed, washed with PBS, and cells acquired by flow cytometry (FACS Canto, BD Bioscience). Data were analyzed using FlowJo software.

Statistics

Graft survivals were graphically expressed using the Kaplan-Meier method, and statistical differences were assessed by Log-rank (Mantel-Cox) test using Prism 6 software (GraphPad software, La Jolla, CA). Unpaired, 2-tailed Student t test was performed, and P less than 0.05 was considered statistically significant.

RESULTS

B-Cell Depletion Prevents Tolerance

To understand the function of the B cells in costimulatory blockade-induced tolerance, we investigated different B-cell–depleting strategies. A single IV injection of 100 μg of rat antimouse CD20 mAb (5D2) into C57BL/6 mice depleted approximately 75% of total CD19⁺ B cells in the spleen and LN and CD19⁺IgM⁺IgD⁺ B cells in bone marrow within 3 days (Figure 1A). This depletion persisted for at least 3 weeks (Figure 1B). Compared to isotype control mAb, even as little as 10 μg anti-CD20 depleted CD19⁺ B cells among total lymphocytes in spleen within 48 hours (Figure S1B). Depletion with 10 μg anti-CD20 mAb was distributed across all subsets of B cells in the spleen and LN (Figure S1B). Injection of 25 μg or 100 μg anti-CD20 mAb also uniformly depleted all B-cell subsets in the spleen and LN (Figure S1B, S1C, and data not shown). Depletion of B cells in presence of the allograft and tolerogen also gave a similar B-cell depletion profile (not shown). A single injection of antihuman CD20 mAb (2H7) in hCD20Tg mice also depleted B cells in spleen and LN (Figure S2A, SDC, https://links.lww.com/TP/B147). B-cell depletion did not induce significant changes in the numbers of CD11c⁺ DCs, plasmacytoid DCs, CD4⁺ T cells or Foxp3⁺ regulatory CD4 T cells in the spleen or LN (Figure S2B, SDC, https://links.lww.com/TP/B147).

FIGURE 1: B-cell depletion prevents tolerance. A, C57BL/6 mice IV injected with anti-mCD20 mAb (100 μg/mouse) or isotype control IgG. After 3 days, CD19⁺B220⁺ cells in the spleen, LN and thymus analyzed by flow cytometry. In bone marrow, IgM⁺IgD⁺ cells analyzed by gating on CD19⁺ cells. Based on FSC-A versus SSC-A profile, all the cells gated on the lymphocytic population. B, C57BL/6 mice IV injected with anti-mCD20 mAb (100 μg/mouse). CD19⁺B220⁺ cells as a percentage of total lymphocytes in spleen, LN and thymus analyzed by flow cytometry at different time points after gating on lymphocytic population. C, C57BL/6 recipient mice given tolerogen (DST, 1 × 10⁷ cells/mouse d-7 relative to transplantation; and anti-CD40L mAb 250 μg/mouse days −7, −4, 0, and +4) and received BALB/c cardiac allografts or C57BL/6 syngeneic grafts on day 0. Recipients given anti-mCD20 mAb (100 μg/mouse) day +1. Graft survival monitored (left). After 5 days, grafts harvested, sections made and stained with H&E, and parenchymal rejection scores were calculated (right). The H&E staining of graft (bottom). D, C57BL/6 recipients given tolerogen and BALB/c allografts as described above, and various doses of anti-mCD20 mAb on day +1. Graft survival plotted (left). After 5 days, grafts harvested, sections stained with H&E, and parenchymal rejection scores were calculated (right). E, Human CD20 transgenic (H-2^d) mice given tolerogen and C57BL/6 (H-2^b) cardiac allografts. Antihuman CD20 mAb (100 μg/mouse) given IV on day +1. Graft survival (left), Masson's trichrome staining and parenchymal rejection score calculated (middle), and H&E staining on day +5 (right).

To investigate the role of B cells in tolerance, C57BL/6 mice received DST day −7; anti-CD40L mAb days −7, −4, 0, +4; and BALB/c vascularized cardiac allografts on day 0. Depletion of B cells with antimouse CD20 mAb 1 day after transplantation prevented tolerance and lead to acute rejection with increased PR scores in the allograft, whereas B-cell depletion did not cause rejection of syngeneic grafts (Figure 1C). Dose–response analysis showed that as little as 10 μg of antimouse CD20 mAb-induced rejection and increased PR scores (Figure 1D). Similarly, hCD20Tg (H-2^d, BALB/c background) mice received the tolerogenic regimen (C57BL/6 DST and anti-CD40L mAb) and C57BL/6 vascularized grafts. Depletion of B cells with antihuman CD20 mAb in hCD20Tg mice also resulted in acute rejection (mean survival time, 8 days; n = 7 mice) (Figure 1E). Injection of antimouse CD19 mAb (1D3), which caused partial B cell depletion, also prevented tolerization of C57BL/6 recipients (mean survival time, 12.5 days; n = 8 mice) (data not shown). Together, these results demonstrated that B-cell depletion with different antibodies and in different strains prevented costimulatory blockade-induced tolerance and led to acute rejection.

B-Cell Depletion Induced Graft Rejection Is Not Due to Serum Cytokine Storm or Alloantibodies

To test if depletion of B cell lead to strong inflammatory cytokine production, we administered tolerogen, transplanted allografts, and depleted B cells as above. Five days after transplantation, various cytokines were analyzed in the serum. The results showed no significant change in almost all inflammatory or anti-inflammatory cytokines tested (Figure 2A). There was reduction in serum IL-6 after B-cell depletion compared to control IgG-treated animals, which would not be expected to impair graft survival because IL-6 promotes Th17 and interferes with Treg induction.²³

FIGURE 2: B cell depletion does not cause serum cytokine storm or alloantibody responses. A, C57BL/6 recipients given tolerogen and BALB/c allografts or C57BL/6 syngeneic grafts on day 0, and anti-mCD20 mAb (100 μg/mouse) day +1. Five days after transplantation, sera were collected, and cytokine levels were quantitated by Luminex bead assay. B, Sera from animals in Figure 1C were collected 5 days after transplantation, and alloantibody responses were measured by flow cytometry. Representative histogram (left) and MFI (right). n = 3–5 mice/group. MFI indicates mean fluorescence intensity.

We also measured the presence of alloantibody in B cell–depleted mice. The results showed that, as expected, B-cell depletion did not result in altered alloantibody levels compared to nondepleted mice (Figure 2B). Together, these results showed that B-cell depletion in presence of costimulatory blockade did not induce an inflammatory cytokine storm or alloantibody in the serum.

B-Cell Depletion Leads to Acute Cellular Rejection of Allograft

Because histopathological analysis of rejecting donor allografts showed mononuclear cell infiltration and increased PR scores in the allografts (Figure 1C and D), we further characterized the allograft-infiltrating cells. Flow cytometric analysis revealed a significantly increased leukocytic in filtration (193.8 ± 32.9 × 10⁴ vs 73.7 ± 18.7 × 10⁴ total leukocytes/heart, P = 0.019). Immunohistological analysis showed increased infiltration of GR1⁺ monocytes, metallophilic macrophages (MOMA)-1⁺ cells, and CD8⁺ T cells (Figure 3A). There were no significant changes in the number of CD4⁺ T cells, but rejecting grafts had decreased numbers of Foxp3⁺Treg compared to tolerated grafts (Figure 3B). No infiltration of B cells, as expected, or CD11c⁺ DCs was observed in the allograft (not shown). Flow cytometric analysis of MOMA-1⁺ cells showed a phenotype of FSC^hiSSC^hiB220⁻CD11c⁻CD11b⁺Gr1^int/lo (Figure S1C, SDC, https://links.lww.com/TP/B147), suggesting a monocytic lineage. Microscopic analysis of isolated cells showed increased cytoplasmic content indicative of activated macrophages (Figure S1D, SDC, https://links.lww.com/TP/B147). It has been reported that under inflammatory conditions CD11b⁺Gr1⁺ myeloid cells promote Th17-cell differentiation²⁴; and MOMA-1⁺ (CD169) macrophages cross-presents antigen from dead cells to CD8⁺ T cells,²⁵ suggesting that this population might contribute to alloantigen presentation and graft rejection. Together these results showed that B-cell depletion induced increased infiltration of CD8⁺ T cells and MOMA-1⁺ macrophages in the rejecting allografts.

FIGURE 3: B-cell depletion causes acute cellular rejection with increased infiltration in the allograft. A, C57BL/6 recipients given tolerogen, BALB/c allografts and anti-mCD20 mAb as in Figure 1C. Grafts harvested 5 days after transplantation. Immunohistochemistry for graft infiltrating cells. Magnification, 200× (left). Quantitative data (right). B, Immunohistochemistry and quantitation of graft infiltrating CD4⁺ and Foxp3⁺ cells. Magnification, 400×. 3 grafts/group, 3 sections/graft, 3 to 4 fields/section.

Costimulatory Blockade Does Not Change CD19⁺CD1d^hiCD5⁺ Breg

It has been shown that IL-10–producing CD1d^hiCD5⁺ B cells function as suppressor Breg to prevent the development of autoimmune diseases.¹⁶ Tolerogen treatment did not induce significant changes in the percentage of CD1d^hiCD5⁺ Breg cells in spleen (Figure 4A), LN, or bone marrow (data not shown). The expression of IL-10 in these cells did not change with tolerogen treatment (Figure 4B and C). In contrast, tolerogen increased IL-10 mRNA expression in non–Breg cells (B220⁺CD19⁺CD5⁻CD1d^lo B cells) (Figure 4B). It has been shown that TIM-1⁺ B cells produce IL-10 and help in the maintenance of tolerance.²⁶ TIM-1 and TIM-4 expression on CD19⁺CD5⁺CD1d^hi B cells was detected; however, their expression did not change after tolerogenic treatment in Breg or non–Breg cells (Figure 4D and E). Thus, the tolerogenic regimen did not change the percentage of CD5⁺CD1d^hi Breg nor their expression of molecules known for their suppressive function.

FIGURE 4: Costimulatory blockade does not increase the CD19⁺CD5⁺CD1d^hi Breg. A, C57BL/6 given no treatment (naive), DST, anti-CD40L mAb, or DST + anti-CD40L mAb; and after 5 days frequency of CD5⁺CD1d^hi B cells among all lymphocytes in the spleen analyzed after gating on B220⁺CD19⁺ cells (upper). Mean percentage of CD5⁺CD1d^hi cells among all CD19⁺ cells (lower). Five mice/group. B, C57BL/6 mice given tolerogen, and after 5 days, B220⁺CD19⁺CD5⁺CD1d⁺ Breg and B220⁺CD19⁺CD5⁺CD1d⁻ non–Breg cells purified from spleen, and IL-10 mRNA expression assessed by qRT-PCR. C, Splenocytes were stimulated with PMA and ionomycin, intracellular IL-10 stained, and analyzed by gating on CD19⁺CD5⁺CD1d^hi cells. Five mice/group. D and E, C57BL/6 mice given tolerogen, and after 5 days TIM-1 and TIM-4 expression analyzed on CD19⁺CD5⁺CD1d⁺ Breg or CD19⁺CD5⁺CD1d⁻ non-Breg in spleen. Four mice/group. qRT-PCR indicates quantitative reverse transcription polymerase chain reaction.

Costimulatory Blockade Increases the Percentage of MZP B Cells and Their IL-10 Expression

Because there was no change in CD5⁺CD1d^hi Breg after tolerogen administration, and because there was increased IL-10 mRNA expression in non–Breg cells, we next enumerated multiple other B-cell subsets in naive and tolerized mice. Various B-cell subsets were gated and defined as shown in Figure S1A (SDC, https://links.lww.com/TP/B147). The results showed that there were significant increases in the percentage of splenic MZP B-cell subsets comparing naive controls and nontolerized mice (Figure 5A), whereas other subsets were unchanged, and there was a nonsignificant decrease in marginal zone B cells. Further analysis of these non–Breg cells showed an increase in IL-10 in the MZP B-cell subset (CD19⁺CD23⁺sIgM^hisIgD^hiCD21/CD35^hi) at both the transcriptional and translational levels, but no increases in IL-10 in other subsets (Figure 5B and C). Anti-CD20 caused depletion of this subset and all other B-cell subsets without preferential subset effects (Figure S1B, SDC, https://links.lww.com/TP/B147). These results revealed that costimulatory blockade increased both the percentage of the MZP B-cell subset and its IL-10 expression.

FIGURE 5: Costimulatory blockade increases MZP B cells and their IL-10 expression. C57BL/6 mice given no treatment (naive), DST, anti-CD40L or DST + anti-CD40L (tolerogen), and after 5 days different subsets of B cells from spleen were analyzed. A, Contour plots show the changes in MZP and MZ cells (left). B-cell subset percentages (right) calculated based on gating scheme in Figure S1A (SDC, https://links.lww.com/TP/B147). Four mice/group. B, IL-10 mRNA expression assessed by qRT-PCR. Error bar is standard deviation. C, Splenocytes (2 × 10⁶ cells/well) stimulated with PMA and ionomycin, intracellular IL-10 stained, and flow gated on various B-cell subsets. Five mice/group. D, IL-21R expression analyzed on MZP and other B-cell subsets. Representative histogram of IL-21R expression on MZP B cells (left). MFI of IL-21R expression on various subsets (right). Four mice/group. E, IL-21R⁺ and IL-21R⁻ B cells purified from control or tolerized mice and IL-10 mRNA analyzed by qRT-PCR. Error bar is standard deviation.

It has been reported that IL-21 together with CD40 regulate B-cell expression of IL-10.^27,28 The MZP B cells from tolerogen-treated mice had significantly increased IL-21R expression compared to other B-cell subsets and compared to nontolerogenic treatment conditions (Figure 5D). The IL-21R⁺ MZP B cells expressed enhanced IL-10 mRNA compared to IL-21R⁻ MZP B cells (Figure 5E), suggesting that IL-21 signaling in the MZP B cells was involved in IL-10 expression.

Deficiency of IL-10 in B Cells Prevents Costimulatory Blockade-Induced Tolerance

To test if IL-10–producing B cells were involved in DST plus anti-CD40L mAb-induced tolerance, we generated CD19Cre^+/−::IL-10 fl^+/+ (B-IL-10^−/−) mice which specifically lacked IL-10 expression in B cells (Figure 6A). These mice had normal B-cell development (Figure 6B) and normal CD4⁺, CD8⁺, and Foxp3⁺ CD4 Treg compared to littermate controls (Figure 6B). B-IL-10^−/− mice were transplanted and received tolerogen, yet deficiency of IL-10 in B cells prevented tolerance in B-IL-10^−/− mice, but not in littermate controls (Figure 6C). Rejecting B-IL-10^−/− mice had increased PR scores suggestive of T cell–mediated rejection, but did not have enhanced alloantibody responses to account for the increased incidence of rejection (Figure 6D). Importantly, adoptive transfer of MZP B cells at the time of transplantation only from IL-10–sufficient littermate mice, but not IL-10–deficient MZP B cells or IL-10–sufficient follicular B cells into B-IL-10^−/−, rescued graft survival (Figure 6C). Parenchymal rejection scores of the allograft were also significantly reduced with the adoptive transfer of wild-type MZP B cells compared to either IL-10^−/− MZP or wild-type follicular B cells (Figure 6C). Transfer of follicular B cells from tolerized wild type mice also did not prevent rejection in B-IL-10^−/− recipients (data not shown). Together, these data demonstrated that IL-10–producing MZP B cells were necessary for induction of costimulatory blockade transplantation tolerance.

FIGURE 6: B cell–specific IL-10 is required for tolerance. A, MZP, MZ and CD19⁺CD5⁺CD1d^hi+ B cells characterized from B-IL-10^−/− or littermates by flow cytometry. IL-10 mRNA expression monitored by qRT-PCR. B, Splenic B-cell subsets (left), CD4⁺ and CD8⁺ cells subsets (upper right), and CD25 and Foxp3 subsets gated on CD4⁺ cells (lower right) analyzed. C, B-IL-10^−/− and littermate control recipients given tolerogen and BALB/c grafts. MZP B cells or follicular (Fol) B cells (4 × 10⁵ cells/mouse) from wild-type or B-IL-10^−/− naïve mice adoptively transferred on the day of transplant. Graft survival (upper) and parenchymal rejection score (lower). Littermate control or B-IL-10^−/− + WT MZP B cells versus B-IL-10^−/− mice (P < 0.05); B-IL-10^−/− versus B-IL-10^−/− + WT Fol B cells or B-IL-10^−/− + B-IL-10^−/− MZP B cells (p ns). D, Alloantibody responses of allograft recipients with the indicated genotypes and treatments measured 14 days after transplantation.

DISCUSSION

The results here demonstrated that costimulation blockade induced a subset of IL-10⁺IL-21R⁺ MZP B cells in the secondary lymphoid organs. Tolerogen did not induce similar changes in other previously characterized suppressive or regulatory B-cell subsets. B cells and B cell IL-10 in the MZP subset were required for costimulatory blockade-induced, alloantigen-specific tolerance, whereas other B-cell subsets were not. We conclude that B cell–produced IL-10 specifically in the MZP subset determines the outcome of costimulatory blockade-induced tolerance.

The role of B cells in immunity and tolerance is varied and diverse. B cells play an important role in the induction of peripheral tolerance.²⁹ Conversely, as revealed by B cell–deficient μMT mice, B cells are required for acute rejection.³⁰ However, μMT animals have several immune abnormalities including decreased size of secondary lymphoid organs, decreased diversity in the T-cell repertoire and number, imbalance of Th1/Th2 cytokines,¹⁴ and lack of macrophages and follicular dendritic cell subsets. It has been shown that tolerated grafts have increased infiltration of B cells compared to rejecting grafts.^31,32 Depletion of B cells with anti-CD19 mAb slightly increases graft survival, whereas anti-CD20 mAb leads to acute rejection of minor histocompatibility-mismatched skin, cardiac, and renal allografts.⁵ In a kidney transplant clinical trial, depletion of B cells induced acute cellular rejection,⁸ yet others have shown that B-cell depletion increases graft survival.⁴ Thus, the depleting strategy affects different subsets of B cells which have diverse immunomodulatory functions. Together, these observations show that no general consensus about the role of B cells in tolerance exists and suggest that simple depletion of B cells may be of limited therapeutic value or even detrimental.

There are several B-cell surface receptors, such as CD19, CD20, CD22, and CD79, which have been targeted to deplete B cells in autoimmune disease models. Functional deficiency of B cells or depletion of B cells can exacerbate inducible autoimmune disease due to increased inflammatory cytokine production³³ or infiltration of mononuclear cells in autoimmune prone organs,^34,35 although B cell–derived IL-10 was not protective in at least 1 model.³⁶ Recently, it was shown that IL-10–producing cells that belong to the MZP B cell population play a protective role in experimental arthritis,³⁷ and mice lacking these populations have increased inflammatory Th1/Th17 and decreased Treg differentiation, leading to exacerbation of disease.³⁷ B-cell costimulatory and regulatory function is required for anti-CD45RB–induced cardiac transplantation tolerance.^38,39 However, absence of IL-10 in B cells increased anti-CD45RB mAb-mediated graft survival and tolerance induction.⁴⁰ In the CD45RB model, there are increased regulatory CD4⁺ T cells in the secondary lymphoid tissues as well as reduced alloantibody production. The cellular and molecular mechanisms of how IL-10 deficiency in B cells prolonged anti-CD45RB mAb-mediated graft survival requires further exploration. Our data showed that depletion of B cells under tolerogenic conditions lead to acute rejection with migration of inflammatory effector cells into the graft. In models in which B cell depletion is protective, combination immunosuppressive regimens using tacrolimus or rapamycin that control the differentiation of T cells are used.^5-8 Thus, the timing of B-cell depletion, the subset depleted, and the precise combination of other immunosuppressive variables in the model may result in either pathogenic or protective B-cell function.

Blocking of costimulatory CD40-CD40L signals provides a robust approach to induce tolerance, and several mechanisms have been proposed to explain CD40-CD40L blockade-induced transplantation tolerance. The mechanisms depend on the type of transplant, subsequent therapy, and the relative importance of CD4⁺ and CD8⁺ T-cell responses.^41-43 However, the cellular and molecular mechanistic contributions of B cells in costimulatory blockade-induced tolerance have not been well studied. We found here that costimulatory blockade induced an IL-10⁺IL-21R⁺ MZP B-cell subset, whereas others have shown that IL-21R deficient mice do not expand the IL-10–producing B cells after antigenic stimulation in an experimental autoimmune encephalomyelitis mouse model.⁴⁴ We did not find increased CD19⁺CD5⁺CD1d⁺TIM-1⁺ B cells or IL-10 production in these cells during tolerization as reported in a different model of allograft tolerance.²⁶ We also did not find the well-described IL-10–producing (B10) Breg subset.¹⁷ However, it is noteworthy that B10 Breg rely on CD40 for their induction or expansion,⁴⁴ so that the tolerogen used in our model may have prevented their induction. Our results and others suggest there are multiple discrete regulatory B-cells subsets that rely on IL-10 or other immunosuppressive mechanisms and that the immunosuppressive regimen may dictate which one(s) is functional in a particular setting.

A theme is emerging in studies of lymphoid organ structure that there are distinct microdomains which dictate the regulation of immune responses, so that Th1, Th2, Th17, T follicular helper cell, Treg, follicular regulatory CD4⁺ T cell, Tc1, and Tc2 responses take place in different places and times.⁴⁵ Interventions or stimuli that alter lymphoid organ structure or lymphocyte migration may have profound effects on immune interactions and the net result of the immune response.⁴⁵ The roles of regulatory B cells and how they shape these microdomains in diverse immune responses, including transplantation tolerance, are not well defined. Our work showed that IL-10–sufficient MZP B cells help in the establishment of tolerance. Because marginal zone and MZP B cells are found in proximity to the follicles and germinal centers,^46,47 our findings suggest that the B-cell subsets may contribute to tolerance through effects on T follicular helper and regulatory subsets.^48,49 Indeed, our preliminary data point to such an effect (not shown). Therapeutic regimens that disrupt the regulatory network in the germinal center, including lymphoid organ domain structure, inhibit tolerance induction, leading to acute inflammation and allograft rejection.⁵⁰ The fine balance among cytokine and chemokine expression and precise localization in the germinal center dictates the subsequent fate of T cells and the inhibition or generation of tolerance.

ACKNOWLEDGMENTS

The authors thank Drs. Christopher Karp, Axel Roers and Mark Shlomchik for reagents. The authors also acknowledge the help from University of Maryland Marlene and Stewart Greenebaum Cancer Center Flow Cytometry Shared Service, and National Centre for Cell Science FACS core facility for help.