The role of B cells in transplant tolerance remains unclear. Although B-cell depletion often prolongs graft survival, sometimes it results in more rapid rejection, suggesting that B cells may have regulatory activity. We previously demonstrated that tolerance induction by anti-CD45RB antibody requires recipient B cells. Here, we show that anti-CD45RB in combination with anti-TIM-1 antibody has a synergistic effect, inducing tolerance in all recipients in a mouse islet allograft model. This effect depends on the presence of recipient B cells, requires B-cell IL-10 activity, and is antigen-specific. These data suggest the existence of a regulatory B-cell population that promotes tolerance via an IL-10-dependent pathway.
B cells have a well-established role in allograft rejection, both in hyperacute and antibody-mediated rejection (1), as well as in promoting cellular immunity (2,3). However, there has been increasing evidence for a subset of B cells that is capable of suppressing, rather than enhancing immunity. In the experimental autoimmune encephalomyelitis (EAE) mouse model, adoptive transfer of a specific subset of CD1dhi CD5+ B cells negatively regulates disease induction in B cell-depleted mice in an IL-10-, CD40- and B7-dependent manner (4,5). IL-10-producing B cells are also associated with increased tumor growth and impaired T-cell responses (4,5). In a collagen-induced arthritis model, CD19+CD21hiCD23hiCD24hi transitional B cells have suppressive activity (6).
Small, resting B cells have been associated with prolonged islet allograft survival in mice (7), though it is unclear whether this is due to true, regulatory activity, or simply induction of anergy resulting from aberrant antigen presentation by the B cells (8–10). However, we previously reported more compelling results that transplant tolerance induced by anti-CD45RB is dependent on the presence of recipient B cells. In our model, tolerance required both CD40 and B7 expression on the B cells, suggesting that lack of co-stimulatory molecules and subsequent anergy did not play a major role in the pathway to long-term graft acceptance (11).
In that model, tolerance occurred in only a fraction of recipients, so we investigated the effect of combining anti-CD45RB treatment with an antibody that could be expected to promote tolerance via T-cell effects. The T-cell immunoglobulin and mucin domain (TIM) family proteins are potent costimulatory molecules in T-cell activation (12). RMT1–10, a monoclonal antibody that blocks TIM-1 signaling, prolongs graft survival and promotes costimulatory blockade-induced tolerance (13). Here, we report that the effects of anti-CD45RB in combination with anti-TIM-1 are not only additive, but synergistic. Their combined effect is dependent on the presence of B cells, regulatory T cells and B cell IL-10.
Materials and Methods
BALB/c, C57BL/6, B6μMT−/−, and IL-10−/− mice were purchased from Jackson Laboratories (Bar Harbor, ME, USA). All mice were housed under specific pathogen-free barrier conditions. All procedures detailed below were performed under the principles of laboratory animal care and approved by the IACUC committee at Massachusetts General Hospital.
Diabetes in C57BL/6 mice or B6μMT−/− mice was induced by a single intraperitoneal injection of 200 mg/kg streptozotocin (STZ; Sigma-Aldrich, St. Louis, MO, USA). Diabetes was defined as blood glucose levels >300 mg/dL for at least 3 consecutive days. Islets from BALB/c donors were isolated by the standard technique of collagenase digestion and Ficoll density gradient purification. Five hundred fresh islets were transplanted under the kidney capsule of diabetic mice. Euglycemia was defined as a nonfasting blood glucose level <200 mg/dL. Rejection was diagnosed when animals became hyperglycemic again, with blood glucose >200 mg/dL for at least 2 consecutive days. Allograft function was confirmed by nephrectomy of the kidney containing the transplanted islets. All recipients with long term grafts became hyperglycemic within 48 h of nephrectomy.
100 μg of anti-CD45RB mAb was administered on days 0, 1, 3, 5 and 7 following transplant, 500 μg of antagonistic anti-TIM-1 mAb (RMT1–10) on day-1, 300 μg on days 0 and 5, 250 μg of anti-CD25 mAb was on days 6 and -1. Antibodies were purchased from Bio Express Inc. (West Lebanon, NH, USA). For IL-10 neutralization, 200 μg rat anti-mouse IL-10 antibody (clone JES5–2A5 from Bio Express) was administered i.p. every other day posttransplantation for a total of five doses. Two hundred and fifty μg anti-CD20 mAb (provided by Biogen IDEC[Cambridge, MA, USA) was administered on day 9 i.p.
Cell sorting and transfer
B cells from C57BL/6 mice bearing long-term islet allografts (BALB/c) after combined antibody treatment were selected using Miltenyi anti-CD19 microbeads (Germany). Purity of the resulting B-cell population exceeded 95%. 5×106 B cells were then injected into either C57BL/6 or B6μMT−/− that received simultaneous BALB/c islet transplants without antibody treatment.
Diabetic B6.RAG recipients received 5 × 106 sorted TIM-1+ or TIM-1− B cells along with 5 × 106 B cell-depleted naive C57BL/6 splenocytes intravenously. Recipients received islet allografts on the day of adoptive transfer.
Lymphocytes were prepared from spleen and peripheral lymph nodes of mice that were transplanted and treated with combined antibody therapy. The following antibodies were used for staining: CD4-PECy7, B220 Pacific Blue, Foxp3 APC, RMT1- 4-biotin (all from eBioscience, San Diego, CA, USA, including Foxp3 fix-perm kit), and streptavidin-A750 (Invitrogen, Grand Island, NY, USA). Cells were analyzed on the LSRII cytometer (Becton Dickinson, San Diego, CA, USA), and the results were analyzed with FlowJo (Treestar, Ashland, OR, USA) software.
For survival studies, Kaplan–Meier survival curves were generated and statistical analysis performed by use of the log-rank test. All data were analyzed by the Prism 5 Program (Graph Pad, San Diego, CA, USA). A value of p < 0.05 was considered significant.
Hundred percent long-term graft survival with combined antibody treatment
Anti-TIM-1 treatment alone did not prolong islet allograft survival. About half of allografts survive long-term in anti-CD45RB treated recipients, consistent with previous reports (14). However, anti-CD45RB and anti-TIM-1 treatments together resulted in long-term survival of grafts in all recipients (Figure 1).
Significant TIM-1 upregulation on B cells upon anti-CD45RB and anti-TIM-1 treatment
Although TIM-1 was originally described as a costimulatory molecule for T cells, it is more strongly expressed on B cells than on CD4+ T cells (15). TIM-1 expression on splenic B cells is significantly upregulated in islet recipients treated with anti-CD45RB and anti-TIM-1 (Figure 2). The absolute number of B cells or total splenocytes is not significantly altered in any of the groups (data not shown). There is no increase in TIM-1 expression following antibody injection alone, of either anti-TIM-1 or anti-TIM-1/anti-CD45RB, in the absence of transplant (data not shown).
Combined antibody-induced graft survival requires B cells
Due to the significant upregulation of TIM-1 on B cells upon transplant with antibody treatment, we investigated graft survival in B-cell-deficient mice. BALB/c islets transplanted into μMT−/− (C57BL/6 background) recipients are rejected consistently and with a tempo comparable to wild-type (WT) recipients (Figure 3A). Graft survival in μMT−/− mice treated with anti-TIM-1 antibody alone or with anti-CD45RB antibody alone is prolonged compared to wild-type recipients. In contrast, combined antibody treatment in μMT−/− recipients resulted in no extension of graft survival over wild-type recipients (Figure 3A). We confirmed the role of B cells by reconstituting the B-cell-deficient recipients with mature syngeneic B cells by iv injection from normal C57BL/6 on the day of transplant. Dual antibody treatment restores tolerance induction to islet allografts in these mice (data not shown and Figure 4).
To further demonstrate B-cell dependence, we performed antibody based B-cell depletion in WT recipients. C57BL/6 recipients were given a standard course of anti-TIM-1 and anti-CD45RB and additionally treated with a single dose of anti-CD20 on day 9 posttransplant. Anti-CD20 treatment results in sustained B-cell depletion more than one week postinjection (data not shown). Sixty percent of these mice rapidly rejected their islet grafts following anti-CD20 administration (Figure 3B). Thus, we conclude that graft acceptance following combined antibody treatment is B-cell-dependent.
Regulatory B cells have been demonstrated to be IL-10-dependent (4,15,16), thus, we tested whether the combined antibody treatment was also IL-10-dependent. On the day of transplant, μMT−/− recipients received standard dual antibody treatment plus either IL-10 competent or deficient purified B cells. Transfer of wild-type B cells results in significant prolongation of graft survival, while in contrast, IL-10-deficient B cells are unable to significantly prolong graft survival (Figure 4). We conclude that in this model graft survival is dependent on IL-10 production by B cells.
Tolerant B cells transfer antigen-specific tolerance
Given the requirement for B cells, we FACS purified B cells from recipients bearing long-term surviving BALB/c donor grafts (>100 days, LTS) and adoptively transferred them to μMT−/− and WT mice that received BALB/c transplants without getting antibody treatment. Indefinite graft survival was seen in these recipients, though the effect was more pronounced in μMT−/− mice than in WT mice (Figures 5A and B). Adoptive transfer of naive B cells does not prolong graft survival (Figure 5A). LTS B cells cannot prolong graft survival of third party C3H grafts, which suggests that these tolerant B cells are antigen-specific. Adoptive transfer of B cells from long-term survivor anti-CD45RB-treated recipients does not prolong graft survival (Figure 5B).
We then sorted TIM-1+ and TIM-1− B220 B cells from C57BL/6 mice 14 days after injecting 20×106 irradiated BALB/c splenocytes and the standard course of anti-CD45RB/anti-TIM-1 antibodies, as an alternate immunizing/tolerizing stimulus. B-cell subpopulations were then adoptively transferred with an equal number of B-cell-depleted naive C57BL/6 splenocytes to diabetic B6.RAG recipients simultaneously transplanted with BALB/c islets. Compared to TIM-1− B cells, co-transfer of TIM-1+ B cells results in significant prolongation of BALB/c islet grafts (Figure 5C). From this, we conclude that TIM-1+ B cells contain a regulatory population and that TIM-1 may be a marker for regulatory B cells.
Tregs are also required for long-term graft acceptance
Regulatory T cells (Tregs) have previously been demonstrated to be involved in anti-CD45RB and anti-TIM-1-mediated allograft survival (13). We found that the percentage of Tregs significantly increases following combined antibody treatment (Figures 6A and B). Depletion of Tregs by anti-CD25 PC61 treatment pretransplant resulted in prompt rejection in nearly all such treated recipients (Figure 6C).
Anti-TIM-1 alone did not prolong survival of islet allografts and anti-CD45RB alone promoted long-term survival in only 40% of recipients, but the two antibodies combined resulted in 100% long-term graft survival. This synergistic effect was completely dependent on the presence of recipient B cells. Although the number of Tregs increased, and Treg depletion led to rejection, isolated B cells, in particular, TIM-1+ B cells from antibody-treated recipients were sufficient to transfer tolerance, without the transfer of Tregs.
Each antibody alone prolongs graft survival in the absence of B cells, which is not unexpected as the antibodies do not target B cells exclusively (13,18,19), and adoptive transfer of B cells from mice treated with either antibody alone into μMT or WT recipient is not capable of increasing graft survival (Figure 5B and data not shown). The synergistic effect of the two antibodies suggests that each antibody is acting on an independent signaling pathway, and that the two pathways interact to generate an effective regulatory, TIM-1+ B cell population. It is interesting to note that anti-CD45RB alone can induce tolerance in μMT mice, but combined anti-CD45RB / anti-TIM-1 cannot. Although this could be explained by Ding and Rothstein's finding that anti-TIM-1 treatment actually accelerates rejection in B cell-deficient mice, in our strain combination, anti-TIM-1 minimally prolongs graft survival. This suggests that either anti-CD45RB and anti-TIM-1 function in a single pathway or that the dependence of anti-TIM-1 on B cells is dominant to the B cell-independent effects of anti-CD45RB.
Ding et al. recently published studies demonstrating regulatory activity by TIM-1+ B cells in the reverse strain combination (C57BL/6 to BALB/c), using anti-TIM-1 treatment in an islet transplant model (15). Using C57BL/6 recipients, with anti-TIM-1 monotherapy, we do not observe significant graft prolongation (Figure 1) whereas we do observe prolongation in the BALB/c background consistent with Ding et al. This may be due to strain-dependent response variation (20). For example, in response to the same immune stimulus, BALB/c are more likely to develop a Th2 response whereas C57BL/6 are more likely to develop a Th1 response.
Differences in vascularization, expression of costimulatory markers, adhesion molecules, cytokines, chemokines and receptors / ligands on leukocytes as well as on endothelial cells have been proposed for understanding organ-specific differences in tolerance and rejection (24–27). Thus, specific organs may produce a unique set of cytokines and/or respond differently to cytokines produced by the immune system. For example, recipients deficient in CCR2, the receptor for chemokine MCP-1, reject heart allografts but mount a limited response against islet allografts. Similarly, we have previously shown that anti-CD45RB cannot induce tolerance to cardiac allografts, but find in this study that it does induce tolerance to islet allografts. IL-10 is not required for tolerance to heart grafts induced by anti-CD45RB, but IL-10 production by B cells is required for tolerance to islets. Furthermore, anti-CD45RB-induced cardiac tolerance requires B-cell expression of B7 and CD40, but does not require expression of antibody (or IL-10; Ref. 11,17), suggesting that in islet tolerance, Bregs interact directly with T cells rather than promoting tolerance by using cytokine expression to influence T-cell function or differentiation. Thus, anti-CD45RB-induced tolerance could serve as an additional model to study organ-specific rejection.
B cells are known to be important as antigen presenting cells (28), providing costimulatory signals (29) and driving Th1 or Th2 differentiation (30). Considering the simultaneous requirement for Tregs in combined antibody induced tolerance and the fact that antigen is required for the development of TIM-1+ B cells, it is tempting to speculate that TIM-1+ B cells act as Treg inducing antigen presenting cells. This will require further work to elucidate. Although regarded as an antagonistic Ab, the mechanism of RMT1–10 action remains unclear, including its ability to increase the percentage of TIM-1+ B cells. Potentially, the TIM-1-TIM-4 interaction downmodulates TIM-1 expression and by blocking this interaction, TIM-1 expression is increased.
In summary, we have shown that combined anti-CD45RB and anti-TIM-1 treatment results in 100% long-term islet allograft survival that is B-cell-dependent, dependent on B-cell production of IL-10, and is associated with upregulation of TIM-1 on B cells. Our data support other's findings that transplant tolerance may depend on a novel TIM-1+ regulatory B-cell subset.
This work was supported in part by NIH grant RO1AI057851–05 (JFM) and K01DK079207 (JIK).
The manuscript was not prepared by or funded by a commercial organization. The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.