Long-Term Control of Alloreactive B Cell Responses by the Suppression of T Cell Help¹

It is widely appreciated that alloantibodies can play a key role in acute and chronic allograft rejection. However, relatively little is known of factors that control B cell responses following allograft tolerance induction. Indeed, most investigations of the induction of allograft tolerance have focused on the regulation of T cell responses (1, 2, 3, 4). We hypothesize that control of B cell responses is necessary for stable long-term allograft survival. This hypothesis is based on the notion that alloantibodies and alloreactive B cells can contribute to allograft rejection by at least three independent mechanisms. First, Abs can directly bind to the allograft and induce injury to the graft endothelium, ultimately causing graft rejection (5). Second, Ag-Ab complexes can bind to activating FcγRs expressed on APCs and induce their maturation into immunogenic APCs (6). These mature APCs can then activate alloreactive T cells, thereby preventing allograft tolerance. Third, activated alloreactive B cells can serve as APCs and directly interact with alloreactive T cells (7). Thus, understanding the mechanisms by which B cell responses are controlled may lead to new ways of inducing and maintaining allograft tolerance.

B cell tolerance to allografts, when achieved, will likely exploit the same mechanisms as those maintaining B cell tolerance to self-Ags. A number of critical checkpoints at which self-reactive B cells are controlled have been previously described (8). These checkpoints include the following. 1) The initial checkpoint comprises clonal deletion and receptor editing of immature self-reactive B cells in the bone marrow (9, 10). In addition, newly formed self-reactive B cells that escape from the bone marrow can be eliminated in the periphery (11, 12). 2) Self-reactive B cells that escape deletion or receptor editing in the bone marrow and emerge in the periphery can enter a state of clonal anergy, a state that has been associated with the down-regulation of membrane IgM and diminished receptor signaling through IgD (12, 14). 3) Absence of T cell help can result in the downstream inhibition of T cell-dependent B cell responses (12, 15). 4) Immune deviation of Th cells, resulting in deficiencies of Th1 or Th2 cytokines, may instruct B cells to secrete selected subclasses of IgG that are less pathogenic (16). 5) Regulatory T cells can inhibit autoreactive humoral responses either by inhibiting Th cells or B cells themselves (17, 18, 19). Collectively, these observations suggest that robust B cell tolerance requires multiple checkpoints acting within the bone marrow, as well as in the periphery.

Using a BCR-transgenic mouse, 3-83 Igi, as recipients, we previously demonstrated that tolerance induction by anti-CD154 and transfusion of donor-specific spleen cells (DST)³ can result in the long-term peripheral deletion of mature alloreactive B cells (20). In this study, we have used these 3-83 B cells to visualize the fate of alloreactive B cells during tolerance induction in nontransgenic mice. Although 3-83 Abs cannot directly elicit vascular rejection, we demonstrate here that tolerance induced by anti-CD154 and DST in nontransgenic BALB/c mice leads to the long-term deletion of 3-83 B cells and inhibition of alloantibody production that is dependent on the dominant suppression of T cell help.

Six- to 10-wk-old BALB/c, BALB/c-scid, and BALB/c RAG2^−/− mice (Taconic Farms or The Jackson Laboratory) were used as recipients, and C57BL/6, C3H, or (C57BL/6 × C3H)F₁ mice (Taconic Farms or The Jackson Laboratory) were used as heart donors. Heterotopic mouse hearts were transplanted into the abdomen of recipient mice by anastomosing the donor aorta and recipient aorta and the donor pulmonary artery and recipient inferior vena cava. Second heart grafts were transplanted into the cervical area of the recipient by anastomosing the donor aorta and recipient carotid artery and the donor pulmonary artery and recipient external jugular vein (end-to-side). The heart grafts were monitored daily for the first 2 wk, then two times a week until rejection; rejection was defined as complete cessation of pulsation.

The 3-83Igi mice had been backcrossed to BALB/c for eight generations and express both the 3-83IgH and 3-83Igκ chains (21). For some indicated studies, 3-83 Igi mice were crossed to the BALB/c RAG2^−/− background. Mice were maintained in specific pathogen-free conditions at the University of Chicago. All experiments were done in accordance with the guidelines of the Institutional Animal Care and Use Committee.

Anti-CD154 mAbs (MR1 mAbs) were purified from protein-free culture supernatants by 45% ammonium sulfate precipitation and dialyzed in PBS. Anti-CD154 was administered at a dose of 1 mg/mouse, i.v. on day 0, followed by i.p. injections on days 7 and 14 posttransplant. DST comprised RBC-lysed donor-derived spleen cells, which was administered i.v. (2 × 10⁷/recipient) on the day of transplantation.

Whole spleen cells from naive or tolerant mice (2 × 10⁷/recipient) were transferred into BALB/c-scid or BALB/c RAG2^−/− on the day of transplantation. For infectious tolerance, the number of naive spleen cells remained at 2 × 10⁷, while the number of tolerant spleen cells varied as indicated in the figure legends. 3-83 B cells were purified by negative selection with a B Cell Isolation Kit (Miltenyi Biotec), and 1.5 × 10⁶ enriched B cells (>97% pure) were transferred i.v. on the day of transplantation. CD25⁺ cells were depleted from naive or tolerant spleen cells with the anti-CD25 (PC61) Ab followed by incubation with rabbit complement, and the CD25-depleted cells contained <0.2% CD4⁺CD25⁺ cells when visualized by flow cytometry using anti-CD25 (7D4) mAbs (BD Pharmingen). These cells were transferred i.v. into BALB/c-scid mice on the day of heart transplantation. In some indicated experiments, 1 × 10⁶ 3-83 B cells (from 3-83 BALB/c RAG2^−/−) along with 2–3 × 10⁷ total or CD25-depleted spleen cells from tolerant BALB/c were transferred to BALB/c RAG2^−/−.

Donor-reactive Abs were determined by flow cytometry as previously reported (22). Briefly, C57BL/6 or C3H lymph node cells were incubated with a 1/10 dilution of mouse serum for 1 h at 4°C, then the cells were washed and incubated with PE-conjugated anti-mouse IgM (Jackson ImmunoResearch Laboratories) or FITC-conjugated anti-mouse IgG (Southern Biotechnology Associates). The mean channel fluorescence of the stained samples was determined by flow cytometry (FACScan; BD Biosciences).

3-83 IgM and IgG titers in sera were determined by ELISA using the anti-idiotypic 54.1 Ab (23). 54.1 mAb-coated plates (Corning) were blocked with 1% BSA/PBS, then diluted serum (1/10 in 1% BSA/PBS) was added to triplicate wells. After 1 h, plates were washed, then incubated with HRP-conjugated anti-mouse IgM (BD Pharmingen) or biotin-conjugated goat anti-mouse IgG (Jackson ImmunoResearch Laboratories) followed by avidin-HRP (BD Pharmingen). The OD were determined on an ELISA plate reader (Bio-Rad) and the results are presented as mean relative OD.

Heart grafts and spleens were surgically removed and snap frozen in Tissue-Tek OCT (Sakura Finetek) using liquid nitrogen. Hearts and spleen were sectioned (5 μm) and stained with H&E for histology. Other sections were immunostained using a standard avidin-biotin peroxidase method (24) or immunofluorescence. A mixture of biotinylated rat anti-mouse IgG1 (A85-1), IgG2a (R19-15), IgG2b (R12-3), and IgG3 (R40-82) was used to detect IgG deposition on cardiac allografts. To detect mononuclear cell infiltration, purified anti-mouse CD8a (53-6.7) was applied as primary Ab and biotinylated goat anti-rat IgG as secondary Ab. For the identification of B cells and development of germinal centers in the spleen, serial sections of each spleen were stained by immunofluorescence with PE-conjugated rat anti-mouse B220 (RA3-6B2) and anti-mouse follicular dendritic cell Ab (FDC-M1). All Abs were obtained from BD Biosciences Pharmingen.

Statistical significance was determined by the unpaired t test or ANOVA followed by and post hoc Dunnett’s multiple comparison or Student-Newman-Kuels tests (Prism 4 for Macintosh (GraphPad) or StatView (Abacus Concepts)). A p < 0.05 was considered statistically significant.

Allogeneic C57BL/6 hearts were rejected in 9 days when transplanted into untreated BALB/c recipients, whereas treatment with anti-CD154/DST induced long-term allograft survival (Fig. 1,a). Histological examination of the long-term allografts (>100 days) indicated normal histology of these long-term surviving allografts (Fig. 2,a). Tolerance was formally demonstrated by the acceptance of a second donor-specific heart graft but not a third-party C3H graft (Fig. 1 b).

Active regulation as the basis for allograft tolerance has been reported in a number of rodent transplantation models (25, 26). However, Bickerstaff et al. (27) reported that BALB/c recipients tolerized to C57BL/6 heart grafts following anti-CD154 therapy do not develop ex vivo delayed-type hypersensitivity regulatory responses, whereas C57BL/6 mice tolerized to BALB/c heart grafts do. We therefore tested whether regulatory mechanisms develop and mediate allograft tolerance in the BALB/c recipient using more conventional tests of peripheral, dominant tolerance, namely, linked suppression and infectious tolerance (28, 29). Linked suppression is defined as the ability to spread tolerance to third-party Ags when those Ags are linked to the tolerant donor Ag on the same APC. Infectious tolerance refers to the ability of tolerant cells to influence naive or primed cells and induce them to become tolerant. Linked suppression was assessed by challenging mice tolerant to C57BL/6 hearts with (C57BL/6 × C3H)F₁ hearts. As illustrated in Fig. 1,b, F₁ grafts exhibited prolonged survival compared with third-party grafts (p < 0.05). Immunohistochemistry showed no CD8⁺ T cell infiltration and no IgG deposition in the tolerant C57BL/6 or F₁ graft. In comparison, there was abundant CD8⁺ T cell infiltration and IgG deposition in the third-party C3H graft (Fig. 2 a).

Infectious tolerance was assessed by testing whether spleen cells (2 × 10⁷/mouse) transferred into BALB/c-scid mice from tolerant BALB/c mice could suppress the ability of naive/non-tolerant BALB/c spleen cells to mediate the rejection of a donor-specific C57BL/6 heart graft. Nontolerant spleen cells rejected C57BL/6 heart grafts, while spleen cells from tolerant mice did not (Fig. 1,c). However, spleen cells from tolerant mice were able to reject third-party C3H hearts (data not shown). When tolerant spleen cells were transferred with different numbers of nontolerant cells into BALB/c-scid mice, allograft rejection was observed at a 1:1 or 2:1 ratio of tolerant:nontolerant spleen cells. At a 4:1 tolerant:nontolerant spleen cell ratio, no rejection was observed (Fig. 1,c). Although these could reflect either inhibition of homeostatic proliferation, by the tolerant spleen cells, of the naive T cells that is necessary for acutely rejecting the allograft, collectively, the data in Fig. 1, b and c, are consistent with the conclusion that BALB/c recipients develop dominant transplantation tolerance following treatment with anti-CD154/DST.

We first demonstrate that anti-CD154 DST after transplant resulted in a significant and prolonged suppression of allo-IgM and allo-IgG responses compared with the untreated nontolerant group, which demonstrated a 2.5-fold increase in allo-IgM and a 55-fold increase in allo-IgG responses (Fig. 3, a and b; p < 0.05). To further assess whether alloreactive B cell function could be maintained by peripheral infectious tolerance mechanisms, we transferred a population of naive, alloreactive B cells into tolerant BALB/c recipients. This use of nontolerant alloreactive 3-83 B cells allowed us to confirm that alloreactive B cells can be actively regulated in tolerant recipients. For a source of alloreactive B cells, we used B cells isolated from naive 3-83 Igi mice that have dual anti-H-2K^b/k specificity (21). These 3-83 B cells were transferred into naive/nontolerant or tolerant BALB/c mice that received a freshly transplanted C57BL/6 heart graft on the same day. A second heart transplant was performed in the tolerant BABL/c recipient to ensure that both naive/nontolerant and tolerant mice received the same alloantigenic challenge. When purified 3-83 B cells (1.5 × 10⁶/mouse) were transferred into naive/nontolerant BALB/c mice that had received a C57BL/6 heart, we observed significant increases in the concentrations of donor-specific allo-IgM and allo-IgG (Fig. 3, c and d; p < 0.05) as well as 3-83-idiotypic IgM (Fig. 3,e; p < 0.05). When 3-83 B cells were transferred into tolerant BALB/c mice that received a second donor-specific heart graft, we observed no significant increase in the concentrations of donor-specific allo-IgM and allo-IgG nor of idiotypic IgM responses (Fig. 3, c–e; p > 0.05). Mice that did not receive 3-83 B cells or mice that received 3-83 B cells only, but no C57BL/6 heart grafts, did not exhibit significant increases in idiotypic IgM (data not shown). We could not assess the suppression of the 3-83 IgG response because of the presence of B cells spontaneously secreting 3-83 IgG in the spleens of naive 3-83 knock-in mice (data not shown). Thus, transfer of splenic B cells from naive 3-83 mice into untransplanted BALB/c recipients resulted in high titers of circulating 3-83 IgG titers, which were unmodified in tolerant recipients (data not shown).

The presence of germinal center development in the spleen and peripheral lymphoid structures is considered evidence of in vivo T-dependent B cell priming and activation. We observed that the reduced Ab titers following challenge with a second donor-specific graft was associated with minimal germinal center formation in tolerant mice (Fig. 2 b). These observations are consistent with the notion that peripheral mechanisms prevent the production of alloantibody in tolerant BALB/c recipients of 3.83 B cells and second C57BL/6 hearts.

Both allo-IgM and IgG responses are dependent on T cell help, because anti-CD154/DST treatment or the absence of CD4 cells in CD4^−/− mice results in significantly reduced alloantibody production (Fig. 3,a and data not shown). Thus, it is possible that the inability of tolerant mice to produce donor-specific Abs, and naive 3-83 B cells transferred to produce 3-83 IgM Abs when transferred into tolerant mice, was due to the absence of T cell help. To test this possibility, we took advantage of the dual specificity of the 3-83 BCR, which binds K^k with the highest affinity and K^b with intermediate affinity (10). We transferred 3-83 B cells into mice tolerant to C57BL/6 hearts, then transplanted these tolerant recipients with a third-party C3H heart graft. As indicated in Fig. 1,b, these grafts were acutely rejected. We reasoned that C3H-specific, self-restricted T cells in these tolerant mice should not be tolerant and would be primed by the indirect pathway by recipient APC presenting C3H Ags. These T cells could be able to interact with alloreactive B cells to provide cognate T cell help. Indeed, we observed that C3H-reactive B cells produced C3H-reactive IgM and IgG alloantibodies in both nontolerant and C57BL/6-tolerant BALB/c (Fig. 4, a and b). We also assessed the ability of naive 3-83 B cells to become activated and to secrete 3-83 IgM in nontolerant and C57BL/6-tolerant recipients. We observed comparable production of 3-83 IgM titers in both nontolerant and C57BL/6-tolerant recipients (Fig. 4 c; p < 0.05).

We also observed the development of germinal centers in the spleens of C57BL/6-tolerant mice challenged with C3H heart grafts (Fig. 2 b), consistent with the increased titers of anti-C3H alloantibody and 3-83 IgM. Thus, a mechanism by which long-term alloantibody production is controlled in mice made tolerant with anti-CD154/DST treatment is through the sustained absence of T cell help.

An important feature of active T cell regulation is linked suppression, in which tolerance to a third-party Ag can be elicited if the third-party Ag is presented in conjunction with the donor-specific Ag. Although this has been demonstrated for allograft rejection, it is not known whether linked suppression can also occur for alloantibody responses. We therefore tested whether linked suppression of 3-83 B cell responses can be observed following the challenge of (C57BL/6 × C3H)F₁ hearts in naive/nontolerant or tolerant BALB/c mice.

We observed the most efficient suppression of anti-C57BL/6 IgM and IgG production in BALB/c mice initially tolerized to C57BL/6 hearts and then challenged with (C57BL/6 × C3H)F₁ grafts (Fig. 5, a and b). We also observed that the 3-83 IgM response was significantly suppressed in tolerant mice challenged with F₁ hearts (Fig. 5,e). The anti-C3H IgG responses also remained significantly suppressed (p < 0.05), while the anti-C3H IgM response was transiently delayed, but eventually reached the levels of naive/nontolerant BALB/c recipients that rejected C3H grafts (Fig. 5, c and d; p > 0.05). Thus, we are were to demonstrate linked suppression of B cells responses in tolerant mice, although the linked suppression of IgG responses was more effective than that of IgM responses. We reason that this may be due, at least in part, to allo-IgG production being more T cell dependent than IgM responses. Examination of the spleens of these recipient mice reveal an inhibition of germinal center formation that was not significantly different from that of the spleens of mice receiving donor-specific heart grafts (p > 0.05), but that was significantly different from those receiving third-party heart grafts (Fig. 2 b; p < 0.05). Collectively, these data are consistent with a model of active regulation of T cell help that results in a sustained suppression of alloreactive B cell responses.

A number of reports have suggested an important role for CD4⁺CD25⁺ regulatory T cells in the suppression of alloreactive T cell responses and in the maintenance of allograft tolerance (30, 31). We therefore tested whether depletion of CD25⁺ cells would reverse tolerance and restore alloreactive B cell responses. Depletion of CD25⁺ cells in the intact tolerant BALB/c mouse did not reverse tolerance (data not shown), leading us to use a model of spleen cell transfer into immune-deficient BALB/c-scid or BALB/c RAG2^−/− mice. In this model, the transfer of naive spleen cells (10⁷/mouse) into BALB/c-scid recipients resulted in the acute rejection of C57BL/6 heart grafts, but that the transfer of tolerant spleen cells resulted in delayed or no rejection (Figs. 1,c and 6,a). To avoid problems that could arise as a result of the immunosuppressive effects of anti-CD25 in vivo, we depleted CD25⁺ cells from the spleen cells from tolerant mice in vitro. These CD25-depleted splenocytes, when transferred into BALB/c-scid recipients of C57BL/6 heart grafts, were able to induce the rejection of C57BL/6 heart grafts, confirming previous reports of a role of CD25⁺ T cells in the prevention of acute allograft rejection (Fig. 6 a and Ref. 32).

We quantified the levels of allo-IgM and IgG in the BALB/c-scid mice receiving naive/nontolerant, tolerant and CD25-depleted spleen cells. No significant increase in allo-IgM was detected (data not shown), but increased titers of allo-IgG were observed in BALB/c-scid mice receiving naive/nontolerant spleen cells and C57BL/6 heart grafts (Fig. 6 b). There was a significant reduction in the titer of allo-IgG in BALB/c-scid mice receiving tolerant spleen cells and C57BL/6 heart grafts (p < 0.05). Surprisingly, depletion of CD25⁺ cells from the tolerant spleen did not result in a statistically significant increase in allo-IgG titers (p > 0.05), despite a restoration of allograft rejection. We hypothesized that this is due to alloreactive B cells being deleted in the periphery of tolerant mice (20) and reasoned that removal of the regulatory T cells alone would not able to restore the alloantibody response.

To test the hypothesis that B cell deletion as well as dominant suppression of T cell help were responsible for controlling B cells responses in long-term tolerant recipients, we cotransferred naive 3-83 B cells along with spleen cells from tolerant BALB/c, with or without CD25 depletion. To avoid transferring contaminating naive T cells and 3-83 IgG-secreting plasma cells, we generated 3-83 Igi on the RAG2^−/− background and isolated 3-83 B cells from these mice. Of the five BALB/c RAG2^−/− recipients that received spleen cells from tolerant BALB/c without CD25 depletion, three had long-term graft survival (>60 days) and showed suppressed 3-83 IgG secretion (Fig. 7). The other two whose grafts were rejected showed comparable 3-83 IgG to the CD25-depleted group (data not shown). Four of the five BALB/c RAG2^−/− recipients (transplanted with C57BL/6 heart grafts) receiving CD25-depleted spleen cells along with 3-83 B cells rejected the heart allograft. All five showed a significant increase in 3-83 IgG titers. In all groups, no 3-83 IgM titers were detected (data not shown). We conclude from these experiments that the removal of CD25⁺ regulatory T cells from the spleen cells of tolerant BALB/c restored their ability to reject cardiac allografts and permitted the activation of alloreactive Th cells capable of providing help to alloreactive B cells. However, alloreactive B cells are deleted in long-term tolerant mice and the restoration of alloantibody responses required both the reconstitution of alloreactive B cells as well as the restoration of T cell help.

In the absence of immunosuppression, transplantation of solid organs across histocompatibility barriers is invariably followed by acute allograft rejection. Nonspecific immunosuppressive agents that inhibit all T cells, also leave the host vulnerable to severe infections and cancers (33). Therefore, a major goal in transplantation immunology is to induce donor-specific tolerance by designing short-term immunosuppressive therapies that will result in extended suppression of the immune responses specific for the allograft, but that will leave the rest of the immune system competent. Therapies that target the costimulatory molecules CD40-CD154, in particular, have been reported to induce long-term allograft survival in mice and in nonhuman primates (34). Different mechanisms directed at T cells have been identified to explain the tolerogenic effects of anti-CD154 mAbs (35). However, most reports currently support a model of dominant suppression by regulatory CD4⁺ T cells as the most important basis for tolerance induced by anti-CD154 mAbs (25, 26).

The effect of anti-CD154 on B cell responses is less well understood. Early studies indicated that anti-CD154 treatment could suppress B cell responses, but not induce B cell tolerance (36). In contrast, we have reported that anti-CD154 therapy, in the context of allograft tolerance models, can induce the long-term suppression of alloantibody and anti-Galactose-α(1–3) Galactose carbohydrate moiety Ab production (37). To address the different effects of anti-CD154 on B cell responses, we have examined the basis for how alloreactive B cell responses are controlled in a mouse heart allograft transplant model. Using a B cell transfer approach, we focused on defining the peripheral mechanisms of B cell tolerance. To circumvent possible contributions of B cell intrinsic mechanisms of tolerance that may have developed during the induction phase of tolerance, we transferred naive alloreactive B cells into tolerant or nontolerant mice that received an allograft on the day of B cell transfer. Using this model, we formally demonstrated that naive alloreactive B cells are constrained from maturing into Ab-secreting cells by the long-term absence of T cell help in mice made tolerant to C57BL/6 allografts. If T cell help was provided in the form of a third-party graft, then alloreactive naive 3-83 B cells were able to mature into Ab-secreting cells. Based on these observations and previous reports that regulatory T cells can control B cell responses (17, 18, 19), we postulate that the alloreactive-specific B cell responses are inhibited by regulatory T cells acting to constrain T cell help.

T cell help to B cells producing donor-reactive Abs is invoked, in this situation, by T cells stimulated by indirect Ag presentation. Although direct Ag presentation to alloreactive T cells can certainly play an important role in graft rejection, in terms of T cell help to B cells, the helper alloreactive T cells have to interact with host B cells presenting donor MHC peptides for alloantibodies to be produced. Certainly helper alloreactive T cells can interact with donor B cells presenting alloantigen by the direct pathway, but such B cells would not be able to secrete graft-reactive Abs, since the allograft would be “self”-relative to donor B cells directly presenting Ag to recipient alloreactive T cells. Therefore, T cells providing help to B cells producing alloantibodies must have been activated by the indirect pathway (by host B cells presenting allopeptide).

Dominant suppression of alloreactive T cell responses in vivo have been demonstrated by infectious tolerance and linked suppression (28, 29). Subsequent studies have revealed that this dominant suppression is mediated by CD4⁺, and, in many cases, by CD4⁺CD25⁺ regulatory T cells (30, 31). In this study, we demonstrate linked suppression of B cell responses in which the generation of Ab responses to a third-party Ag was constrained when the Ag was presented in conjunction with the tolerizing donor Ag. We note that the control of anti-C57BL/6 IgM and IgG, as well as 3-83 IgM, was better suppressed than the anti-C3H alloantibody response following the challenge with an F₁ graft. We speculate that this may be due, at least in part, to allo-IgG production being more T cell dependent than IgM responses. In addition, the weaker control of C3H responses, compared with the C57BL/6-specific responses, following transplantation of an F₁ graft can be explained by the efficacy of infectious tolerance (Fig. 8). We postulate that regulatory T cells inhibit the activation of Th cells by modifying the capacity of APCs to simulate conventional Th cells, as well as by directly suppressing Th cells when both T cell subsets engage with the same alloreactive B cell at the T-B cell interface. Thus, C57BL/6-specific regulatory T cells can function at both the APC and B cell levels when anti-C57BL/6-specific B cells present C57BL/6-derived peptides or when 3-83 B cells present both C57BL/6-derived and C3H-derived peptides to effector or regulatory T cells. However, C57BL/6-specific regulatory T cells can only function at the level of the APCs presenting both C57BL/6-derived and C3H-derived peptides to inhibit C3H-specific Th cells, but are unable to do so when C3H-specific Th cells interact with anti-C3H-specific B cells.

An important question in the field of allograft tolerance is where and how regulatory T cells exert their function in vivo. Early studies on regulatory T cells have focused on their ability to inhibit T cell activation and proliferation in vitro (38). These in vitro studies are supported by recent observations that regulatory T cells can inhibit priming events in the lymph node (39, 40). Using intravital two-photon microscopy to visualize the effect of regulatory T cells in vivo, both Tang et al. (39) and Tadokoro et al. (39) reported that the Ag-driven interaction between conventional T cells and APCs was significantly diminished in the presence of Ag-specific regulatory T cells. The consequence of the diminished interactions between conventional T cells and APCs is reduced priming and generation of pathogenic effector cells in vivo (39).

In contrast to the notion that regulatory T cells inhibit the priming of conventional CD25⁻ T cells, an increasing number of reports suggest that regulatory T cells may in fact mediate tolerance by inhibiting effector T cells in the nonlymphoid tissue sites (41). Chen et al. (41) reported that FoxP3⁺ regulatory T cells had no effect on the priming of diabetogenic T cells in the pancreatic lymph node, but prevented diabetes by inhibiting the function of effector T cells in the pancreatic islets. In transplantation models, the presence of cells capable of functional regulation was demonstrated by the ability of tolerant skin to confer donor-specific suppression when retransplanted into RAG^−/− mice (25). Subsequently, higher levels of FoxP3 mRNA, a marker of regulatory T cells, were observed in tolerized skin or heart grafts than in rejecting graft or isograft controls (42, 43). This was more recently confirmed by observations by Chen et al. (44) of increased numbers of CD4⁺FoxP3⁺ T cells in the tolerant compared with rejected heart grafts. Lee et al. (42) reported that inhibiting the accumulation of regulatory T cells in the graft prevents the development of allograft acceptance. Finally, Waldman and colleagues (45) reported that alloreactive TCR-transgenic CD8⁺ T cells proliferate normally but their ability to induce allograft rejection was associated with the censoring of immune effector function by regulatory T cells (45). Collectively, these data are consistent with the notion that the maintenance of long-term allograft tolerance is dependent on CD4⁺CD25⁺FoxP3⁺ regulatory T cells that inhibit the effector function of conventional T cells within the allograft.

Although T cells are primed in the secondary lymphoid organs and then migrate to the allograft to mediate rejection, optimal alloreactive B cell priming, class switching, and somatic hypermutation normally occur in secondary lymphoid organs, specifically in the germinal center (46). Ab-secreting cells and plasma cells remain in the spleen or migrate to the bone marrow, but generally do not have to infiltrate the allograft to facilitate graft rejection (47). We provide here histological evidence that the suppression of T cell help, and ultimately B cell responses, occurs within the spleen. The regulation of B cell responses can be visualized by the inhibition of the germinal center development in tolerant mice challenged with donor-specific or F₁ heart grafts. In contrast, a vigorous development of germinal centers was observed following challenge with third-party grafts. Based on these observations, we hypothesize that a subset of regulatory T cells within the secondary lymphoid organs inhibits the development of donor-specific T cell help in mice tolerant to allografts following treatment with anti-CD154/DST. This hypothesis is consistent with the in vitro observations by Lim et al. (19) that regulatory CD4⁺CD25⁺ T cells can migrate to B cell follicles and regulate Th cells that facilitate germinal center reactions and B cell responses, and with the report of Fields et al. (17) that regulatory CD4⁺CD25⁺ T cells can inhibit autoreactive anti-chromatin B cell Ab production, and that this inhibition is associated with inhibition of Th cell cytokine production of IL-10 and IFN-γ.

A unique feature of alloreactive B cell tolerance induced by anti-CD154/DST treatment is the restoration of acute allograft rejection following the removal of CD4⁺CD25⁺ T cells, but not the restoration of alloantibody responses. We interpret these observations as consistent with our recent report that alloreactive B cells are deleted in tolerant recipients (20). In those studies using 3-83 Igi mice as recipients of C57BL/6 heart grafts, we observed that the deletion of alloreactive B cells was at the mature B cell stage. This is in contrast to other models of B cell tolerance to self-Ags, which results in the deletion of self-reactive B cells at the immature/transitional stage in the periphery (48, 49). A number of caveats are associated with the use of the 3-83 Igi mice, with the high frequency of alloreactive B cells being the most critical. We were also unable to demonstrate changes in the titers of 3-83 Abs in naive 3-83 mice because of the spontaneously high titers of 3-83 IgG (data not shown). The basis for the high titers of 3-83 IgG is not currently understood, but may reflect the presence of plasma cells that secrete 3-83 IgG independently of T cell help.

We extend our previous observations by demonstrating here that the long-term inhibition of alloreactive B cell responses following the induction of allograft tolerance in wild-type BALB/c recipients with anti-CD154/DST treatment is maintained by the dominant suppression of T cell help as well as the deletion of alloreactive B cells in the periphery. In intact tolerant mice, the provision of T cell help permits new alloreactive B cell emigrants from the bone marrow to mature into Ab-secreting cells. In contrast, in the spleen cell transfer experiments into BALB/c RAG2^−/− recipients, the depletion of CD25⁺ regulatory T cells was able to restore T cell alloreactivity and allograft rejection, but not alloantibody production. We reasoned that this was because alloreactive B cells had been deleted in the periphery and that no new alloreactive B cells could be reconstituted by the BALB/c RAG2^−/− recipients. Tolerant spleen cells were able to suppress Ab production by cotransferred 3-83 naive alloreactive B cells, while the depletion of CD25⁺ regulatory T cells from the tolerant spleen cell population resulted in productive alloantibody responses. These observations are therefore consistent with the conclusion that the long-term inhibition of alloreactive B cell responses following the induction of allograft tolerance through anti-CD154/DST treatment is through dominant inhibition of T cell help and the deletion of alloreactive B cells in the periphery.

In conclusion, we report that active regulation and peripheral deletion of alloreactive B cells contribute to the control of alloantibody production in recipients made tolerant to cardiac allografts by treatment with anti-CD154/DST. We speculate that the absence of T cell help in the presence of allograft facilitates the deletion of mature alloreactive B cells in the periphery. The reduced frequency of alloreactive B cells, likewise, ensures a tight control of Ab responses in the periphery and reduces the probability of the inadvertent activation of alloreactive B cells responses by T-independent mechanisms.

We thank Dr. Roberta Pelanda (Department of Immunology, National Jewish Medical and Research Center, Denver) for generously providing 3-83 Igi mice, the anti-Id 54.1 mAb, and for technical advice. We also thank Jamie Kim, Lucy Deriy and Ting-ting Zhou for technical assistance, and Dr. Ian Boussy for help in preparation of this manuscript.

The authors have no financial conflict of interest.