Anti-CD40L mAb plus bone marrow transplantation (BMT) and recipient CD8 T-cell depletion permits long-term mixed hematopoietic chimerism and systemic donor-specific tolerance to be achieved across full MHC barriers. Initial tolerance is characterized by peripheral deletion of donor-reactive CD4 cells. In regimens using costimulatory blockade without BMT to achieve allograft survival, cyclosporine inhibited graft survival, suggesting that the combination may not be clinically applicable. We assessed the role of cyclosporine-sensitive mechanisms and the mechanisms of T-cell apoptosis involved in the induction of early peripheral CD4+ T-cell tolerance by BMT with anti-CD40L. Neither a short course of cyclosporine (14 days) nor the absence of FAS-mediated activation-induced cell death (AICD) blocked the induction or maintenance of donor-specific tolerance. IL-2 production was not associated with tolerance induction, consistent with the lack of a role for Fas-mediated AICD. Mice in which passive T-cell death was impaired because of constitutive expression of a Bcl-xL transgene did not develop tolerance with this protocol. These data confirm that deletion of donor-reactive T cells is critical for the induction of mixed chimerism and tolerance. However, the mechanisms involved may differ from those involved in costimulatory blockade regimens that do not include BMT.
Blockade of the CD40-CD40L pathway either alone or in combination with other costimulatory blocking reagents can lead to the significant prolongation and, under some conditions, functional tolerance to various allografts [reviewed in (1)]. However, with few exceptions (2), most of these approaches do not lead to primary skin graft acceptance across full MHC barriers in euthymic animals, the most stringent measure of tolerance. Recently, we and others have combined the use of costimulatory blockade with bone marrow transplantation (BMT) to induce high levels of donor hematopoietic chimerism across full MHC barriers (3–6), with or without the addition of multiple minor histoincompatibilities. Mixed chimerism establishes a systemic state of donor-specific tolerance that lasts the life of the recipient, leads to specific in vitro unresponsiveness to the donor, and permits the acceptance of primary donor skin grafts in euthymic mice [reviewed in (7)]. Long-term donor-specific tolerance in mixed chimeras is maintained through the establishment of central deletion, in which newly developing thymocytes are subject to negative selection by donor-derived APCs within the thymus (3,4,8–10). The use of costimulatory blocking reagents with BMT obviates the need for peripheral T-cell-depleting mAbs, a necessity in previous protocols (8–11), for the induction of chimerism, indicating that the pre-existing T-cell repertoire is tolerized with this approach.
The mechanisms by which peripheral T cells are rendered nonresponsive to the BMT donor in such recipients are undefined. In various experimental models using costimulatory blockade, anergy (12), immune deviation (13), suppression (14), and apoptosis (2,15) have all been implicated as playing a role in prolonging graft survival. In a model involving BMT with costimulatory blockade (MR1 and CTLA4Ig), we have previously described the thymus-independent early deletion of donor-reactive T cells in the peripheral lymphoid tissues (3,16). In this model, the deletion of donor-reactive CD4+ T cells progresses over time, and the early (1 week post-BMT) deletion has been shown to have features of both passive cell death (PCD) and activation-induced cell death (AICD) (17). The suggestion that both pathways may play an important role is of interest, as their respective mechanisms are still incompletely defined for many in vivo systems. While both pathways share a similar end result of apoptosis, passive cell death, or ‘death by neglect’, is the result of necessary survival signals (antigen stimulation, costimulatory receptor signaling, cytokines, etc.) being withheld from T cells and the lack of expression of anti-apoptotic proteins (Bcl family), whereas activation-induced cell death occurs when activated T cells express receptors that may deliver pro-apoptotic (e.g. Fas) signals to the T cell [reviewed in (18)]. Data involving costimulatory blockade protocols have also suggested a role for both pathways of T-cell apoptosis in the establishment of peripheral tolerance (2,15,19,20) and impairment of graft prolongation has been reported in recipients in which AICD is blocked by calcineurin inhibitors (2,21,22). Therefore, it is important to understand the role of each type of apoptosis in the induction of lasting chimerism and tolerance using costimulatory blockade and BMT.
We have previously demonstrated that when recipient CD8+ T cells are depleted, mixed chimerism can reliably be established in mice receiving only 3 Gy TBI and one injection of anti-CD40L mAb (23), and that the peripheral CD4+ T-cell population is rapidly tolerized by donor antigens by blocking CD40/CD40L interactions (24). The low dose of TBI can be removed from the regimen if the marrow dose is increased approximately 10-fold (H. Ito et al., unpublished data). In light of the early ‘peripheral’ tolerance component that affects the CD4+ T-cell population, we wished to address the relative roles played by AICD and PCD in the induction of CD4+ T-cell tolerance using BMT and anti-CD40L.
Materials and Methods
Female C57BL/6 (B6; H-2b), B10.A (H-2a), B10.RIII (H-2r), A.SW (H-2s) and C57BL/6lpr/lpr (B6-lpr) mice were purchased from Frederick Cancer Research Center (Frederick, MD) or from The Jackson Laboratory (Bar Harbor, ME). B6-Bcl-xL and littermate control mice were received from Dr Turka, University of Pennsylvania. Mice were housed in a specific pathogen-free microisolator environment, as previously described (25).
Conditioning and bone marrow transplantation
Age-matched (8–14 weeks old) B6 (or, where indicated, B6-Bcl-xL or B6-lpr) recipient mice were treated with a nonmyeloablative dose of total body irradiation (TBI, 3 Gy) followed by an intravenous injection (tail vein) of unseparated MHC-mismatched B10.A bone marrow cells (BMCs). Hamster antimouse CD40L mAb (MR1; 0.5 mg or 2.0 mg; some studies were performed with the higher dose because occasionally a mouse failed to achieve chimerism with the lower dose) was administered i.p. on Day 0 with respect to BMT. The MR1 hybridoma was kindly provided by Dr Randolph J. Noelle (Dartmouth University, NH). Depleting anti-CD8 mAb (2.43; 0.35 mg/mouse) was administered i.p. on Day −1 where indicated. CsA (SandImmune; Novartis Pharma AG, Basel, Switzerland) was administered s.c. at a dose of 20 mg/kg/day for the first 2 weeks after BMT where indicated.
Flow cytometric analysis of multilineage chimerism in white blood cells
Flow cytometric (FCM) analysis of multilineage chimerism was performed as previously described (26). Single-cell peripheral white blood cell (WBC) suspensions were prepared in medium consisting of 1X HBSS, 0.1% sodium azide, and 0.1% BSA. In brief, forward angle and 90° light scatter properties were used to distinguish lymphocytes, monocytes, and granulocytes in peripheral WBC. Two-color FCM was used to distinguish donor and host cells of particular lineages, and the percentages of donor cells were calculated as previously described (26), by subtracting control staining from quadrants containing donor and host cells expressing a particular lineage marker, and by dividing the net percentage of donor cells by the total net percentage of donor plus host cells of that lineage. Dead cells were excluded using propidium iodide staining. Donor-derived cells were identified by biotinylated anti-H-2Dd mAb 34-2-12 developed with phycoerythin-streptavidin (PEA). Non-reactive mAb HOPC1-biotin was used as a negative control. Cell lineages were distinguished by fluorescein isothiocyanate (FITC)-conjugated mAbs, including anti-CD4, anti-CD8, anti-B220 (all purchased from PharMingen, San Diego, CA), and anti-Mac1 (purchased from Caltag, San Francisco, CA). Negative control mAb HOPC-FITC was prepared, purified, and conjugated in our laboratory. Non-specific FcγR binding was blocked by antimouse FcγR mAb 2.4G2 (27).
FCM analysis of T-cell receptor V β usage
Peripheral blood lymphocytes (PBL) were stained with FITC-conjugated anti-Vβ 5.1/2, Vβ11, and Vβ8.1/2 mAbs vs. PE-conjugated anti-CD4 mAb (all purchased from PharMingen). Nonspecific PE-conjugated rat IgG2a (PharMingen) served as a negative control. Two-color FCM analysis was performed on gated CD4+ cells. Background staining (as determined by nonreactive mAb HOPC-FITC) was subtracted from the percentage of cells staining with each anti-Vβ mAb.
Mixed lymphocyte reaction assay
Mixed lymphocyte reactions (MLR) were performed as described (28). In brief, splenocytes suspended in medium consisting of RPMI 1640 (Mediatech, Herdon, VA), 10% CPSR-2 serum-free medium (Sigma, St. Louis, MO), 0.09 mm nonessential amino acids, 2 mm l-glutamine, 1 mm sodium pyruvate, 100 U/mL penicillin, 100 µg streptomycin, 0.05 mm 2-mercaptoenthanol, and 0.01 m HEPES buffer, were cultured in triplicate wells containing 4 × 105 responder cells with 4 × 105 stimulators (irradiated with 3000 cGy). After 3–4 days, the wells were pulsed with [3H] thymidine, and then harvested 15–18 h later. Stimulation indices (SI) were calculated by dividing c.p.m. obtained for anti-donor and anti-third party responses by those obtained for anti-self (host in the case of bone marrow chimeras) responses. These control responses were similar in c.p.m. to those obtained with responder cells cultured in the absence of stimulator cells.
ELISA spot assay
Ninety-six-well ELISA spot plates (Polyfiltronics, Rockland, MA) were coated overnight with a capture mAb in sterile PBS. Anti-IL-2 was used at 3 mg/mL (PharMingen). On the day of the experiment, the plates were washed twice with sterile PBS, blocked for 1.5 h with PBS containing 1% BSA, then washed three times with sterile PBS. Responder cells were added to wells previously filled with intact donor cells. Cells were incubated for 48 h. The plates were washed three times with PBS, then four times with PBS containing 0.025% Tween (PBST). Biotinylated anti-IL-2 detection mAb were added at 2 mg/mL (JES6–5H4) (PharMingen) and incubated either for 5 h at room temperature or overnight at 4 °C. After washing three times with PBST, avidin-horseradish peroxidase (1/2000) was added to each well for 1.5 h. Four washes with PBS were performed before the spots were revealed by the addition of the developing solution composed of 800 μL of 3-amino 9-ethylcarbazole (AEC) (Sigma, 10 mg dissolved in 1 mL dimethylformamide) in 24 mL of 0.1 m sodium acetate, pH 5.0, catalyzed by 12 μL H2O2. The resulting spots were counted and analyzed on a computer-assisted ELISA spot image analyzer (C.T.L., Cleveland, OH).
Full thickness tail skin (∼1.0 cm2) from B10.A (donor-specific) and fully MHC-mismatched third party (either A.SW or B10.RIII) mice was grafted on the dorsal thoracic wall, sutured with 5–0 silk and bandaged, and followed by daily visual inspection from Day 7 onwards. Grafts were defined as rejected when <10% of the graft remained viable.
Short course of CsA administration does not impair the establishment of mixed chimerism or tolerance
When recipient CD8+ cells are depleted, a single injection of anti-CD40L mAb MR1 is able to completely overcome CD4+ T cell-mediated resistance to allogeneic BMT in 3 Gy irradiated mice, and lasting mixed chimerism and tolerance are reliably induced (23). This donor-specific tolerance develops rapidly, and is not the result of specific targeting or signaling of activated CD4+ T cells by anti-CD40L mAb (24). Other groups using costimulatory blocking reagents in regimens aimed at inducing tolerance have reported that the inclusion of calcineurin inhibitors (i.e. CsA, FK506) impairs the prolongation of graft survival (2,21,22). To address this in our protocol for inducing mixed chimerism and tolerance, we treated B6 mice with our standard regimen (3 Gy TBI, anti-CD8 mAb, 20 × 106 B10.A BMCs, and MR1) and compared the incidence of chimerism and the deletion of donor-reactive CD4+ T cells with that in a group receiving the same protocol plus daily CsA treatment (20 mg/kg) beginning on the day of BMT and continuing for 14 days. This CsA dosing regimen has been shown by other groups to impair allograft survival in mice receiving costimulatory blockade (2,21). In recipients of the standard regimen, 80% of the mice demonstrated long-term multilineage chimerism for at least 27 weeks post-BMT (Figure 1A). In the group of mice that received the standard regimen plus CsA, 90% of recipients became long-term chimeras (Figure 1A), suggesting that the addition of a calcineurin inhibitor did not impair the induction of CD4+ T-cell tolerance. Treatment with CsA without MR1 did not permit the achievement of chimerism (Figure 1A).
To further evaluate the effect of CsA on the induction of tolerance, we assessed the deletion of donor-reactive CD4+ T cells in the peripheral blood of recipients of BMT and anti-CD40L with or without CsA treatment. In this experiment, the donor strain B10.A expresses the MHC class II gene I-E, whereas the recipient strain, B6, does not. I-E is necessary to present superantigens encoded by the mammary tumor virus (Mtv)-8 and -9 retroviruses, which are encoded in both the B10.A and B6 genomes. These superantigens bind to thymocytes expressing T-cell receptors (TCRs) containing Vβ5 or Vβ11, resulting in their deletion during intrathymic development and establishing a hole in the repertoire of I-E+ strains, such as B10.A (29–31), but not of I-E– strains, such as B6 (31,32). Thus, Vβ5.1/2 and Vβ11 expression on B6 CD4+ T cells can be used as a marker for the presence of donor-reactive cells. Unfortunately, CD8+ cells appear to play a critical role in producing the Mtv determinants required to delete Vβ5+ and Vβ11+ T cells, so CD8+-cell depletion precludes the ability to examine early (week 1) Vβ5+ and Vβ11+ deletion in the peripheral blood after BMT in mice treated with anti-CD40L and anti-CD8 mAb (23). Such deletion is usually apparent between 2 and 3 weeks post-BMT (24). CsA treatment delayed the peripheral deletion of donor-reactive CD4+ T cells, as the early (week 2) deletion of Vβ11+ (Figure 1B) and Vβ5+ (not shown) CD4+ T cells was significantly impaired in mice treated with CsA compared with mice that received only the standard protocol, as previously reported (17). Over time, the deletion of these cells progressed in a similar fashion to that in mice not receiving CsA (p > 0.05). In mice receiving anti-CD8 mAb and CsA alone (without MR1), no deletion of donor-reactive Vβ was observed at any time and long-term chimerism was not achieved (data not shown).
To further evaluate the establishment of tolerance in mice treated with BMT, costimulatory blockade, and CsA, MLR assays were performed using spleen cells at the time of sacrifice (27 weeks post-BMT). As shown in Figure 1 (C), mice receiving TBI, anti-CD40L, and anti-CD8 with or without CsA showed donor-specific MLR tolerance. Maintenance of tolerance depended on the persistence of chimerism, as the one mouse that received the standard treatment and CsA but had no detectable chimerism at the time of sacrifice demonstrated a significant response to donor antigens, similar to the nonchimeric recipient that did not receive CsA (data not shown). Responses of all chimeric animals to third party (B10.RIII) stimulator cells were comparable, regardless of CsA treatment. Cell-mediated lympholysis (CML) assays were also performed at the time of sacrifice to assess cytotoxic T lymphocyte (CTL) responses to donor and third party antigens. All chimeras demonstrated unresponsiveness to donor (B10.A) antigens, while showing measurable antithird party responses (data not shown), further supporting the presence of specific tolerance in these mice.
Mixed allogeneic chimerism and donor-specific tolerance are established in Fas-deficient mice depleted of CD8+ cells and treated with anti-CD40L mAb plus 3Gy TBI
As CSA blocks AICD, a major pathway of which is dependent on Fas, we evaluated the ability to achieve chimerism and tolerance in Fas-deficient recipients using the anti-CD8/MR1 BMT regimen. We have previously demonstrated that lpr (Fas-deficient) recipients show an impairment of the early (week 1) deletion of donor-reactive CD4+ T cells in a regimen using anti-CD40L and CTLA4Ig plus allogeneic BMT (17). To determine whether or not this impairment would affect the induction of tolerance in the CD4+ population and the establishment of mixed chimerism, we used Fas-deficient (B6-lpr) and WT (B6) mice that were depleted of CD8+ T cells with anti-CD8 mAb 2.43 as recipients of fully MHC-mismatched allogeneic BMCs with 3Gy TBI and one injection of anti-CD40L. As shown in Figure 2 (A), all lpr mice that received CD8-depleting mAb, 3 Gy TBI, and MR1 developed lasting donor chimerism (5/5), similar to WT control mice receiving the same treatment (5/5). Of note, the levels of donor chimerism in various cell lineages differed between the lpr and the WT recipients (see below). In contrast, lpr mice receiving BMT and anti-CD8 mAb treatment without anti-CD40L rejected donor marrow. These results suggest that Fas-mediated AICD is not essential in order to overcome the CD4+ T-cell barrier to allogeneic marrow engraftment with anti-CD40L.
To assess tolerance in an additional assay, recipient mice were grafted with both donor (B10.A) and third party (B10.RIII) tail skin 1 day post-BMT. All lpr and WT mice receiving BMT and anti-CD40L following treatment with anti-CD8 mAb and 3Gy TBI accepted donor skin grafts >140 days, while rejecting third party grafts (Figure 2B). In contrast, lpr mice treated with only anti-CD8, TBI, and BMT (without anti-CD40L) rejected their donor skin grafts, demonstrating an essential role for the MR1 in tolerance induction. All mice, regardless of treatment, were able to reject third party grafts in a similar fashion, demonstrating that the mice were not globally immunosuppressed and that BMT played a critical role in tolerance induction in mice receiving 3Gy TBI and MR1. Thus, Fas-deficient recipients of fully MHC-mismatched allogeneic BMCs with this regimen develop specific tolerance to donor antigens.
Deletion of donor-reactive CD4+ T cells with BMT and anti-CD40L in Fas-deficient recipients
We also assessed the deletion of donor-reactive CD4+ T cells in the peripheral blood of both WT and lpr recipients of BMT with anti-CD40L to determine whether or not Fas expression on recipient CD4+ T cells was required for their deletion. In this experiment, all chimeric mice, including lpr and WT recipients, demonstrated similar and statistically significant deletion of Vβ5 and Vβ11 CD4+ T cells by 4 weeks post-BMT (earliest time point checked in peripheral blood; Figure 3 and data not shown). This deletion progressed over time to levels that were comparable to those in normal B10.A mice (by week 16 post-BMT). In lpr mice that received BMT without anti-CD40L mAb, no deletion or expansion of either subset was observed at any time (data not shown and Figure 3). As a specificity control, the percentages of Vβ8.1/2+ CD4+ T cells were followed, and no significant changes were observed (Figure 3 insert). Thus, the deletion of donor-reactive CD4+ T cells occurs in both lpr and WT mice treated with anti-CD40L, suggesting that Fas-mediated AICD is not necessary for the progressive deletion or establishment of tolerance in mice treated with MR1 and BMT.
Donor-specific tolerance develops rapidly in Fas-deficient and WT recipients of BMT with anti-CD40L
As noted earlier, lpr and WT recipients treated with anti-CD8 mAb, 3 Gy TBI, and anti-CD40L specifically accepted donor skin grafted on Day +1, suggesting that tolerance was established early in these mice. However, third party skin grafted on Day +1 was rejected in a relatively slow manner, making it difficult to ascertain the early timing of the development of donor-specific tolerance. We have previously shown that in WT recipients treated with anti-CD8, 3 Gy TBI, and MR1, tolerance develops rapidly in the periphery, as measured by MLR assay (24). We therefore evaluated the establishment of CD4+ T-cell tolerance in both lpr and WT mice sacrificed at either 2 or 20 weeks post-BMT. For this purpose, MLR assays were performed using spleen cells from mixed chimeras at the time of sacrifice. As shown in Figure 4 (C,F), nonresponsiveness to donor antigens was observed in both lpr and WT chimeras at 2 and 20 weeks post-BMT, respectively, while responses to third party were measurable, although somewhat weak for the lpr chimeras. These results are consistent with the specific acceptance of donor skin grafted 1 day after BMT. Additionally, this evidence of early donor-specific tolerance in the absence of significant detectable deletion of donor-reactive CD4+ T cells in both lpr and WT recipients observed at 2 weeks post-BMT, as shown in Figure 4 (B), further supports the conclusion that Fas-mediated AICD is not required for the induction of CD4+ T-cell tolerance using BMT and anti-CD40L mAb treatment. By 20 weeks post-BMT, the levels of Vβ5+ and Vβ11+ CD4+ T cells in spleens of lpr-mixed chimeras were comparable to those in normal B10.A mice (Figure 4E).
Lack of T-cell chimerism in lpr recipients of allogeneic and syngeneic BMT
Although long-term lpr mixed chimeras demonstrated lasting donor-specific tolerance, there was a discrepancy in the levels of donor chimerism in various lineages. At 2 weeks post-BMT, lpr and WT recipients that received the full regimen demonstrated similar levels of donor CD4+, B (CD19+), and Mac1+ cells in the bone marrow (Figure 4A). Very few mature donor thymocytes (TCRhigh, MHChigh) were observed in animals in either group. Among mice sacrificed 20 weeks post-BMT, WT chimeric recipients demonstrated high levels of donor CD4+, CD8+, and B cells in the spleen, mature thymocytes, and Mac1+ cells in the BM, whereas chimeric lpr recipients demonstrated few to no CD4+ or CD8+ donor T cells in the spleen. While variable levels of B cells in the spleen, Mac1+ cells in the bone marrow, and mature thymocytes (Figure 4D) were observed in this group, there was a striking disparity between T-cell chimerism in the periphery (both spleen and peripheral blood, Figure 5A) and chimerism of other lineages (B cells, granulocytes, and Mac1+ BMCs) in the same compartments (Figures 4D and 5A) in animals with high levels of chimerism of the latter types. These animals also had high levels of thymocyte chimerism, suggesting that the low peripheral T-cell chimerism results from an event in the periphery.
This lack of stable donor T-cell chimerism in lpr recipients led us to question whether incomplete tolerance or a competitive advantage for lpr T cells over WT T cells might be responsible for this selective defect in T-cell chimerism. To distinguish these possibilities, we performed BMT in a syngeneic model system in which alloreactivity could not limit donor engraftment. 20 × 106 WT BMCs from B6-CD45.1+ donors were administered to either B6-lpr CD45.2+ or WT B6 CD45.2+ recipients treated with 3 Gy TBI, and various WBC populations were analyzed over time by flow cytometry. In the WT (CD45.1)(→)WT (CD45.2) combination, high levels of CD45.1+ cells were detected in all lineages, including T cells, at all times (Figure 5B) as previously reported (33). In contrast, in the WT (CD45.1)(→)lpr (CD45.2) combination, only very low levels of WT T cells were detectable (Figure 5B) in the peripheral blood, similar to results in the allogeneic model. WT B cells, monocytes, and granulocytes were detectable at all times, although the levels of B-cell and monocyte chimerism were significantly lower than those seen in the WT(→)WT group, similar to results in the allogeneic model. As in the allogeneic model (Figure 4D), both WT and lpr recipients in the syngeneic model had high levels of donor WT CD45.1 mature (MHChighTCRhigh) thymocytes detectable by FACS at the time of sacrifice (week 23; 90.3–97.7% and 32.8–78.9% CD45.1+, respectively), demonstrating that donor Fas+ T-cell progenitors were able to populate the recipient thymus but not the periphery. Collectively, these results suggest that the lack of T-cell chimerism observed in the B10.A(→)B6-lpr chimeras is most likely the result of a proliferation or survival advantage in the periphery of lpr-derived cells over WT cells, and that it is not the result of an incomplete tolerance to the allogeneic donor in the allogeneic model.
Specific absence of early IL-2 production in response to donor antigens in mixed chimeras
High concentrations of the growth factor IL-2 enhance the expression of Fas ligand on activated T cells and increase sensitivity to Fas-mediated apoptosis by down-regulating the inhibitory protein, FLIP, that is associated with the Fas receptor (34–36). Thus, IL-2 production in response to donor antigens would be expected to play a role if AICD were important in the induction of tolerance with anti-CD40L and BMT. We therefore investigated the ability of peripheral CD4+ T cells from animals treated with anti-CD40L, anti-CD8 mAb, 3 Gy TBI and BMT to produce IL-2 when stimulated in vitro with donor antigens at various times post-BMT. Animals that received the full regimen and various control groups were sacrificed 2, 4, 8, 15, or 36 days following BMT, and spleen cells were used in ELISPOT assays to assess the number of IL-2-producing cells upon stimulation with donor, host, or third party alloantigens. In all animals sacrificed, flow cytometric analysis was also performed to assess levels of donor chimerism, and these were similar to previously published results (24). As shown in Table 1, animals that received the full regimen plus BMT had similar numbers of IL-2-producing cells in response to B10.A stimulation as to self (B6) stimulation. By Day 4 post-BMT, this nonresponsiveness was specific, as chimeras mounted IL-2 responses to third party stimulators that were similar to those of naïve B6 mice and nontolerant conditioned controls. The specific lack of IL-2 production required persistent donor chimerism, as demonstrated by the presence in a recipient that failed to become chimeric after receiving the full protocol (Chimera 3, Day 15) of IL-2-producing cells at a frequency similar to that in an animal receiving anti-CD8, anti-CD40L, and TBI without BMT. The necessity of anti-CD40L tolerance is evident in the ability of mice that received TBI, anti-CD8, and BMT to produce IL-2 in response to B10.A stimulation at 4 days post-BMT, and in their sensitized responses to donor antigens, which becomes apparent at 15 days post-BMT.
Table 1. Splenic IL-2 ELISPOT assay in mice receiving BMT with anti-CD8, anti-CD40L, and 3 Gy TBI
Each data represents the mean number of spots (2 wells) obtained with 1 × 106 initial splenocytes
“Chimeras” received 3 Gy TBI, anti-CD8 mAb, MR1 and BMT; “Conditioned controls” received TBI, anti-CD8, MR1
“BMT with no MR1” received TBI, anti-CD8, and BMT
“nd” means not done
Conditioned control (no BMT)
BMT with no MR1
Role for passive cell death in the induction of mixed chimerism and tolerance
We have previously suggested that both AICD and PCD play a role in the early peripheral deletion of donor-reactive CD4+ T cells in mice receiving costimulatory blockade and BMT (17). As the data described earlier suggest that CsA-sensitive mechanisms and FAS-mediated AICD are not involved in the induction of donor-specific peripheral CD4+ tolerance with anti-CD40L and BMT, we addressed the possible role of PCD in the induction of mixed chimerism and tolerance. WT B6 recipients and B6 mice that constitutively expressed Bcl-xL in the T-cell lineage (37) were treated with the full regimen and followed by flow cytometric analysis and skin graft acceptance. As shown in Figure 6 (A), six of seven WT mice that received CD8-depleting mAb, 3 Gy TBI, and MR1 developed lasting donor chimerism. In contrast, only one of nine Bcl-xL recipient mice receiving the full treatment developed lasting mixed chimerism. Of note, several Bcl-xL recipients initially had detectable donor cells (by flow cytometry) that steadily declined in percentage to undetectable levels by week 20.
In this experiment, recipient mice were skin grafted with both donor (B10.A) and third party (A.SW) tail skin 1 day post-BMT to assess the induction of tolerance. All WT long-term chimeras produced with BMT and anti-CD40L plus anti-CD8 mAb accepted donor skin grafts >200 days, while rejecting third party grafts (Figure 6B). In contrast, all Bcl-xL mice rejected their donor skin grafts within 55 days post-BMT. Although the transgenic mice that had the most prolonged detectable chimerism also had the most prolonged donor skin graft survival, the donor skin graft was fully rejected even in the tg mouse that showed sustained chimerism. Both WT and Bcl-xL tg recipients rejected third party grafts in a similar fashion, and no significant difference was found between the rejection of donor and third party grafts by Bcl-xL recipients. When the deletion of donor-reactive Vβ11+ CD4+ T cells in the peripheral blood was examined, WT long-term chimeras again showed progressive deletion over time to levels comparable to those in normal B10.A mice (Figure 7). The single WT recipient that lost chimerism showed initial deletion similar to long-term chimeras while donor cells could be detected (Figure 7, gray bars), but these Vβ returned to levels similar to those in normal B6 mice after donor chimerism was lost (weeks 15–28, dotted bars). In Bcl-xL tg recipients, mice that never had detectable donor chimerism also had no detectable deletion of donor-reactive CD4+ T cells at any time point (Figure 7, non-chm). Those Bcl-xL tg recipients that had some donor chimerism (but that became undetectable after Week 13 post-BMT) showed partial deletion over time, but these Vβ recovered to normal levels after chimerism was lost (Figure 7, transient chm). The level of deletion in these mice at week 3 was significantly less than that observed in WT chimeras (p < 0.005), although levels of donor chimerism were also significantly lower (data not shown). The single Bcl-xL tg recipient that developed long-term chimerism demonstrated a deletion profile that was similar to that of WT recipients, although these Vβ were slightly higher than those of WT recipients at week 3 (e.g. %Vβ11+ CD4+ = 3.16; data not shown).
At 29 weeks post-BMT, mice from both groups were sacrificed and assessed for chimerism and tolerance in vitro. WT chimeras demonstrated chimerism in all lineages tested and were specifically nonresponsive to donor antigens in MLR and CML assays, as previously reported (24) (data not shown). Bcl-xL recipients that had no detectable chimerism had measurable responses to both donor and third party antigens in MLR and CML assays, and the single long-term chimeric Bcl-xL recipient was nonresponsive to both donor and third party antigens (data not shown). Overall, these data suggest that long-term mixed chimerism and specific tolerance cannot be readily achieved in Bcl-xL mice using BMT and anti-CD40L.
Previous protocols for the induction of mixed hematopoietic chimerism have required the exhaustive depletion of peripheral T cells to prevent the rejection of donor hematopoietic cells and allow the repopulation of the periphery with a newly developed pool of T cells that have undergone negative selection (central deletion) in the presence of donor-derived thymic immigrants (7). The utility of costimulatory blocking protocols for inducing peripheral hyporesponsiveness of recipient T cells to donor antigens has been demonstrated in numerous models (1,38). Depending on the model being studied, costimulatory blockade appears to exert its effect on peripheral T cells through various mechanisms, including anergy (12), immune deviation (13), suppression (14), and apoptosis (2,15,19,20). When combined with BMT, costimulatory blockade allows the establishment of long-term mixed chimerism and systemic donor-specific tolerance using the most stringent tests of tolerance (3–6,23). In this model, we have previously demonstrated early, thymus-independent deletion of donor-reactive CD4+ T cells, which leads to the eventual elimination of these cells from the periphery (3,16). We have now evaluated the mechanisms involved in the establishment of early donor-specific deletional tolerance, specifically the roles of CsA-sensitive, FAS-dependent and other mechanisms of CD4 T-cell death.
The question of CsA sensitivity of tolerance induction is of particular importance, as the ability to combine novel tolerance-inducing therapies with current immunosuppressive drugs would greatly facilitate the adoption of these protocols within the clinical setting. CsA, which inhibits calcineurin-induced NFAT activation via TCR signaling and blocks the induction of T-cell anergy and apoptosis (2,39,40), has been reported to be detrimental to the prolongation of graft survival in various protocols using costimulatory blockade (2,21,22). One critical function of calcineurin is the dephosphorylation and activation of NFAT, a key transcription factor involved in the production of IL-2 (41), which promotes FAS-mediated AICD by down-regulating the amount of FLIP, the down-regulator of apoptosis, associated with the Fas molecule (36). When we added a short course of CsA (14 days) to treatment with TBI, anti-CD8, anti-CD40L and BMT, we found that the early deletion of donor-reactive CD4+ T cells was impaired (Figure 1B) as previously described (17). However, long-term mixed chimerism and donor-specific tolerance was still established (Figure 1A,C). This contrasts with other costimulatory blocking regimens that do not include BMT, in which a similar dose of CsA impaired graft survival (21). Furthermore, the result contrasts with those in our own studies involving MR1 and CTLA4Ig in combination with allogeneic BMT, without CD8 depleting mAb, in which the addition of a similar course of CsA to the regimen completely blocked tolerance induction and the long-term maintenance of chimerism (42). As a major difference between the two protocols is the requirement for toleration of peripheral CD8 cells in the model in which this subset is not depleted, we hypothesize that CsA may specifically block the tolerization of CD8, but not CD4 T cells, in the periphery of mice receiving BMT with costimulatory blockade. Our results show similarities to and differences from those of Blazar et al. in a model involving BMT with repeated doses of anti-CD40L alone, in which donor skin grafts were accepted in less than half the control recipients and CyA had no obvious effect on tolerance, but decreased donor T-cell reconstitution (43), which we did not observe in our model.
CsA-sensitive pathways have been implicated in playing a critical role in the mechanism of activation-induced cell death (18). Activation-induced cell death can be triggered by restimulating activated T cells with high concentrations of antigen (18,44). Activation-induced cell death has been suggested to play a role in restricting the expansion of recently activated T cells, and in the maintenance of self-tolerance by eliminating auto-reactive lymphocytes (44,45). Activation-induced cell death is often Fas (CD95)-dependent (44–46) (though Fas-independent pathways for AICD also exist), is promoted by the presence of IL-2 (35), and can be inhibited by CsA (39,40). In protocols for inducing peripheral T-cell tolerance, AICD has been suggested to be necessary to allow the long-term survival of donor tissues, most likely through the contraction of the donor-reactive T-cell pool (18,47). Furthermore, in a regimen using BMT with anti-CD40L and CTLA4Ig, we have demonstrated that Fas-deficient T cells are impaired in the early (1 week post-BMT) deletion of donor-reactive CD4+ T cells (17), suggesting that AICD may also play an important role for the induction of mixed chimerism using costimulatory blockade. To investigate the role of AICD in the induction of peripheral CD4+ T-cell tolerance in our BMT plus anti-CD40L protocol, we compared B6 WT and B6.lpr (Fas-deficient) mice as recipients of allo-BMT. In these experiments, Fas-deficient recipients developed long-term mixed chimerism, donor-specific tolerance and deletion of donor-reactive CD4+ T cells as readily as WT mice. Thus, Fas-mediated apoptosis is not crucial for either the deletion of these donor-reactive cells, or for other mechanisms establishing tolerance in this model. Although these results differ from those in the model involving CD40L plus CTLA4Ig, the addition of CD8-depleting mAb precludes the observation of the same early (1 week post-BMT) deletion of donor-reactive T cells (23). While Fas-mediated deletion may still occur in this model, these new data demonstrate that it is not necessary for the induction of mixed chimerism or CD4 cell tolerance following BMT. These results concur with those of Honey et al. in a minor antigen plus Mls-mismatched BMT model using nonlytic anti-CD4 and anti-CD8 mAbs, in which Fas was shown not to be required for the deletion of antigen-specific T cells (48). The lack of a role for Fas-mediated mechanisms in the early induction of donor-specific tolerance is further demonstrated by the long-term acceptance of donor skin grafted on Day +1 following BMT (Figure 2B), and by the specific nonresponsiveness to donor antigens in MLR assays performed 2 weeks post-BMT (Figure 3C).
The observation that donor T cells did not persist in the periphery of lpr recipients of allo-BMT was shown not the result of residual alloresponsiveness, but instead of an intrinsic advantage of lpr T cells over WT T cells, as shown in the syngeneic WT CD45.1 into WT and lpr CD45.2 experiment. In the allogeneic combination, donor thymocytes were detected at levels in some cases as high as in the WT recipients, although after 5 weeks post-BMT, no detectable donor T cells were observed in the periphery. This defect is most likely the result of a peripheral advantage, as in the syngeneic model, both WT and lpr recipients had substantial levels of donor WT CD45.1 mature thymocytes at the time of sacrifice (week 23; 90.3–97.7% and 32.8–78.9% CD45.1+, respectively). In the periphery of these animals, lpr recipients had little to no detectable WT CD45.1+ T cell chimerism in the blood (Figure 5B) or the spleen at the time of sacrifice (data not shown). These findings extend those described by Laouar et al. (49), in which a competitive advantage was also seen for Fas-deficient cells in a syngeneic model.
An additional hallmark of Fas-mediated AICD is its promotion by the addition of exogenous IL-2 (18). When we examined production of IL-2 in response to donor antigens by ELISPOT assay, we found that IL-2 was not produced in response to donor antigens at any time point in mice receiving BMT with our regimen (Table 1), suggesting that IL-2 production is not required for either the deletion of donor-reactive T cells or the establishment of donor-specific tolerance. Additionally, we have not found evidence for the early expansion of donor-specific CD4+ T cells (J.K and M.S., unpublished data), which has been reported to precede AICD in several studies (47). Overall, these data suggest that ‘classical’ AICD is not required for the establishment of peripheral CD4+ T-cell tolerance in mice receiving BMT with anti-CD40L mAb treatment. This result is in contrast to those of other groups, but may be related to the use anti-CD40L specifically in combination with BMT.
To address the relative role that passive cell death plays in the establishment of mixed chimerism and tolerance using BMT and costimulatory blockade, we attempted to produce this state in Bcl-xL tg recipients constitutively expressing Bcl-xL in T cells. Bcl-xL has been shown to protect T cells from apoptosis caused by cytokine withdrawal (37), and we have previously shown that Bcl-xL tg mice have an impairment in the early deletion of donor-reactive CD4+ T cells (17). But in contrast to the results found in Fas-deficient recipients, mixed chimerism was established in only one mouse and donor-specific tolerance could not be achieved in any of the Bcl-xL Tg mice, as measured by skin graft acceptance. This suggests that one of the major mechanisms involved in establishing tolerance using this approach requires a passive death pathway, which has also been implicated in other tolerance-inducing protocols using costimulatory blockade (15).
Various groups using costimulatory blockade as a means of inducing tolerance have implicated a suppressive or regulatory cell population in promoting graft acceptance (14,50,51). One primary difference between the induction of peripheral tolerance with costimulatory blockade and an organ allograft and the use of costimulatory blockade with BMT is that the long-term maintenance of tolerance in mixed chimeras developing with BMT occurs primarily via central deletion of donor-reactive cells (8–10). However, as the peripheral CD4+ T cells are not depleted at the time of BMT, the early mechanisms involved in CD4+ T-cell tolerance with BMT and anti-CD40L may include similar processes to those involved in other peripheral tolerance protocols, including anergy and regulatory/suppressive cells. Donor-specific tolerance is induced rapidly, as evidenced by the permanent acceptance of donor-type skin grafted on Day +1 post-BMT, and by in vitro results showing complete donor specific tolerance in MLR and IL-2 production assays within 8 days of BMT. As deletion of donor-reactive Vβ is incomplete by this time, mechanisms in addition to peripheral deletion must contribute to this initial tolerance.
Interestingly, in the Bcl-xL recipients, several mice had detectable donor chimerism in the blood longer than 13 weeks post-BMT, even though they had already rejected donor skin grafts (Figure 6). One possible explanation for this result is that early post-BMT, a regulatory cell population or some other type of specific nonresponsiveness is established in the peripheral T-cell pool as a result of BMT with anti-CD40L, and this may be sufficient to prevent the rapid rejection of donor BMCs in some Bcl-xL recipients. As the anti-CD40L mAb levels decrease, this nonresponsive state may disappear, and the donor-reactive CD4+ T cells that are deleted in WT recipients as a result of PCD are still present in the Bcl-xL tg recipients, and mediate rejection of donor skin grafts. The relative lack of early deletion in Bcl-xL recipients that have donor chimerism compared with WT-mixed chimeras is consistent with this interpretation (Figure 7). Another possibility is that Bcl-xL mice have a deficiency in the ability to generate regulatory or suppressive cell populations necessary for the long-term maintenance of donor-specific tolerance. However, studies in WT recipients do not implicate such a population in the maintenance of long-term tolerance in recipients of BMT with anti-CD40L, anti-CD8, and 3 Gy TBI (J.K et al., manuscript in preparation).
In these studies, we have further defined the relative roles that T-cell death play in establishing mixed chimerism and specifically elucidated its role in inducing CD4+ T-cell tolerance using anti-CD40L and BMT. In contrast to other reports, we have demonstrated that the traditional AICD pathways involving CsA, Fas, or IL-2 do not seem to play an essential role in either the early induction or the late maintenance of donor-specific tolerance in mixed chimeras generated using this protocol, but that passive cell death appears to be crucial in allowing the long-term acceptance of donor-specific grafts. These results will have implications in the design of novel strategies for tolerance induction in the clinical setting.
Supported in part by NIH Grant ROI HL49915, NIH Training Grant IT32-07529 and a sponsored research agreement with Bio Transplant, Inc. We thank Drs Henry Winn and John Iacomini for critical review of this manuscript, and Ms. Robin Laber for expert assistance in its preparation.