Strategic nonmyeloablative conditioning: CD154:CD40 co-stimulatory blockade at primary BMT promotes engraftment for secondary BMT after engraftment failure

Abstract

There is an increased risk of failure of engraftment following nonmyeloablative conditioning. Sensitization resulting from failed bone marrow transplantation (BMT) remains a major challenge for secondary BMT. Approaches to allow successful retransplantation would have significant benefits for BMT candidates living with chronic diseases. We used a mouse model to investigate the effect of preparative regimens at primary BMT on outcome for secondary BMT. We found that conditioning with TBI or recipient T-cell lymphodepletion at primary BMT did not promote successful secondary BMT. In striking contrast, successful secondary BMT could be achieved in mice conditioned with anti-CD154 co-stimulatory molecule blockade at first BMT. Blockade of CD154 alone or combined with T-cell depletion inhibits generation of the humoral immune response after primary BMT as evidenced by abrogation of production of anti-donor Abs. The humoral barrier is dominant in sensitization resulting from failed BMT, as almost all CFSE-labeled donor cells were killed at 0.5 and 3 hr in sensitized recipients in in vivo cytotoxicity assay, reflecting antibody-mediated cytotoxicity. CD154:CD40 co-stimulatory blockade used at primary BMT promotes allogeneic engraftment in secondary BMT after engraftment failure at first BMT. The prevention of generation of anti-donor Abs at primary BMT is critical for successful secondary BMT.

Introduction

Bone marrow transplantation (BMT) has the potential to provide novel, cell-based therapies for treatment of autoimmune diseases and hemoglobinopathies, and to induce tolerance to solid organ and cellular transplants (1–3). The morbidity associated with ablative conditioning has prevented the widespread application of BMT to these conditions. To reduce the risk of BMT, reduced-intensity conditioning has been pursued for treatment of a number of disorders (4–8). Nonmyeloablative conditioning regimens have the added benefit of allowing autologous reconstitution of the recipient hematopoietic cells if donor engraftment fails. However, a significant challenge that has emerged from these reduced-intensity approaches is the fact that a certain percentage of recipients fail to engraft at primary BMT. Although endogenous hematopoiesis ensues, if sensitization occurs these patients are often subjected to disease-progression as a result. Little information is available evaluating re-transplantation after engraftment failure with nonmyeloablative conditioning. The purpose of the present studies was to study the effects of nonmyeloablative conditioning regimens used during primary BMT on the success of re-transplantation after primary graft failure.

The original studies of nonmyeloablative conditioning were performed using mouse models. The most common immunosuppressive regimens shown to promote engraftment with reduced myelotoxic intensity include: 1) antibodies to target host immune-reactive cells, including antilymphocyte globulin, anti-CD4 anti-CD8, and anti-αβ-TCR, anti-NK1.1; 2) antibodies for costimulatory blockade: anti-CD154, CTLA4 Ig, anti-OX40L; and 3) immunosuppressive drugs including cyclophosphamide, rapamycin, mycophenolate mofetil, fludarabine, and cyclosporine A. The combined use of these treatments has significantly decreased the dose of total body irradiation (TBI) required for engraftment, with the findings successfully translating to larger animals (dogs and pigs) and humans (9–11).

We previously demonstrated in a mouse model that conditioning with 700 cGy of TBI is required to achieve engraftment in MHC-disparate allogeneic recipients (12). The minimum TBI dose can be significantly reduced by targeting recipient T cells with T cell-specific mAbs (13,14), and/or when immunosuppressive drugs (15,16), or agents that block costimulatory molecule interactions (15,17) are added to the conditioning approach. In a recent report, stable engraftment was achieved in an allogeneic mouse model without irradiation using a combination of costimulation blockade plus rapamycin (18). These promising mechanistically-based nonmyeloablative conditioning strategies have been successfully translated to the clinic (4,8,19,20). One limitation to reduced-intensity conditioning is that there is an increased rate of failure of engraftment.

In the present studies, we used a mouse model of BMT to study the impact of the approach for primary conditioning in recipients who failed engraftment after BMT and require re-transplantation. We found that the dominant mechanism for secondary BMT failure is due to sensitization and the generation of an adaptive humoral response. Our data demonstrate that TBI alone at primary BMT does not promote engraftment of secondary BMT in the recipients who failed primary BMT. Recipient T cell lymphodepletion also does not prevent sensitization if engraftment failure results at primary BMT. Co-stimulatory blockade with anti-CD154 alone or in combination with anti-αβ-TCR at the time of primary BMT prevents allosensitization and promotes allogeneic engraftment in secondary BMT after engraftment failure in primary BMT. Conditioning of BMT recipients with anti-CD154 alone or combined with anti-αβ-TCR prevents generation of anti-donor MHC antibody after failed primary BMT. In contrast, significantly higher Ab titers resulted in mice conditioned with anti-αβ-TCR or without mAb treatment at primary BMT. Pre-existing anti-donor antibodies represent the critical mechanism for secondary graft rejection, as >98% of donor cells were eliminated within 30 min in in vivo cytotoxicity assay in the sensitized recipients after first BMT. Our data suggest that the prevention of sensitization, especially generation of anti-donor Ab by agents used in conditioning for first BMT, is critical for successful secondary BMT in nonmyeloablative conditioning. This finding may significantly benefit candidates for BMT with otherwise untreatable chronic disease states.

Materials and Methods

Animals

Male C57BL/6 (B6; H-2^b) and BALB/cJ (BALB/c; H-2^d) mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Animals were housed at the Institute for Cellular Therapeutics under specific pathogen-free conditions, and cared for according to National Institutes of Health guidelines.

Preconditioning and bone marrow transplantation

Primary BMT was performed to mimic BMT with insufficient conditioning (Fig. 1A). Recipient B6 mice were conditioned with 0–700 cGy TBI (γ-cell 40; Nordion, Ontario, Canada). In the mAb-treated groups, B6 mice were pre-treated i.p. with anti-αβ-TCR (H57–597: hamster anti-mouse IgG, generated in our laboratory) and anti-CD154 (MR-1: hamster anti-mouse IgG3, Bioexpress, Lebanon, NH) alone or in combination. Anti-αβ-TCR mAb (30 μl per mouse) was injected on day −3 before BMT and anti-CD154 mAb was given (0.5 mg per mouse) on day 0 and +3 with respect to BMT. Anti-αβ-TCR mAb was unpurified and concentrated by ammonia sulfite precipitation. A dose titration for in vivo depletion of αβ-TCR⁺ cells was previously performed (14). For the secondary BMT, recipients were ablatively conditioned with 950 cGy TBI (Fig. 1B) or nonmyeloablative conditioning comprised of anti-αβ-TCR, anti-CD154, rapamycin (3 mg/kg, day 0, +1, +2, +3, and +4; LC Laboratory, Woburn, MA), and 300 cGy TBI (Fig. 1C). Conditioned mice were transplanted with 15 × 10⁶ untreated donor bone marrow cells (BMC) via lateral tail vein injection at least 4–6 h after TBI as previously described. Briefly, tibias and femurs were harvested from donors. Bone marrow was expelled from the bones with medium 199 (Life Technologies, Grand Island, NY) containing 10 μg/ml gentamicin (Life Technologies) and gently processed into a single-cell suspension as previously described (21).

Figure 1.

A, A nonmyeloablative BMT model was established to provide a threshold for engraftment. Some recipient B6 mice were conditioned with 0–700 cGy TBI 4–6 hr before BMT. Some recipients were preconditioned with mAb via i.p. injection with anti-αβ-TCR and/or anti-CD154. B, For secondary BMT, recipients that failed primary transplants were preconditioned with 950 cGy TBI. Conditioned mice were transplanted with 15 × 10⁶ untreated donor BMC via lateral tail vein injection at least 4–6 h later. Chimerism testing was performed 1 mo after BMT. C, Secondary BMT was also performed after nonmyeloablative conditioning consisting of anti-αβ-TCR, anti-CD154, rapamycin, and 300 cGy TBI.

Chimerism testing

Recipients were characterized for chimerism using flow cytometry to determine the relative percentages of donor-derived peripheral blood leucocytes (PBL) at 4 wk as previously described (21). Peripheral blood was stained with Abs specific for MHC class I Ags of donor (FITC-conjugated anti-H-2K^d, 36-7-5, mouse IgG2a) and recipient (PE-conjugated anti-H-2K^b, AF6–88.5, mouse IgG2a) origin. Multilineage production was assessed by four-color staining for FITC-conjugated, anti-donor-specific Ab and different fluorochrome (PE, PerCP, and APC)-conjugated lineage makers, including T cells (anti-CD4, RM4–5; anti-CD8α, 53-6.7; and anti-αβ-TCR, H57–597), B cells (anti-B220, RA3-6B2), NK cells (anti-DX5, DX5), dendritic cells (DC, anti-CD11c, HL3), and myeloid cells (anti-GR-1, RB6-8C5; and anti-MAC-1, M1/70) (All antibodies were obtained from BD/PharMingen, San Jose, CA). Nonspecific background staining was controlled using isotype control Abs directed against irrelevant Ag conjugated with the same fluorochrome.

Characterization of residual host cell populations after total body irradiation and anti-CD154 treatment

To assess the changes in cell subpopulations in recipient marrow and spleen after TBI and anti-CD154 treatment, the absolute number of T (CD4⁺ and CD8⁺), B (CD19⁺), NK (NK1.1⁺), NK/T (αβ/γδ-TCR⁺/NK1.1⁺), and DC (CD11c⁺) cells in each mouse was determined. Briefly, recipient B6 mice were irradiated with 100, 300, and 600 cGy TBI and bone marrow and spleens were harvested one day after TBI. Anti-CD154 treatment was performed on day 0 and +3, and bone marrow and spleens were harvested two days after the final dose of mAb. Naïve mice served as controls. Single-cell suspensions were incubated with ammonium chloride lysing buffer to eliminate erythrocytes and was washed twice. 1 × 10⁶ cells were stained and analyzed by fluorescence-activated cell sorter (LSR, Becton Dickinson, San Jose, CA). The phenotypic analysis was performed by four-color flow cytometry using anti-CD19, anti-αβ-TCR, anti-γδTCR, anti-CD4, anti-CD8, anti-NK1.1, and anti-CD11c.

Flow cross-match assay (FCXM)

Sera were taken from mice 4 wk after BMT. 0.5 × 10⁶ splenocytes from naive BALB/c mice were incubated with 10 μl of sera for 30 min. The cells were washed and incubated with fluorescein isothiocyanate (FITC)-conjugated polyclonal goat anti-mouse Ig (Immunology Consultants Laboratory, Inc., Newberg, OR), followed by a third incubation with PE-conjugated anti-mouse CD4 plus CD8 (PharMingen, San Diego, CA). Levels of circulating alloantibodies were determined by FACSCalibur or LSR (BD Biosciences, Mountain View, CA), gating on the CD4⁺ and CD8⁺ T-cell fraction, and were reported as mean fluorescence intensity (MFI).

CD8⁺ effector/memory T-cell enumeration

BMT was performed (day 0) from BALB/c to B6 mice with or without anti-CD154 mAb treatment (day 0 and +3). Recipient PB was collected at selected time points up to 30 days after BMT. Cells were stained with anti-CD44-FITC (IM7), antk-CD62L-PE (MEL-14), anti-CD4-PerCP, and anti-CD8-APC (all from BD Pharmingen). Flow cytometry was performed using a FACSCalibur and/or LSR flow cytometry system (BD Biosciences) and analyzed using Cell Quest software (BD Biosciences). Percentages of CD8⁺/CD44^high/CD62L^low/−effector/memory T cells were measured.

Flow cytometric analysis of TCR Vβ families

BALB/c mice express superantigen I-E, resulting in the clonal deletion of superantigen-specific Vβ5.1/2⁺ and Vβ11⁺ T cells. As recipient B6 mice do not express I-E, they do not delete these two Vβ subfamilies. To investigate whether clonal deletion is operational in our model, peripheral blood (8–100 μl) from mixed chimeras and naive control mice was stained with anti-Vβ5.1/2-FITC (MR9-4), Vβ6-FITC (RR4–7), Vβ8.1/2-FITC (MR-5-2), or Vβ11-FITC (RR3–15) vs anti-host H2K^b-PE, anti-CD8-PerCP, and anti-CD4-allophycocyanin (all from BD PharMingen) for 30 min at 4°C. A minimum of 50,000 gated events was collected within the total lymphoid gate. Background staining was determined by FITC-conjugated isotype mAbs. Relative expression indicates the percentage of Vβ-positive cells within the CD8⁺ or CD4⁺ T cell subsets of the host (H2K^b) lymphoid gate in peripheral blood.

In vivo cytotoxicity assay

Single splenocyte suspensions were prepared by gently crushing the spleens and filtering through nylon mesh (30 μm; BD Biosciences). Target and internal control splenocytes (100 × 10⁶/mL in PBS) were then incubated with 4.0 μM or 0.2 μM carboxyfluorescein diacetate succinimidyl ester (CFSE; Molecular Probes, Leiden, The Netherlands), respectively, at room temperature for 10 min. Equal volume of fetal bovine serum (FBS; Gibco BRL) was added to quench the reaction. After washing, cells were mixed in a 1:1 ratio and resuspended in PBS, and 20 × 10⁶ cells from each were injected intravenously. Peripheral blood was collected from individual mice at selected time points after cell infusion: 0.5hr, 1hr, 3hr, day 1, day 2, and day 3. After lysis of red blood cells, peripheral-blood lymphocytes were analyzed for CFSE expression by fluorescence-activated cell sorting (FACS). The percentage of killing was determined by calculating the ratio between target and internal control cells.

Statistical analysis

Data are presented as the average ± SD. The two-tailed t test (two-sample, assuming unequal variances) was used to evaluate statistical differences. The difference between groups was considered significant at P < 0.05.

Results

Conditioning that prevents generation of the anti-donor humoral response enhances likelihood of success in second transplants

For primary transplantation, recipient B6 mice were conditioned with 0–700 cGy TBI and transplanted with BALB/c bone marrow cells. Engraftment occurred in 53.3% and 20% of mice that received 700 and 600 cGy TBI, respectively (Fig. 2A). Engraftment failed in all mice conditioned with 0, 100, and 300 cGy TBI. All mice survived. Those who failed to engraft exhibited endogenous hematopoiesis. FCXM assay was performed in sera collected 4 wk after BMT to detect donor-specific antibodies (Fig. 2B). Donor-specific antibody levels were not significantly increased (P > 0.05) in groups conditioned with 600 (MFI: 5.4 ± 3.8) or 700 cGy TBI (5.2 ± 0.8) compared with serum from naïve controls (4.6 ± 0.7; P > 0.05). There was also no difference in serum MFI between mice that engrafted and those that did not in the groups conditioned with 600 or 700 cGy TBI (data not shown). Significantly higher antibody titers were detected in recipients conditioned with 100 (218.8 ± 69.8; P = 0.00005) or 300 cGy TBI (149.9 ± 108.2; P = 0.006) compared with sera from naïve controls and their values were similar to the control B6 mice (197.6 ± 102.1) who received BMT but no TBI.

Figure 2. Total body irradiation alone at primary BMT did not promote engraftment of secondary BMT.

Recipient B6 mice were conditioned with 0, 100, 300, 600, and 700 cGy TBI 4–6 hr prior to infusion of BALB/c 15 × 10⁶ bone marrow cells. A, Recipients were analyzed for engraftment using flow cytometric analysis 4 wk after BMT by determining the relative percentages of donor-derived PBL. Results are summarized from three separate experiments. B, Anti-donor Ab was measured by FCXM assay. Sera were collected 4 wk post primary BMT from same mice in A. Sera collected from naïve mice served as a control. Levels of circulating alloantibodies were determined by FACSCalibur or LSR (Becton Dickinson), gating on the CD4⁺ and CD8⁺ T-cell fraction, and are reported as MFI. To assess the changes in cell populations in recipient marrow and spleen after conditioning, another group of recipient B6 mice was prepared and irradiated with 100, 300, and 600 cGy TBI, and bone marrow (from femurs and tibias) and spleens were harvested one day after TBI. Naïve mice served as controls, single-cell suspension was prepared and 1 × 10⁶ cells were stained and analyzed. Cells were first incubated with normal mouse serum to eliminate nonspecific binding. The absolute number of T (CD4⁺, and CD8⁺), B (CD19⁺), NK (NK1.1⁺), NK/T (αβ/γδTCR⁺/NK1.1⁺) and dendritic (CD11c⁺) cells in each mouse were enumerated in BM (C) and spleen (D). Results are summarized from two experiments. E, Secondary BMT was performed 5–7 wk after primary BMT in mice from A and B. Naïve and sensitized B6 mice served as controls. Recipients were conditioned with 950 cGy TBI and transplanted with 15 × 10⁶ untreated BALB/c donor BMC via lateral tail vein injection 4–6 h after irradiation. Animals were analyzed for engraftment 4 wk after BMT.

To determine the host cell populations targeted by TBI and to explore the correlation between sensitization with the cell population changes after TBI, B6 recipients were conditioned with selected doses of TBI (100, 300, 600 cGy). Splenocytes and BMC were analyzed by flow cytometry 24 h thereafter. The absolute number of total splenocytes, BMC, and subpopulations declined steadily with increasing TBI dose (Fig. 2C and 2D). With 600 cGy TBI, the most dramatic reduction occurred in B cells in bone marrow, decreasing 77.4-fold (P < 0.01) from 8.9 ± 2.6 to 0.12 ± 0.05 million. A decrease was also observed in CD4 T cells (6.5-fold; P < 0.005), CD8 T cells (6.6-fold; P < 0.005), NK cells (3.3-fold; P < 0.005), NK-T cells (1.6-fold; P < 0.05), and DC (27-fold; P < 0.005). These populations decreased in a similar fashion in recipient spleen.

Secondary BMT was performed 5–7 wk after the primary BMT in the mice that failed to engraft to evaluate the effect of TBI at primary BMT on secondary transplant. B6 recipients were conditioned with 950 cGy TBI and transplanted with 15 × 10⁶ BALB/c bone marrow cells (Fig. 2E). Naïve B6 mice served as controls. Engraftment did not occur in mice treated with 0, 100, or 300 cGy TBI alone at the first BMT, as expected, due to sensitization from the first BMC infusion. Interestingly, significantly less engraftment occurred than was expected in recipients that had received 600 or 700 cGy TBI conditioning at first BMT, even though no anti-donor antibody was detected in these animals. Only 20.0% and 12.5% of mice conditioned with 700 or 600 cGy TBI engrafted, respectively.

Co-stimulatory blockade of CD154:CD40 alone or combined with T-cell depletion at primary BMT prevents allosensitization after the failure at first BMT

We next examined whether blockade of CD154/CD40 interactions with or without T-cell lymphodepletion enhanced prevention of allosensitization in primary BMT candidates who failed to engraft. Primary BMT was performed to mimic the insufficient conditioning in BMT. B6 mice were conditioned with: 1) anti-αβ-TCR alone; 2) anti-CD154 mAb alone; or 3) a combination of both mAbs and transplanted with 15 × 10⁶ BALB/c BMC. The effect of anti-CD154 treatment on cell populations was analyzed 2 days after the last dose of anti-CD154. There were no significant differences detected in B, T, NK, NK-T, and dendritic cell subpopulations in the spleens (Fig. 3A) of anti-CD154-treated vs. naïve mice. The levels of anti-donor antibody were measured 4 wk after primary BMT (Fig. 3B). As expected, none of the recipients treated with anti-αβ-TCR alone, anti-CD154 alone, or both antibodies in combination engrafted at primary BMT (data not shown). Antibody titers in mice treated with anti-CD154 mAb alone (MFI: 5.5 ± 1.6; P > 0.05) or in combination with anti-αβ-TCR (5.3 ± 2.7; P > 0.05) were not significantly different from naïve mice (4.6 ± 0.7). In contrast, mice treated with anti-αβ-TCR mAb alone produced anti-donor antibody at significantly higher levels (47.8 ± 59.7; P = 0.03) compared with naïve mice, but significantly lower levels (P = 0.0002) than the positive controls that received BMC only without conditioning (148.9 ± 72.3).

Figure 3. The effect of anti-CD154 mAb and/or anti-αβ-TCR used at first BMT on secondary engraftment.

A, Anti-CD154 treatment (n = 4) was performed on day 0 and +3, and spleens were harvested two days after the last dose. The absolute number of T (CD4⁺, and CD8⁺), B (CD19⁺), NK (NK1.1⁺), and NK/T (αβ/γδTCR⁺/NK1.1⁺) cells were enumerated. Untreated mice served as controls (n = 4). B, Recipient B6 mice were pretreated with anti-αβ-TCR mAb (0.1 mg/day; day −3) and/or anti-CD154 mAb (0.5 mg/day; day 0 and +3) at primary BMT (day 0) from BALB/c donors. Anti-donor antibody production was analyzed with FCXM assay from sera collected 4 wk after primary BMT. Antibody titers are reported as MFI. C and D, To evaluate the influence of the conditioning used at primary BMT on secondary BMT, secondary BMT was performed 5–7 wk after primary BMT in B6 recipients who were failed engraftment at primary BMT. In secondary BMT, recipients were conditioned myeloablatively (950 cGy TBI) or nonmyeloablatively (anti-αβ-TCR, anti-CD154, rapamycin, and 300 cGy TBI). 15 × 10⁶ BALB/c BMC were transplanted to each recipient. Animals were analyzed for engraftment by flow cytometric analysis. The frequency of engraftment 1 mo after secondary BMT (C) and the level of chimerism in animals that engrafted (percentage of donor cells in PBL) at 1 mo (D) were shown. The results are the summary of three experiments.

Administration of anti-CD154 mAb with conditioning at first BMT promotes secondary allogeneic engraftment in mice who experience failure at primary transplant

To evaluate the influence of conditioning at primary BMT on the success of secondary BMT, secondary BMT was performed in B6 recipients (used in Fig. 3B) 5–7 wk after primary BMT. Recipients were conditioned with 950 cGy TBI and transplanted with 15 × 10⁶ BALB/c BMC (Fig. 3C). Engraftment occurred in 22.2% of mice (n = 9) treated with anti-αβ-TCR mAb alone at first BMT and these two engrafted recipients had low and close to normal level of anti-donor Ab (MFI: 6.5 and 7.0). All mice with no treatment at first BMT failed engraftment after secondary BMT. In contrast, 100% of mice treated with anti-αβ-TCR plus anti-CD154 engrafted, and 81.8% of mice treated with anti-CD154 alone engrafted. Engraftment was stable up to 6 mo follow-up (data not shown) and multilineage (Table 1, Group A). TCR Vβ repertoire could not be assessed in these chimeras because the levels of donor chimerism approached 100%.

Table 1.

Multilineage Analysis^a

	nb	% donor	% in donor (H2Kd+) lymphoid gate c
			αβ-TCR	CD8	CD4	B220	DX5	Mac-1	Gr-1	CD11c
A	11	99.7±0.5	20.0±5.5	2.5±0.9	17.6±5.7	73.7±5.2	6.3±5.0	2.2±0.5	1.0±0.3	1.1±0.1
B	10	15.6±8.9	13.9±6.2	1.8±1.1	11.9±5.8	80.2±5.9	0.7±0.2	2.7±0.8	N/A	0.6±0.1

A: myeloablative conditioning at secondary BMT

B: nonmyeloablative conditioning at secondary BMT

^aMultilineage typing was performed in mice between 2 and 3 mo after 2nd BMT.

^bMice were conditioned with anti-CD154 alone or combined with anti-αβ-TCR at the primary BMT

^cMean ± SD percentage of donor cells in donor lymphoid gate.

We next asked whether secondary BMT could be successfully achieved with nonmyeloablative conditioning in those recipients who failed engraftment. The sensitization of these mice was prevented by treatment with anti-CD154 alone (MFI: 5.0 ± 1.1) or in combination with anti-αβ-TCR (MFI: 4.9 ± 0.4) after primary BMT. In secondary BMT, recipients were conditioned with anti-αβ-TCR, anti-CD154, rapamycin, and 300 cGy TBI, and transplanted with 15 × 10⁶ BALB/c BMC (Fig. 1C). Engraftment occurred in all mice treated with anti-CD154 alone or anti-αβ-TCR plus anti-CD154 at primary BMT (Fig. 3C). The levels of donor chimerism were similar to naïve recipients who only received the nonmyeloablative conditioning (Fig. 3D). Twenty percent of mice (n = 5) treated with anti-αβ-TCR mAb alone at first BMT and none without treatment at first BMT engrafted after secondary BMT. This engrafted recipient with anti-αβ-TCR mAb at first BMT had relative low level of anti-donor Ab (MFI: 6.1) compared with other four without engraftment (MFI: 13.1, 15.1, 44.8, and 26.5). Moreover, the engraftment was stable for up to 6 mo follow-up (data not shown) and multilineage (Table 1, Group B). Moreover, all chimeric mice demonstrated deletion of Vβ5½ and Vβ11 in host-derived CD4 and CD8 T cells (Fig. 4A), a critical marker for the mechanism of central tolerance to donor alloantigens.

Figure 4. A, Relative TCR Vβ expression in mixed chimeras with nonmyeloablative conditioning.

Expression of Vβ5.1/2, Vβ6, Vβ8.1/2, and Vβ11 on PBL from unmanipulated hosts (B6), unmanipulated donors (BALB/c), mixed chimeras with nonmyeloablative conditioning atcsecondary BMT in Figure 3C and 3D prepared preconditioning with anti-CD154 alone, anti-αβ-TCR alone or both mAb at first BMT was measured by FACS analysis 2–3 mo after secondary BMT. Relative expression represents the percentage of Vβ-positive cells within the CD8 or CD4 T-cell subsets of the host (H2K^b) lymphocytes in peripheral blood. Vβ expression in either chimeric group was compared with that in B6 mice using one-tailed t test (two-sample, assuming unequal variances). Significant p values are indicated above the respective data bars (Ψ, p < 0.01; ж, p < 0.001). The results are the summary of two experiments. B, Effect of anti-CD154 mAb and/or anti-αβ-TCR used at first BMT on antibody mediated cytotoxicity in vivo. To determine the effect of donor-specific Ab on rejection of donor cells, new group of B6 mice was transplanted and treated with anti-CD154 and/or anti-αβ-TCR. In vivo cytotoxicity assays were performed 5–7 wk after first BMT. 20 × 10⁶ CFSE labeled each naïve donor BALB/c splenocytes (as target cells) and naïve recipient B6 splenocytes (as internal controls) were infused into each experimental mouse. Peripheral blood was collected from individual mice at selected time points after cell infusion: 0.5hr, 1hr, 3hr, day 1, day 2, and day 3. After lysis of red blood cells, peripheral-blood lymphocytes were analyzed for CFSE expression by a FACSCalibur or LSR scanner and CellQuest software (BD Biosciences). The results are the summary of four experiments.

Anti-CD154 mAb treatment prevents the generation of antibody-mediated cytotoxicity in vivo

The success of engraftment in secondary BMT correlated inversely with generation of anti-donor Ab detected in the recipient serum post-transplantation. To determine the effect of donor-specific Ab on rejection of donor cells, in vivo cytotoxicity assays were performed in mice treated with anti-CD154 and/or anti-αβ-TCR at the first BMT to define the mechanism for our observation. 20 × 10⁶ CFSE labeled naïve donor BALB/c splenocytes (as target cells) and naïve recipient B6 splenocytes (as internal controls) were infused into mice previously transplanted 5–7 wk after primary BMT (Fig. 4B). Recipients of only BMC infusion at first BMT and naïve B6 mice served as controls. B6 mice initially infused with BALB/c BMC only rapidly eliminated almost all CFSE-labeled BALB/c splenocytes within 0.5 h (98.1 ± 1.41%), and the anti-donor Ab titer was 164.7 ± 82 in these as per the in vivo cytotoxicity assay. The dramatically increased cytotoxicity strongly suggests that antibody-mediated killing represents the predominant barrier for alloreactivity in sensitized recipients. Mice previously conditioned with anti-αβ-TCR eliminated 78.5±14.0% and 92.5± 10.0% CFSE labeled BALB/c cells at 1 and 3 hr, respectively. In contrast, B6 mice treated with anti-CD154 or both mAb displayed in vivo cytotoxicity similar to the naïve mice, confirming that anti-CD154 mAb treatment prevents the generation of antibody mediated cytotoxicity by functionally impairing sensitization in vivo.

CD154 blockade inhibits the generation of CD44^high/CD62L^low/− effector/memory CD8⁺ T cells

CD8⁺ effector/memory T cells are generated in adaptive immune responses and identified as CD44^high/CD62L^low/− (22–24). To investigate the role of CD154:CD40 interactions in generating memory T cells, B6 mice were treated with anti-CD154 mAb at primary BMT and CD44^high/CD62L^low/− effector/memory T cells were enumerated at 11 to 13 days after BMT (Fig. 5A). CD8⁺ effector/memory T cells were detected in non-mAb treated controls and the percentage of CD8⁺ effector/memory T cells in CD8⁺ cell gate was significantly higher than in naïve mice (11.0 ± 1.6% vs. 2.7± 1.5%, respectively; P < 0.000001). Blockade of CD154 completely inhibited the generation of CD8⁺ effector/memory T cells and the percentage of CD8⁺ effector/memory T cells in CD8⁺ cell gate (3.1± 2.2%) was similar as naïve controls (P = 0.59).

Figure 5. Effect of anti-CD154 mAb and/or anti-αβ-TCR used at first BMT on effector/memory T-cell development.

A, B6 mice treated with anti-CD154 mAb (day 0, +3) were transplanted with BALB/c BMC (day 0). Naïve and non-mAb treated B6 mice served as controls. Recipient peripheral blood was collected 11 to 13 days thereafter and stained with anti-CD44-FITC, anti-CD62L-PE, anti-CD4-PerCP, and anti-CD8-APC mAbs. The CD44^high/CD62L^low/−effector/memory T cells were enumerated (n = 8–14 mice per group). B, The expression of effector/memory T-cell phenotype in CD8⁺ T cells in the context of homeostatic proliferation after T-cell depletion and alloresponse. B6 mice treated with anti-αβ-TCR mAb (day −3) and anti-anti-CD154 mAb (day 0, +3) were transplanted with BALB/c BMC (day 0). Naïve and anti-αβ-TCR mAb treated B6 mice served as controls. Recipient peripheral blood was collected on day13 or 30 and the CD44^high/CD62L^low/− effector/memory T cells were enumerated (n = 5–8 mice per group).

CD154 blockade impairs generation of CD8⁺ effector/memory T cells induced by both alloantigen exposure and lymphopenia

T-cell depletion leads to lymphopenia-induced homeostatic proliferation of T cells, driving naive T cells to differentiate directly into memory T cells (25,26). As shown in Figure 5B, more CD8⁺ effector/memory T cells (22.7 ± 10.8%, P = 0.02) were detected at day +11 to +13 in mice treated with anti-αβ-TCR (day −3) and BMC infusion (day 0) compared with mice that received BMC only (11.0 ± 1.6%, Fig. 5A). The level of CD8⁺ effector/memory T cells remained higher even 30 days after BMT (7.2 ± 1.0%) in mice treated with anti-αβ-TCR. The expression of T cells of CD8⁺ effector/memory phenotype in the context of homeostatic proliferation and alloresponse was decreased by CD154 blockade. The levels of CD8⁺ effector/memory T cells were significantly lower in the mice treated with anti-αβ-TCR plus anti-CD154 at day +11 to +13 (11.9 ± 8.1, p = 0.04 ) or day +30 (2.6 ± 1.0, p =0.005; Fig. 5B) compared with CD8⁺ effector/memory T cells in mice treated with anti-αβ-TCR alone.

Discussion

Bone marrow transplantation has the potential to treat a number of chronic conditions, including autoimmune disease, hemoglobinopathies, enzyme deficiency states, and to induce tolerance to organ transplants (3,27–30). In order to reduce the toxicity of conditioning for BMT reduced-intensity or nonmyeloablative approaches are being tested in the clinic with the goal to minimize risk and maximize outcomes. As a result, rejection of donor stem cells has emerged as a significant complication of BMT following nonmyeloablative conditioning (7,8). Strategies to successfully re-transplant these individuals are of high priority if optimal therapeutic potential is to be realized in these chronic disorders. This has prompted us to study the influence of conditioning regimens used at primary BMT on the outcomes following secondary BMT. The results presented here demonstrate that secondary BMT can be successfully performed in recipients whose first BMT has failed, and this is significantly influenced by the conditioning used at primary BMT.

We first evaluated the effect of TBI conditioning alone on outcome in secondary BMT. In our current study, 600 TBI significantly decreased the major cell populations of B, T, NK, NKT and DC cells in B6 recipients within 1 day after administration. The proportion of these cell populations on the various doses of TBI directly correlated with sensitization. Coincident with this observation, the production of anti-donor Ab inversely correlated with TBI doses from 0 to 700 cGy used at BM cell infusion. We hypothesize that when host B cells and T cells, especially CD4⁺ cells, decreased to a certain level with a given TBI dose (≥ 600 cGy) there are not enough viable B- and helper T cells to initiate the adaptive humoral immune response. The residual radiation-resistant NK, NK-T, and even cytotoxic T cells after 600 cGy TBI are sufficient to induce cellular immunoresponse evidenced by the BMC rejection in the majority of recipients. When irradiation doses increased to 700 cGy, more cell populations responsible for cellular immune responses are eliminated. Thus, both cellular and humoral immune responses are inhibited and allogeneic engraftment occurs in a higher proportion of recipients. When lower doses of TBI (≤ 300 cGy) are used, they only partially decreased the B and T-cell populations, and there were still significant cells remaining to initiate cellular and humoral immune responses, as evidenced by the generation of donor-specific Ab and BMC rejection. As expected, engraftment did not occur in recipients conditioned with 100 or 300 cGy TBI at primary BMT and 950 cGy at following secondary BMT due to the generation of Ab against donor after primary BMC infusion. However, we were surprised to find superior engraftment was not achieved in recipients conditioned with 600 and 700 cGy TBI at primary BMT who underwent ablative secondary BMT (Fig. 2E), even though a humoral response was not generated the first BMT. In considering the relatively high combined TBI dose used at primary and secondary BMT with a short interval (5–7 wk), we hypothesize that the failure of engraftment observed in these recipients might be due to the accumulation of toxicity from irradiation.

It has been well known that T cell depletion combined with other conditioning strategies enhances allogeneic BM engraftment (34,35). T cell depletion alone without TBI or other myelotoxic agents is not sufficient to achieve allogeneic BM engraftment (13,36,37). Our data further demonstrate that the remaining T cells after αβ-TCR T-cell depletion are still be able to initiate humoral and cellular immune responses as evidenced by the generation of Ab against donor MHC antigens and the rejection of BMC at primary BMT (Fig. 2B). In this light, it is not unexpected that the mice treated with anti-αβ-TCR mAb at first BMT failed to engraft at secondary BMT. These results also confirm a dominant role for the humoral immune barrier in rejection of transplanted BMC in sensitized recipient.

Co-stimulatory blockade of CD154:CD40 interactions induces acceptance of murine skin, heart and islet grafts (13,39–41) when combined with T-cell depletion in vivo. CD154 (CD40 ligand) is rapidly up-regulated on recently activated T cells (42,43). CD40, the receptor for CD154, is constitutively expressed on APC, including B cells (44). Approaches that target the CD40–CD154 costimulation components of T cell activation have shown promise in development of tolerance-based transplantation protocols in both rodent and nonhuman primate models (45) by induction of apoptosis or deletion of Ag-activated T cells (46). Anti-CD154 is a non-depleting mAb, as no significant changes in T cell populations are detected after anti-CD154 treatment (Fig. 3A). In contrast with these other grafts, the present studies show that anti-CD154 alone or combined with T cell depletion (without irradiation or cytoreduction agents) did not induce tolerance to allogeneic BMC without irradiation at primary BMT. The disparity in outcomes between solid organ and BMT may be related to the systemic nature of the infused BMC, which circulate in the blood and come in contact with recipient immune cells. Moreover, the donor hematopoietic stem cells need to overcome the recipient immune system for successful engraftment to occur. Therefore, bone marrow cells may face stronger immune barriers than solid organ grafts and require more conditioning for alloengraftment. Our data have confirmed a previous report that the interaction between CD154 and CD40 is primarily for B cell development and function, and contributes only a modest degree in adaptive T cell immune responses (47,48). Our findings indicate that with proper management, patients who have failed BMC grafts may still be capable of accepting secondary BMT.

Blockade of the CD40 costimulatory pathway is known to abort the generation of antigen-specific Ab (49–51). Another possible mechanism for anti-CD154 mAb treatment leading to the inhibition of humoral immune responses is an ability to inhibit T cell activation by affecting the interaction between antigen presenting cells and T cells. As shown in Fig 5 A, CD154 blockade totally inhibited the generation of effector/memory CD8⁺ T cells after allogeneic BMC infusion. This result confirms previous finding in response to alloantigens after skin graft (Xu, Blood 2008). The relevant new finding is that CD154 blockade impaired generation of CD8⁺ effector/memory T cells induced by both alloantigen exposure by BMC and lymphopenia by T cell depletion (Fig. 5A and 5B). T cell depletion has been recently reported to lead to lymphopenia-induced homeostatic proliferation of T cells, driving naive T cells to differentiate directly into memory T cells sharing not only phenotype but also function (Wu, Nat Med. 2004 Jan 10(1): 87; Neujahr, et al., J. Immunol 2006,176:4632). Our data showed that CD154 blockade was enough to block the anti-donor Ag generation in the context of both alloantigen lymphopenia.

Sensitization to MHC antigens is among the most critical challenges to clinical transplantation (33,52,53). Sensitization significantly increases the risk of rejection of solid organ and bone marrow cells. The dominance of the humoral immune barrier in sensitized recipients is reflected by the fact that allogeneic donor marrow engraftment was abrogated in naive mice with the passive transfer of as little as 25 μL serum from sensitized recipients (33). This observation has been further demonstrated using green fluorescent protein (GFP)-labeled BMC and bioluminescence imaging showing that the rejection mediated by preformed antibody occurs within 3 h of transplantation (53). Our current data extends these observations and point to the dominance of humoral sensitization as a barrier to successful transplantation. The BMC-primed recipients rapidly eliminated a moderate donor splenocyte dose in less than 30 min after infusion in an in vivo cytotoxicity assay (Fig. 4B). Therefore, it is almost impossible to achieve allogeneic bone marrow engraftment in sensitized recipients (33). Fortunately, sensitization can be prevented with two treatments of anti-CD154 mAb given at the same time as exposure to alloantigens from BMC. Even when engraftment failed in anti-CD154 mAb-treated recipients, no anti-donor MHC Ab were detected. In addition, there was a marked abrogation of rapid cytotoxicity of donor-specific cells.. Strategies to simultaneously prevent sensitization in nonmyeloablative conditioning regimens at the time of transplant would have the potential to allow a successful secondary transplant to be performed when in the event of failed engraftment. Our findings suggest that the agents used at primary BMT have a critical influence on the outcome of second BMT and an approach should be selected to prevent the generation of donor-specific Ab. Co-stimulatory blockade of CD154:CD40 interactions with anti-CD154 at the time of primary BMT is one approach that serves very well for this purpose. These observations have confirmed our recent work that allosensitization costimulatory blockade of CD154 results in prevention of allogeneic sensitization with skin graft (Xu, et al., Blood 2008). The CD154 blockade can inhibit allogeneic humoral immune response not only to high antigenic skin tissue but also to whole bone marrow cells containing heterogeneous populations. The BMCs have been used in current study to establish a more clinically relevant mouse model to study secondary BMT after the failure in first BMT. With anti-CD154 mAb preconditioning alone or combined anti-CD154 and anti-αβ-TCR mAbs at the time of first failed BMT, allogeneic engraftment was readily achieved with ablative conditioning at subsequent secondary BMT, a result that cannot occur without the anti-CD154 mAb treatment. We then explored whether nonmyeloablative conditioned secondary transplant recipients would also engraft in pre-sensitized recipients. We found that with anti-αβ-TCR, anti-CD154, rapamycin, and as low as 300 cGy TBI at secondary BMT, engraftment occurred in all mice treated with anti-CD154 alone or anti-αβ-TCR plus anti-CD154 at primary BMT. The mixed chimerism observed in these recipients had similar features in durability of donor chimerism and TCR Vβ clonal deletion, suggesting a deletional mechanism.

In conclusion, the results of our study demonstrate that CD154:CD40 co-stimulatory blockade at the time of primary BMT promotes allogeneic engraftment in secondary BMT after engraftment failure. The mechanism by which anti-CD154 enhances the engraftment of secondary BMT is to inhibit B cell activation to generate alloantibody. The prevention of anti-donor Ab generation by agents used at first BMT is critical for successful secondary BMT. These findings could have a significant impact in the clinic on management of BMT in pre-sensitized recipients exposed to allogeneic MHC via a previous, failed transplant.