Treg-therapy allows mixed chimerism and transplantation tolerance without cytoreductive conditioning

Abstract

Establishment of mixed chimerism through transplantation of allogeneic donor bone marrow (BM) into sufficiently conditioned recipients is an effective experimental approach for the induction of transplantation tolerance. Clinical translation, however, is impeded by the lack of feasible protocols devoid of cytoreductive conditioning (i.e. irradiation and cytotoxic drugs/mAbs). The therapeutic application of regulatory T cells (Tregs) prolongs allograft survival in experimental models, but appears insufficient to induce robust tolerance on its own. We thus investigated whether mixed chimerism and tolerance could be realized without the need for cytoreductive treatment by combining Treg therapy with BM transplantation (BMT). Polyclonal recipient Tregs were cotransplanted with a moderate dose of fully mismatched allogeneic donor BM into recipients conditioned solely with short-course costimulation blockade and rapamycin. This combination treatment led to long-term multilineage chimerism and donor-specific skin graft tolerance. Chimeras also developed humoral and in vitro tolerance. Both deletional and nondeletional mechanisms contributed to maintenance of tolerance. All tested populations of polyclonal Tregs (FoxP3-transduced Tregs, natural Tregs and TGF-beta induced Tregs) were effective in this setting. Thus, Treg therapy achieves mixed chimerism and tolerance without cytoreductive recipient treatment, thereby eliminating a major toxic element impeding clinical translation of this approach.

Figure 1

Phenotypical and functional characterization of FoxP3-Tregs, nTregs and iTregs. B6 CD4⁺ cells were transduced with FoxP3/GFP virus and sorted for GFP expression to generate FoxP3-Tregs (A) or in vitro cultured with TGFβ to yield iTregs (C). B6 CD4⁺CD25⁺ nTregs (B) were isolated from spleen and lymph nodes and activated in vitro. Prior to in vivo transfer, cells were phenotypically characterized by flow cytometry. All three populations expressed high levels of FoxP3, CD25 and CD62L. Typical histograms (gated on CD4⁺ cells) for intracellular expression of FoxP3 and surface expression of CD25 and CD62L are shown. (D) FoxP3-Tregs suppressed proliferation of B6 splenocytes in response to allogeneic stimulation (Balb/c splenocytes) (as do freshly sorted control CD4⁺CD25⁺ Tregs), whereas coculture with CD4⁺CD25⁻ sorted cells resulted in enhanced proliferation. SI (stimulation index) was calculated by dividing the mean cpm from responses against syngeneic (B6) or allogeneic (Balb/c) stimulator cells by mean background cpm.

Figure 2

Treg treatment together with rapamycin induces mixed chimerism without cytoreductive conditioning. Groups of B6 mice were grafted with 2 × 10⁷ Balb/c BM cells under the cover of costimulation blockade (anti-CD154, CTLA4Ig) and rapamycin and were additionally treated (A) with (, 4 × 10⁶, n = 7) or without (◊, n = 6) recipient-derived FoxP3-Tregs, (B) with (, 3 × 10⁶, n = 4) or without (◊, n = 7) recipient-derived nTregs or (C) with (, 5 × 10⁶, n = 6) or without (◊, n = 7) recipient-derived iTregs, respectively. Only recipients treated with any of the Treg populations developed chimerism. Donor (H-2D^d+) chimerism among leukocytes of the myeloid (Mac1⁺) lineage was assessed by flow cytometry of peripheral blood at multiple time points and is shown as mean percent (error bars indicate standard deviation; nTreg and iTreg groups were done in the same experiment, therefore the control group is shown twice in B + C). Repeat experiments performed with nTregs and iTregs showed similar results. **p < 0.005, *p < 0.05 with versus without Tregs (two-sided Student's t-test). (D) B6 mice received 2 × 10⁷ Balb/c BM cells under the cover of costimulation blockade and 5 × 10⁶ iTregs with or without rapamycin. Recipients treated without rapamycin failed to develop chimerism (• iTregs with rapamycin n = 8; iTregs without rapamycin, n = 8; p = 0.0002, Fisher's exact test). Donor (H-2D^d+) chimerism among leukocytes of the myeloid (Mac1⁺) lineage is shown 2 weeks post-BMT as scatter plot (mean and standard deviation are indicated).

Figure 3

Treg treatment leads to multilineage chimerism and stem cell engraftment. Chimerism in Treg-treated mice was of multilineage nature as shown by the presence of donor cells among the T-cell (CD4, CD8), B-cell (B220) and myeloid (Mac1) lineages. Two-color flow cytometry plots are shown from representative BMT recipients ∼3 months post-BMT treated with (A) FoxP3-Tregs, (B) nTregs and (C) iTregs. (D) BM was recovered from iTreg-treated chimeras (8 weeks post-BMT; n = 3) and transplanted into myoablated secondary B6 recipients. Donor chimerism was detectable in the T-cell (CD4, CD8), B-cell (B220) and myeloid (Mac1) lineages 12 weeks postsecondary BMT, indicating that donor hematopoietic stem cells had engrafted in primary recipients. Flow cytometry plots are shown from one secondary recipient. Numbers indicate the net percentage of donor chimerism in the depicted lineage (for calculation algorithm, please see section ‘Materials and Methods’).

Figure 4

Chimeras induced through Treg treatment develop donor-specific skin graft tolerance and hyporesponsiveness in vitro. Donor-specific tolerance was assessed by grafting of full thickness donor and third party (C3H) tail skin 4–6 weeks post-BMT. Grafts were considered to be rejected when less than 10% remained viable. Donor skin graft survival was significantly prolonged in BMT recipients treated with (A) FoxP3-Tregs (▪, n = 7; p = 0.0002) compared to BMT recipients without Tregs (♦, n = 6); and in recipients treated with (B) nTregs (▪, n = 4; p = 0.0360) and (C) iTregs (▪, n = 6; p = 0.0043) compared to the controls without Tregs (♦, n = 7, same control group for B + C). Third-party grafts were promptly rejected in all groups (Treg-treated •, control ▵). Survival was calculated according to the Kaplan–Meier product limit method and compared between groups using the log-rank test. Repeat experiments performed with nTregs and iTregs showed similar results. (D,E) Mixed lymphocyte reaction results from selected BMT recipients were obtained early (1 week post-BMT) and at the end of follow-up (22 weeks post-BMT). (D) Early MLRs (1 week post-BMT) demonstrate hyporesponsiveness toward the donor in Treg-treated recipients (n = 2) in comparison to controls without Tregs (p = 0.0120, SI antidonor compared to controls). (E) Long-term chimeras of Treg-treated mice (nTreg, n = 2; iTreg, n = 3) showed specific hyporesponsiveness toward the donor in vitro (p = 0.0446 for nTregs; p = 0.0063 for iTregs, SI antidonor compared to naïve B6 mice). SIs were calculated by dividing the mean cpm from responses against recipient (black column; B6), donor (white column; Balb/c), or third-party (gray column; C3H) stimulator cells by mean background cpm (i.e. cpm with no stimulator population). Error bars indicate standard deviation.

Figure 5

Donor skin grafts of Treg-treated chimeras show high frequencies of mast cells and Tregs. (A) Histopathology of donor skin grafts from Treg-treated chimeras revealed high frequencies of mast cells (Giemsa) and FoxP3 positive cells (immunohistochemistry) (HE, hematoxylin and eosin stain, magnification 100×; FoxP3, immunohistochemistry with specific FoxP3 antibody, magnification 400×; Giemsa, Giemsa staining, magnification 200×; representative graft shown 5 months post-BMT). (B) Genomic DNA isolated from skin grafts of FoxP3-Treg treated mice (n = 3) was subjected to PCR analysis specific for GFP. Grafts lacked detectable GFP expression, indicating that graft-infiltrating FoxP3 Tregs did not originate from the therapeutically administered Tregs. FoxP3-transduced NIH3T3 cells and FoxP3 vector were used as positive controls. DNA levels of GFP (upper panel) and β-actin (lower panel) are shown.

Figure 6

Chimeras induced through Treg treatment demonstrate partial central and peripheral deletion of donor-reactive T cells late after BMT. (A) Percentages of Vβ11⁺ and Vβ5⁺ (but not Vβ8⁺) CD4⁺ splenocytes (SPL) were significantly reduced in Treg-treated mice (red bars, n = 7) in comparison to naïve B6 mice (gray bars, n = 5; p = 0.0003 for Vβ11⁺, p = 0.0309 for Vβ5⁺) and in comparison to BMT recipients without Tregs (black bars, n = 6; p = 0.0026 for Vβ11⁺, p = 0.0128 for Vβ5⁺; 25 weeks after BMT). (B) CD8⁺ splenocytes (SPL) were also significantly deleted in Treg-treated chimeras in comparison to naïve B6 controls, indirectly indicating central deletion (as mature superantigen-reactive CD8 cells are not deleted extrathymically) (p = 0.0002 for Vβ11⁺, p = 0.0141 for Vβ5⁺). (C) Percentages of Vβ11⁺ and Vβ5⁺ (but not Vβ8⁺) single-positive CD4⁺ thymocytes (THY) were significantly reduced in Treg-treated chimeras (red bars, n = 3) in comparison to naïve B6 mice (gray bars, n = 3; p = 0.0330 for Vβ11⁺, p = 0.0187 for Vβ5⁺) and in comparison to BMT recipients without Tregs (black bars, n = 7; p = 0.0102 for Vβ11⁺, p = 0.0151 for Vβ5⁺; 22 weeks after BMT). Deletion was assessed by multicolor flow cytometry in selected mice. Gray bars denote naïve B6 controls, white bars denote naïve Balb/c controls. p values are shown for comparison between groups (two-sided Student's t-test), error bars indicate standard deviation.