Natural CD4+CD25+ regulatory T cells (Treg) contribute to tolerance induction after transplantation. We previously showed that the adoptive transfer of donor-derived Treg cells prevents lethal graft-versus-host disease (GVHD) after allogeneic bone marrow transplantation (BMT) in murine disease models. In contrast, host-type Treg cells failed to protect when co-transplanted under identical conditions. We now examined whether MHC compatibility between Treg cells and conventional CD25CD4+ and CD8+ T cells (Tconv) is required for the suppression of alloresponses, or whether elimination of host-type Treg by allo-aggressive donor Tconv cells occurred. To address this issue, mixed lymphocyte reactions were performed in which CFSE-labelled responder T cells (Tresp), Treg cells and antigen presenting cells (APC) were systematically varied with regard to their MHC haplotype. When BALB/c (H-2d) Tresp cells were stimulated with mixed BALB/c and C57BL/6 (H-2b) APC, cultures contained 26.0 ± 3.1% and 86.2 ± 2.2% proliferating CD4+ and CD8+ T cells, respectively, on 6 d. In the presence of syngeneic BALB/c Treg cells, proliferation was decreased to 9.1 ± 4.7% and 25.1 ± 4.9% for CD4+ and CD8+ Tresp cells, respectively. In contrast, in cultures with allogeneic C57BL/6 Treg cells, proliferation remained at 22.1 ± 1.8% for CD4+ and 89.6 ± 0.4% for CD8+ Tresp cells. Comparable results were obtained with C57BL/6 Tresp cells after stimulation with F1 (C57BL/6 × BALB/c; H-2b/d) or 3rd party (DBA/1; H-2q) APC. Lack of suppression in co-cultures of MHC-mismatched Tresp and Treg cells was not caused by an early elimination of allogeneic Treg cells, as those were still detectable after 6 d of allostimulation. In corresponding in vivo studies, CB6F1 or DBA/1 recipients were protected from lethal GVHD only when Tconv and Treg cells were derived from MHC-identical donors, but not when they were from two MHC-disparate strains. Transplantation of 1 × 106 C57BL/6 Tconv cells resulted in 100% lethality of CB6F1 recipients by d56. When co-transplanted with 1 × 106 C57BL/6 Treg cells, all recipients survived for 100d, whereas only 40% survived after co-transfer of the same number of BALB/c Treg (n = 15; p = 0.004). Similarly, when 1 × 106 BALB/c Tconv cells were transplanted into CB6F1 recipients, all animals died from GVHD by d46. In contrast, all recipients of BALB/c Tconv and Treg cells (ratio1:1) survived for 100d, but only 10% of recipient mice survived after co-transfer of C57BL/6 Treg (n = 10; p < 0.001). Similar results were obtained after BALB/c and C57BL/6 T cell transfer into DBA/1 (3rd party) recipients. In conclusion, these data indicate that MHC-identity between Tconv and Treg cells is required for maximum suppression of an alloresponse and that Treg cells isolated from a 3rd party donor might not be suited for the prevention of GVHD after allogeneic BMT.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author

Sign in via your Institution