In murine disease models, the adoptive transfer of donor CD4+CD25+ regulatory T cells (Treg) prevents lethal graft-versus-host disease (GVHD) after allogeneic bone marrow transplantation. We recently described methods for the isolation and in vitro expansion of human Treg from adult peripheral blood and found that only CD45RA+ but not CD45RA CD4+CD25high Treg (RA+ and RA Treg, respectively) maintain FOXP3 expression as well as suppressive activity after prolonged culture (

Hoffmann et al.
Blood
108
:
4260
;
2006
). Here, we compared the immunoregulatory capacity of these two different Treg subpopulations in vivo in a xenogeneic GVHD (xGVHD) model. Transplantation of 20x106 human peripheral blood mononuclear cells (huPBMCs) into immunodeficient BALB/c Rag2−/−γc−/− mice led to engraftment and extensive proliferation of human CD45+ cells (huCD45+), which was accompanied by clinical signs of xGVHD. Histological and flow cytometric evaluation revealed that mainly CD8+ T cells within PBMC caused xGVHD and that main targets were liver, lung and spleen as well as mouse hematopoiesis. When expanded CD45RA+ Treg were co-transplanted at a 1:1 ratio with CD4+ and CD8+ T cells contained in huPBMC, the frequency of huCD45+ cells in peripheral blood (PB) at d14 was reduced to 12 ± 4.9 % (n=15), while animals that received expanded RA Treg had 30 ± 7.5 % (n=15) and animals without co-transfer of Treg 40 ± 8.3 % (n=14) huCD45+ cells. The expansion of conventional T cells, however, was only delayed but not prevented, as by d21 the percentage of huCD45+ cells in PB increased to 35 ± 9.2 % (n=15) when RA+ Treg were co-transplanted as compared to 40 ± 9.3 % (n=15) with RA Treg and 50 ± 6.4 % (n=14) without Treg administration. Subsequently, all mice developed clinical signs of xGVHD and 73.3 % of mice co-transplanted with RA+ Treg, 86.7 % co-transplanted with RA Treg and 78.6 % of mice transplanted with huPBMCs alone died within 100 days. Since Treg mainly inhibited early expansion of huPBMC, we examined the splenic cellularity of xeno-transplanted animals at d10 after cell transfer. While spleens of mice that received only huPBMC contained 20 ± 3x106 huCD45+ cells (n=10), co-transfer of RA Treg reduced huCD45+ cells to 14 ± 4.8×106 (n=9) whereas animals that received RA+ Treg contained only 5 ± 1.1×106 huCD45+ cells (n=11) at that time. Furthermore, IL-2 and IFN-γ production of conventional T cells re-isolated from the spleen on d10 was reduced 5fold in mice co-transplanted with RA+ Treg as compared to recipients of only huPBMCs or additional RA Treg. Thus, in vitro expanded human RA+ Treg suppress the proliferation and cytokine production of conventional T cells in early phases of xGVHD, but ultimately do not prevent disease onset and mortality.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author

Sign in via your Institution