CCL22-producing CD8α- myeloid dendritic cells mediate regulatory T cell recruitment in response to G-CSF treatment.

G-CSF prevents type 1 diabetes in the NOD mouse by promoting the local recruitment of T regulatory cells (Tregs). This is an indirect effect because adoptive transfer of G-CSF-induced tolerogenic dendritic cells (DCs) promotes Treg accumulation. However, the identity of the particular DC subset and the molecule(s) mediating this effect remain unknown. We demonstrate in this study that the adoptive transfer of CD11c(high)CD8α(-) DCs isolated from pegylated G-CSF (pegG-CSF) recipients, but not that of other DC subtypes, enhanced the pancreatic recruitment of CD4(+)CD25(+)Foxp3(+) Tregs, which generated increased amounts of TGF-ß. Likewise, only CD11c(high)CD8α(-) DCs from pegG-CSF recipients secreted the chemokine CCL22 at levels that effectively attracted Tregs. PegG-CSF was more efficient at enhancing the synthesis of CCL22 by CD11c(high)CD8α(-) DCs from the pancreatic lymph nodes compared with those from the spleen. Accordingly, CD11c(high)CD8α(-) DCs from the pancreatic lymph nodes of pegG-CSF recipients were more efficient than their splenic counterparts in the recruitment of Tregs upon adoptive transfer. Predictably, CD11c(high)CD8α(-) DCs failed to recruit these Tregs both in vivo and in vitro following intracellular neutralization of CCL22. These data assign a key role to CD8α(-) DCs and CCL22 in Treg recruitment in the protection of NOD mice against type 1 diabetes following the treatment with G-CSF.

Early posttransplantation donor-derived invariant natural killer T-cell recovery predicts the occurrence of acute graft-versus-host disease and overall survival.

Invariant natural killer T (iNKT) cells can experimentally dissociate GVL from graft-versus-host-disease (GVHD). Their role in human conventional allogeneic hematopoietic stem cell transplantation (HSCT) is unknown. Here, we analyzed the post-HSCT recovery of iNKT cells in 71 adult allografted patients. Results were compared with conventional T- and NK-cell recovery and correlated to the occurrence of GVHD, relapse, and survival. We observed that posttransplantation iNKT cells, likely of donor origin, recovered independently of T and NK cells in the first 90 days after HSCT and reached greater levels in recipient younger than 45 years (P = .003) and after a reduced-intensity conditioning regimen (P = .03). Low posttransplantation iNKT/T ratios (ie, < 10(-3)) were an independent factor associated with the occurrence of acute GVHD (aGVHD; P = .001). Inversely, reaching iNKT/T ratios > 10(-3) before day 90 was associated with reduced nonrelapse mortality (P = .009) without increased risk of relapse and appeared as an independent predictive factor of an improved overall survival (P = .028). Furthermore, an iNKT/T ratio on day 15 > 0.58 × 10(-3) was associated with a 94% risk reduction of aGVHD. These findings provide a proof of concept that early postallogeneic HSCT iNKT cell recovery can predict the occurrence of aGVHD and an improved overall survival.