Myeloid-derived suppressor cells (MDSCs) are proposed to control graft-versus-host disease (GVHD) in allogeneic hematopoietic stem cell transplantation (allo-HSCT). However, the definition of human MDSCs has not yet reached consensus. Granulocyte colony-stimulating factor (G-CSF) has been routinely used to mobilize stem cells to peripheral blood in healthy donors. It was also recognized as a novel mediator of T-cell tolerance. However, the effects of G-CSF administration on donor-derived MDSCs and the further regulatory effects of these MDSCs on GVHD remained unclear.


The aim of this study is to evaluate the in vitro and in vivo effects of G-CSF expanded, donor-derived MDSCs (HLA-DR-/lowCD16-CD33+) in preventing acute GVHD after allo-HSCT.


The frequency and cell numbers of different kinds of MDSCs in peripheral blood before and after G-CSF administration from 10 healthy donors were analyzed by flow cytometry. Cells morphological features were detected by May-Grünwald-Giemsa cytospin. Secondly, the suppressive and regulatory functions of HLA-DR-/lowCD16-CD33+ population on CD3+ T cells were assessed via in vitro experiments. A humanized xenogeneic acute GVHD model was established to determine whether this population could prevent acute GVHD in vivo. Furthermore, a clinical prospective cohort study enrolled one hundred consecutive transplant recipients was performed to assess the effects of HLA-DR-/lowCD16-CD33+ contained in HSC grafts on the occurrence of acute GVHD.


The findings of this study include: First, a novel phenotype of HLA-DR-/lowCD16-CD33+ MDSCs with suppressive function and morphological features similar to those of immature monocyte was identified. The median of percentages of this subset were significantly increased both in peripheral blood (PB, 6.5% vs. 4.6%, P=0.0122) and peripheral blood stem cells harvest (PBSCs, 15.5% vs. 4.6%, P<0.0001) after treating healthy donors with G-CSF than those of PB before mobilization. The median of percentage of interleukin-10 (IL-10) and transforming growth factor beta (TGF-β) positive cells among HLA-DR-/lowCD16-CD33+ population were both significantly higher than those of PB before mobilization. (IL-10+, 10.1% vs.0.62%, P=0.002; TGF-β+, 60.24% vs. 10.39%, P=0.0003). Donor-derived HLA-DR-/lowCD16-CD33+population inhibited the proliferation of T cells in a TGF-β-dependent manner. In vitro cell co-culture experiments also showed that this MDSCs subset could promote regulatory T cells (Treg, CD4+CD25+Foxp3+) expansion and induce T helper 2(Th2, CD4+IL-4+) differentiation (The median of percentage of Treg in co-culture with MDSCs group, 21.4% vs. without MSDCs group, 8.35%, P=0.0048. The median of fold change of Th2/Th1 in co-culture with MDSCs group, 1.35 vs. without MSDCs group 0.98, P=0.0159. The median of fold change of Th2/(Th1+Th17) in co-culture with MDSCs group, 1.28 vs. without MSDCs group 1.00, P=0.0095. Th1 (CD4+IFNγ+), Th17 (CD4+IL-17A+)). Second, we demonstrated that these cells could prevent acute GVHD in a humanized mouse model. Adoptive transfer human G-CSF-mobilized HLA-DR-/lowCD16-CD33+ cells significantly prolonged the survival and ameliorated the weight loss and tissue damage in GVHD mice. Third, clinical cohort results showed that the number of HLA-DR-/lowCD33+CD16- cells in the donor graft was the only independent risk factor inversely correlated with the incidence II-IV acute GVHD in recipients (HR 0.388, 95% CI: 0.158-0.954, P=0.039).


Our results suggest that MSDCs with HLA-DR-/lowCD16-CD33+ phenotype in G-CSF-mobilized PBSCs have monocytic features and immune-regulatory properties, which could alleviate acute GVHD in the allo-HSCT settings.

Key words: Myeloid-derived suppressor cells; granulocyte colony-stimulating factor; graft-versus-host disease


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.