Treatment of high-risk leukemia with transplantation of allogenic bone marrow or stem cells from peripheral blood is limited by the scarcity of HLA-matched related donors; only 30% of patients were eligible. Unrelated cord blood transplantation (UCBT) have become emerging therapies for adult patients without matched donors, We previously demonstrated that NK cells generated after haplo-mismatched hematopoietic stem cells transplantation are blocked at an immature state characterized by specific phenotypic features and impaired functioning, having negative impact for clinical relapse. In the present work, we studied NK cells after UCBT. This study included 27 adults’ patients (median age: 42 years), with high-risk hematologic malignancies, following a non-myeloablatif regimen. Fourthy one percent (11/27) of patients experienced initial engraftment, whereas 16 subjects (59%) experienced either primary or secondary graft failure. Neutrophil engraftment occurred at a median of 20.5 days. The incidence of the relapse was 26% (7/27) and was a major cause of death in 5 subjects. The cumulative incidence of grades II, and III acute GVHD was 33.3%. Infection occurred only in 2 patients (7.4%). The survival rate was 73% at 6 months. Analysis of DNA polymorphisms in the peripheral blood revealed full (median 99.7) donor-type chimerism. Peripheral blood samples were collected during the first year following UCBT in order to realize an extensive phenotypic study of NK cells. This analysis was realized on the cytotoxic NK CD56dim cell subset, which reaches a count number at one-month post graft (M1) close to healthy volunteer. Our results shown that NK cells phenotype post-UCBT was closed to adult controls, contrasting with NK cells generated after haplo-mismatch allograft. Indeed, excepted for the receptors NKG2A and KIR2DL1/DS1, which are respectively increased (M1: p<0.0001) and decreased (M1: p=0.0077), following the engraftment, we observed an over-expression of the activator receptors (NCR) NKp30 (M1: p=0.0012) and NKp46 (M1: p=0.0202), and a remarkable stability of KIR2DL2-3/DS2-3 and KIR3DL1/DS1. These results sharply contrast with the data previously observed after haplo-mismatch allografts. Furthermore, other studyied NK receptors (ILT2, NKG2D, NKG2C, NKp44, NKp80, LAIR-1, 2B4 and Pen5) display a similar pattern, at one month post UCBT, as compared to NK from healthy controls. Moreover, following engraftment we observed a striking transient over-expression of the activation markers CD69 (M1: p<0,0001) and HLA-DR (M1: p=0,0098) on CD56dim NK cells, probably reflecting an in vivo activation. To determine the significance of these findings, we assessed the overall functional ability of NK cells following UCBT, we have observed an increase ability of NK cells to produce IFN_amma following UCBT, using intra-cytoplasmic staining (M1: p=0.0357), as well as ELISA assay on plasma samples (M1: p=0.0013). Furthermore, the functional potential of NK cells was confirmed by their ability to kill K562 and primary blasts from various HLA-mismatched AML patients, at a level currently observed with healthy controls. These results led us to consider that NK cells could play a key role in graft versus leukemia (GvL) effect, decreasing the risk of relapse following engraftment. Thus, we have observed a correlation between relapse and low expression of both KIR2DL2-3/DS2-3 (p=0.0007) and KIR3DL1/DS1 (p=0.0133), at one-month post graft. Together, these data could reflect an early NK cells maturation following engraftment. Collectively, this work could suggest a key role for the NK cells on the clinical paradox observed following UCBT consisting of a strong GvL effect associated with a low incidence of GvHD, despite the persistent immaturity of T cells. The most striking finding in this study is the significant correlation between the decreased risks of leukemic relapse among patients, with high expression of specific KIR on CD3-CD56dim cytotoxic NK cells early after transplantation. These data reflect the rapid reconstitution of a mature NK cell subset, fully functional to exert an efficient GvL effect improving clinical outcome following UCBT.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author