Receptor Repertoire Reconstitution Suggests Acquisition of Patient-Specific Tolerance by Natural Killer Cells Arising from Hematopoietic Progenitor Stem Cells after Haploidentical Transplantation

Donor-recipient incompatibility for human leukocyte antigen (HLA) ligands of killer cell immunoglobulin-like receptors (KIRs) in haploidentical hematopoietic stem cell transplantation (HSCT), has been associated with a selective graft versus leukemia (GvL) effect mediated by donor-derived alloreactive natural killer (NK) cells expressing KIRs whose ligands are missing in the recipient. In this study, we show that NK cells arising from hematopoietic stem cell progenitors after transplantation into haploidentical recipients, acquire a receptor repertoire that is compatible with patient-specific tolerance due to engagement of patient HLA ligands by inhibitory NK receptors. Using four-color immunofluorescence with monoclonal antibodies (mAbs) specific for the receptors CD94/NKG2A, KIR2DL1/2DS1, KIR2DL2/2DL3/2DS2 and KIR3DL1, we have analyzed NK receptor reconstitution kinetics in eleven adult patients affected by acute myeloid (n=9) or lymphoblastic (n=2) leukemia, who underwent HSCT from a KIR ligand matched (n=5) or mismatched (n=6) haploidentical family donor, using high doses (median 12.5x106/kg) of purified CD34+ progenitors. Nine patients achieved long-term (>150 days) complete remission of disease, independently from disease status at time of transplantation, and, importantly, from the presence (n=5) or absence (n=4) of donor NK alloreactivity. Within the first two months after transplantation, the vast majority (96% at 30 days, 86% at 60 days; SD 2% and 11%, respectively) of NK cells arising in the patients expressed the inhibitory receptor CD94/NKG2A, whose ligand HLA-E is ubiquitously expressed by cells positive for classical HLA class I molecules including leukemic blasts. As shown by mAb inhibition studies, lysis of patient-derived phytohemagglutinin-activated T cell blasts by these early arising NK cells was specifically inhibited by engagement of CD94/NKG2A. KIR expression was restored with variable kinetics in the later post-transplantation phase (3–9 months). Interestingly, however, during this period, NK cells devoid of CD94/NKG2A were found to express at least one KIR specific for an HLA ligand present in the patient, suggesting functional silencing of NK cells arising in the later phases after transplantation by acquisition of specific KIRs. Taken together, these data challenge current broad view on putative antileukemic effect of alloreactive NK cells reconstituting from haploidentical donor CD34+ cells and suggest that optimal exploitation of NK alloreactivity for GvL requires the presence of NK cells matured in the context of the donor’s rather than the recipient’s HLA repertoire. Ultimately, these findings provide a rationale for emerging clinical evidence in favor of efficacy of NK-based immunotherapy with mature donor NK cells.