MLL-rearranged ALL is associated with an extremely poor prognosis despite intensive chemotherapy and hematopoietic stem cell transplantation (SCT). We have reported that MLL-rearranged ALL is resistant to death-inducing ligands TRAIL and FasL expressed on cytotoxic T lymphocytes, and therefore T-cell mediating graft-versus-leukemia (GVL) effect is not expected post allogeneic SCT. We recently demonstrated that MLL-rearranged ALL cells were effectively killed by allogeneic NK cells from adult peripheral blood (PB) in a perforin-dependent manner when KIR (killer cell Ig-like receptor) ligand incompatibility exists between ALL cells and NK cells. This KIR ligand incompatibility has been reported to possibly reduce the relapse rate of acute myeloid leukemia post SCT. To pursue the clinical implication in KIR ligand incompatible cord blood transplantation (CBT) for the treatment of MLL-rearranged ALL, we examined the KIRs expression on allogeneic NK cells from umbilical cord blood (CB) by flow cytometry and their in vitro cytotoxic activities against MLL-rearranged ALL cell lines by a standard Cr-release assay at an effector-to-target ratio of 20 to 40. The results were compared between NK cells from adult PB and CB, and between KIR ligand compatible and incompatible NK cells. NK cells from CB were enriched by negative selection and classified into 2 groups based on HLA-C alleles; C1/C1 type (n=5) both alleles belonging to group I (Cw1, Cw3 et. al.) and C1/C2 type (n=4) each of alleles belonging to group I and group II (Cw2, Cw4, et. al.). NK cells from adult PB were similarly classified into C1/C1 type (n=4) and C1/C2 type (n=5). All of the MLL-rearranged ALL cell lines established in our laboratory (n=10) were C1/C1 type, and two cell lines with MLL-ENL (KOPN1, KOCL50) were used as targets. K562 lacking HLA class I expression was used as a positive control target. Although there was no significant difference in the HLA-C group II receptor (KIR2DL1, CD158a) expression between CB- and adult PB-NK cells (17.8±6.3% vs. 26.5±15.2%), the HLA-C group I receptor (KIR2DL2/L3, CD158b) expression on CB-NK cells was significantly lower than on adult PB-NK cells (24.3±10.2% vs. 47.4±19.2%, p=0.009). The CD158b expression showed no difference between C1/C1 and C1/C2 types of CB-NK cells, but it expressed higher on C1/C1 type of adult PB-NK cells than on C1/C2 type (58.9±17.4% vs. 35.9±14.0%. p=0.047), suggesting that the CD158b expression on NK cells increases as getting older particularly in C1/C1 type individuals. In the cytotoxic assay, CB-NK cells irrespective of C1/C1 and C1/C2 types exhibited a lower cytotoxicity against K562 compared with adult PB-NK cells (42.0±19.2% vs. 63.6±9.5%, p=0.009). Of importance, although both C1/C1 and C1/C2 CB-NK cells showed a similar cytotoxicity against K562, C1/C2 CB-NK cells exhibited a significantly higher cytotoxicity against C1/C1 MLL-rearranged ALL cell lines than did C1/C1 CB-NK cells when assessed by a relative cytotoxicity to K562 (KOPN1, 0.84±0.19 vs. 0.47±0.13, p=0.028; KOCL-50, 0.87±0.27 vs. 0.40±0.14, p=0.028), suggesting that a loss of inhibitory signal to CD158a on NK cells from leukemia cells can specifically enhance their alloreactivity. Taken together, MLL-rearranged ALL cells are sensitive to killing by KIR ligand incompatible allogeneic CB-NK cells, and therefore the maximal GVL effect against this leukemia could be expected if the specific CB whose NK cells can exert their alloreactivity is selected for CBT.

Author notes

Disclosure: No relevant conflicts of interest to declare.