Abstract

Approximately half of the patients with acute myeloid leukemia (AML) harbor occult disease during therapy, leading to overt relapse. Novel treatments are needed to advance cure rates. AML cells are sensitive to natural killer (NK) cell cytotoxicity if they express HLA Class I molecules that do not bind killer-inhibitory receptors (KIR) on NK cells. The demonstration that haploidentical NK cells can expand in vivo and exert anti-AML activity when infused after non-myeloablative conditioning (

Miller et al.,
Blood
105
:
3051
,
2005
), provided impetus to further explore their clinical potential and develop ways to increase their efficacy. The success of NK cell therapy depends on: i) mismatch in recipient HLA and donor KIR phenotype, allowing NK cell alloreactivity; ii) infusion of sufficient numbers of NK cells to achieve an effector: target (E:T) ratio that produces a significant leukemia cytoreduction. We found that K562 cells genetically modified to express membrane-bound IL-15 and 4-1BB ligand (K562-mb15-41BBL) induced expansion of human NK cells (
Imai et al.,
Blood
106
:
376
,
2005
). In the present study, we first tested the stimulatory capacity of irradiated K562-mb15-41BBL in 34 additional healthy donors: CD56+ CD3 NK cell expansion after 7–10 days of culture was 5–87 fold (median, 22); after 21 days, NK cells could expand >1000 fold. CD3+ T cells expanded minimally or not at all. NK cells derived from 12 healthy donors were tested against the AML cell lines K562, KG-1, U937 and HL-60. Expanded NK cells were consistently cytotoxic at low E:T ratios. Thus, mean (± SD) cytotoxicity after 4 hrs at 4: 1 was 85.1% ± 8.7% for K562, 83.7% ± 9.4% for KG-1, 78.8% ± 15.2% for U937 and 94.8% ± 5.1% for HL-60. Expanded NK cells were effective even when outnumbered by target cells: at a 0.5: 1 ratio, cytotoxicities were 34.1% ± 14.7% with K562, 51.5% ± 16.5% with KG-1, 24.5% ± 14.8% with U937 and 52.1% ± 9.8% with HL-60. We next tested cytotoxicity of expanded NK cells from 10 donors against primary cells obtained from the bone marrow of 9 newly diagnosed patients with AML. Median cytotoxicity after 4 hrs of culture at a 4: 1 ratio was high, although interdonor variability was observed, with cytoxicities ranging from 22% to 90%. When expanded NK cells were cultured for 7 days with primary AML cells in the presence of bone marrow mesenchymal cells (to prevent spontaneous apoptosis of the AML cells) we could detect cytotoxicity at a 0.01:1 E:T ratio. Expanded NK cells were consistently more cytotoxic than primary NK cells from the same donor. Gene expression studies revealed marked changes in expression of adhesion molecules and cytokine transcripts after expansion. Expanded NK cells exerted considerable antileukemic effect in NOD-SCID-IL2Rgammanull mice engrafted with human AML cells, providing a strong rationale for their clinical testing. To this end, the K562-mb15-41BBL stimulatory cell line is currently being made under cGMP conditions and conditions for large-scale NK cell expansion have been established in support of a pilot protocol in which expanded haploidentical NK cells with be administered to patients with refractory AML.

Author notes

Disclosure: No relevant conflicts of interest to declare.