NK-92 Preferentially Targets Acute Myeloid Leukemia Stem Cells

Outcome of patients with acute myeloid leukemia (AML) remains poor because many relapse from residual leukemic stem cells (LSCs) enriched in the CD34+CD38- fraction of AML blasts. NK-92 is a human permanent natural killer (NK) cell line in phase I clinical trials for relapsed and refractory malignancies. We recently showed that NK-92 targets LSCs preferentially over bulk leukemia in the cell line KG1 (Cytotherapy. 2010 Mar 15). Here, we evaluate the action of NK-92 and another NK cell line, KHYG-1, against five AML cell lines and five primary AML samples by the chromium release and methylcellulose cytotoxicity assays to determine the mechanism of recognition and killing and the effect on leukemic stem cells.

Using a 4 hour chromium release assay (CRA), the highest effector:target (E:T) ratio tested (25:1) for NK-92 and KHYG-1 against cell line targets revealed % lysis as follows: K562: 81.2+/−5.4%; 82.2 +/−3.9%; KG1: 41.0+/−7.2%; 37.3% +/−3.6; OCI/AML2: 37.3+/−22.7%; 33.0+/−13.9%; OCI/AML3: 33.8+/−4.5%; 51.0+/− 6.8% and OCI/AML5: 99.0+/−4.9%; 52.9 +/− 5.6%, respectively. Killing by NK-92 and KHYG-1 was completely inhibited by calcium chelation using 4 mM EGTA for all cell lines tested. Blockade of class I HLA on target cells using 10 ug/ml of anti-class I monoclonal antibody did not affect killing by NK-92 and KHYG-1 except for a decrease in killing of OCI/AML5 by both NK-92 and KHYG-1 from 99.0 +/− 4.9% to 39.0 +/−3.2% and 52.9+/−5.6% to 19.6 +/− 6.25%, respectively. Blockade of NKG2D using 10 ug/ml of anti-NKG2D monoclonal antibody did not significantly affect killing of AML cell lines by NK-92 or KHYG-1. Five primary AML blast samples treated with NK-92 at 25:1 E:T yielded only slight to moderate degrees of cytotoxicity by CRA: 15.6 +/−12.7%, 42.3+/−3.6%, 29.8 +/−3.6%, 43.9 +/− 1.47%, 42.6 +/− 0.1%lysis. KHYG-1 at 25:1 E:T had minimal killing of this panel of AML blasts by CRA: 1.27 +/−21.9%, 9.8+/−2.2%, 5.2+/−2.5%, 17.1+/−2.8%, 8.5+/−3.3% lysis. Blockade of class I HLA did not affect killing of primary AML blasts by NK-92, but the fourth sample only was rendered more sensitive to killing by KHYG-1 from 17.1 +/−2.8% to 35.8 +/−1.2% lysis. Blockade of NKG2D with 10 ug/ml anti-NKG2D monoclonal antibody did not significantly affect killing of primary AML samples by NK-92 or KHYG-1. To further assess killing by NK-92 of LSCs from primary AML, we used a methylcellulose cytotoxicity assay (MCA) established previously by our lab. The MCA showed that NK-92 at 25:1 E:T eliminated clonogenic growth of 3/5 primary AML blast samples with minimal colony growth in 2/5 with % cytotoxicity of 100 +/−0%, 86.3 +/− 2.3%, 98.4 +/−2.8%, 100 +/− 0% and 100 +/−0%, demonstrating higher toxicity than obtained with the CRA. Primary AML derived CD34+CD38- sorted LSCs were more sensitive to killing than CD34+CD38+ blasts by NK-92 in a 4 hour CRA at 1:1 E:T: 58.9+/−11.5%, 20.3 +/−1.71%; 5:1: 78.3 +/−9.7%, 43.5+/−11.1% and 10:1: 72.9+/−5.6%, 38.5 +/−2.4% lysis.

We demonstrate cytotoxicity of NK-92 and KHYG-1 against a range of AML targets and show greater killing of primary AML blasts by NK-92. The mechanism of cytotoxicity for both cell lines is primarily by granule exocytosis and is NKG2D independent. Unexpectedly, blockade of class I HLA resulted in reduced killing of OCI/AML5 by both NK-92 and KHYG-1, suggesting the presence of an activating KIR receptor common to both cell lines. One of five primary AML samples had increased killing by KHYG-1, suggesting KIR mediated inhibition. Further assessment of NK-92 against LSCs by the MCA demonstrated greater cytotoxicity than with the CRA and indicated preferential killing of LSCs. This finding was confirmed using the CRA with sorted immunophenotypically defined CD34+CD38- LSCs and CD34+CD38+ blasts. Our findings support the use of NK-92 in the treatment of AML.

No relevant conflicts of interest to declare.