Despite intensive chemotherapy, acute lymphoblastic leukemia (ALL) recurs in approximately 20% of children and 65% of adults. For most of these patients, allogeneic hematopoietic cell transplantation (HCT) is the only curative treatment option but risk of relapse after transplantation is high. Donor lymphocyte infusions can suppress leukemia relapse but they are generally ineffective in ALL.

We and others have shown that T lymphocytes transduced with anti-CD19 chimeric receptors have remarkable anti-ALL capacity in vitro and in vivo, suggesting the clinical testing of receptor-modified autologous T cells in patients with persistent minimal residual disease. However, the use of allogeneic receptor-modified T lymphocytes after HCT might carry the risk of severe graft-versus-host disease (GvHD). In this setting, the use of CD3-negative NK cells is attractive because they should not cause GvHD.

Spontaneous cytotoxicity of NK cells against ALL is weak, if measurable at all. To test whether anti-CD19 chimeric receptors could enhance it, we developed a method to specifically expand human primary NK cells and induce high levels of receptor expression. Specific NK cell expansion has been problematic to achieve with established methods, because CD3+ T cells expand preferentially; even with T-cell depletion, residual T cells typically become prominent after stimulation. We overcame this obstacle by generating a genetically-modified K562 myeloid leukemia cell line that expresses membrane-bound interleukin-15 (IL-15) and 4-1BB ligand (CD137L) (K562-mb15-137L). Peripheral blood mononuclear cells from 8 donors were cultured with K562-mb15-137L in the presence of 10 IU/mL IL-2. After 1 week of culture with K562-mb15-137L, CD3- CD56+ NK cells expanded by 16.3 ± 5.9 fold, whereas CD3+ T cells did not expand. The stimulatory effect of K562-mb15-137L was much higher than that of K562 cells transduced with control vectors, K562 expressing membrane-bound IL-15 or CD137L alone, or K562 expressing wild-type IL-15 instead of membrane-bound IL-15.

NK cell exposed to K562-mb15-137L were transduced with a retroviral vector and the anti-CD19-BB-ζ receptor, consisting of the single-chain variable domain of an anti-CD19 monoclonal antibody, the hinge and transmembrane domains of CD8α, and the signaling domains of CD3ζ and 4-1BB. 4-1BB mediates signals that are crucial for immune response to tumors in vivo and significantly improves chimeric receptor signaling. In 27 experiments, mean (± SD) transduction efficiency after 7–14 days was 67.5% ± 16.7%. Seven days after transduction, 92.3% (range 84.7%–99.4%) of cells were CD3- CD56+ NK cells; expression of receptors on the cell surface was high. NK cells expressing anti-CD19-BB-ζ had powerful cytotoxicity against NK-resistant B-lineage ALL cell lines and primary ALL cells. NK cells transduced with anti-CD19-BB-ζ had consistently higher cytotoxicity than those transduced with receptors lacking 4-1BB.

The method described here allows specific expansion of primary NK cells and highly efficient transduction of chimeric receptors. Expression of anti-CD19-BB-ζ receptors in NK cells markedly enhances their anti-ALL activity. This approach could be a valuable treatment option for patients with refractory or relapsed B-cell malignancies after HCT.

Author notes

Corresponding author