Simultaneous Use of Histone Deacetylase Inhibitor Valproic Acid Potentiates Natural Killer Cell Cytotoxicity against T-Lineage Acute Lymphoblastic Leukemia

Leukemic cells could survive after aggressive therapies including mega-dose chemotherapy. Natural killer (NK) cells have great promise for treatment of such refractory disease because many kinds of hematopoietic malignancies and solid tumors are reported to be intrinsically sensitive to NK cell lysis. Indeed, NK cell allo-reactivity has been shown to reduce leukemia relapse after haploidentical stem cell transplantation (SCT) and cord blood transplantation (CBT) in patients with myeloid leukemia (Ruggeri et al. Science 2002; Willemze, et al. EBMT 2008 abstract). However, NK cell allo-reactivity against acute lymphoblastic leukemia (ALL) has been reported to be generally weak regardless of the presence of killer immunoglobulin receptor (KIR) ligand mismatch. We previously showed that genetic modification of primary NK cells with anti-CD19 chimeric receptor could overcome the refractoriness of B-lineage ALL cells to NK cell lysis (Imai et al, Blood 2005). For T-lineage ALL, however, development of antigen-specific artificial receptors against T-ALL cells has been hampered by the similarity in immunophenotype between NK cells and T cells. Novel strategy to augment anti-T-ALL activity of NK cells, such as combined use of a kind of agents which has anti-leukemic activity and does not inhibit NK cell activity, is needed. Valproic acid (VPA) has been widely used as an anticonvulsant and a mood stabilizer for a long time, and its safety has been established. Recently, a number of studies have shown that VPA could act as a histone deacetylase (HDAC) inhibitor and possess anti-cancer activity. We thus studied the effect of VPA on NK cell cytotoxicity against T-ALL cell lines and asked whether VPA could be incorporated into NK cell therapy. NK cells were activated and expanded by coculture of peripheral blood mononuclear cells with K562 cells transfected with 4-1BB ligand and membrane-bound interleukin-15. After 7 days, cells were subjected to CD3-depletion procedure to earn highly pure (>99%) NK cells, and then cultured in the presence of high dose interleukin (IL)-2 (1000 U/ml). Efficient killing of T-ALL cell lines MOLT4, Jurkat and CEM-C7 by activated NK cells were observed in 4-hour cytotoxicity assay, even if KIR ligand mismatch were not present. VPA suppressed leukemic cell proliferation and induced apoptosis in MOLT4, Jurkat and CEM-C7 at clinically achievable concentrations and in a dose-dependent fashion. Exposure of activated NK cells to VPA at 100 mg/ml for 48 hours in the presence of IL-2 resulted in mild inhibition of NK cell proliferation by approximately 20% and decreased surface expression of activating receptors such as NKG2D, 2B4, and natural cell cytotoxicity receptors p30, p44, p46 (10–40% decrease in mean fluorescence intensity). However, the addition of VPA to coculture of NK cells with MOLT-4, Jurkat and CEM-C7 did not result in interfering the anti-leukemia activity of NK cells. When anti-leukemic activity was evaluated in more prolonged culture experiments at very low effector: target (E/T) ratio, the number of residual leukemic cells was significantly lower in culture with VPA compared to those without VPA (MOLT-4: residual leukemia cells 12.2 ± 1.9% vs 46.0 ± 2.3% of control culture at E/T ratio of 0.13:1; CEM-C7: 21.0 ± 0.4% vs 40.1 ± 1.1% of control culture at E/T ratio of 0.13:1, 72-hour incubation). Similar results were also obtained when VPA was substituted by another HDAC inhibitor Suberoylanilide hydroxamic acid (SAHA). We found that upregulation of NK-activating receptor ligands MICA/B, ULBP1 and ULBP2 occurred in MOLT-4, Jurkat and CEM-C7 after 24–48 hour exposure to VPA at 100 mg/ml. On the contrary, T-ALL cells did not show increased sensitivity to NK cell lysis after pretreatment by VPA for 48 hours. Therefore, we thought that the synergistic effect of VPA observed in prolonged culture experiments could be mainly attributed to growth suppression of continuously growing leukemic cells, not to its immunomodulatory activity. These data suggest that simultaneous use of VPA would not abrogate NK cell cytotoxic activity and VPA might be used in combination with infusion of NK cells for patients bearing VPA-sensitive tumors. In patients after haploidentical SCT and CBT, NK cell allo-reactivity might be potentiated by the use of VPA.