TNFα sensitizes AML cells to the antileukemic activity of DNTs. (A-B) AML cell lines (top), primary AML blasts, and allogeneic PBMCs (bottom) were pretreated with or without rTNFα, washed with phosphate-buffered saline (PBS), then cocultured with DNTs for 2 to 4 hours. Percentage specific killing of AML cells are shown with a red line indicating the 10% specific killing as resistance threshold (A). Linear regression analysis performed between percentage specific killing and the relative percentage increase in specific killing of rTNFα-sensitized AML (B). Each symbol represents an AML cell line or primary AML sample. Experiments were done in triplicates and were performed with at least 2 DNT donors. Numbers represent the ID of patients with AML. (C-D) Primary AML samples (130794 and 140176) were treated with or without rTNFα, washed with PBS, then cocultured with DNTs. Gating strategy (C) and percentage specific killing of rTNFα-treated and nontreated groups of the primary AML samples gated on specified leukemic blast populations in triplicates are shown (D). (E) DNTs were cocultured with AML cell lines for 24 hours, in the presence of increasing (for OCI-AML3) or fixed (for OCI-AML2 and MV4-11, 10 μg/mL) concentration of anti-TNFα blocking antibody or isotype control antibody. The graphs shown are representative of 2 independent experiments. (F-G) Sublethally irradiated (225 cGy) NSG mice were IV injected with OCI-AML2 untreated or treated with rTNFα, followed by 2 infusions of DNTs or PBS. Schematic of the in vivo xenograft mouse model with treatment schedule (F). Bar graphs represent the mean AML bone marrow engraftment levels (left) and engraftment levels normalized to the PBS group (right) (G). Each symbol represents an individual mouse. Data represent the mean ± standard deviation (SD) and pooled from 3 independent experiments (n = 3-5 per group). Student t test, 1-way/2-way ANOVA, and linear regression analysis were used. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ns, nonsignificant.