Atovaquone induces apoptosis of pediatric patient AML blasts but not normal bone marrow samples and produces an additive effect in combination with cytarabine, daunorubicin, and CPX-351. (A) Twelve primary pediatric AML samples supported with HS5 TW coculture demonstrated increased apoptosis over time when treated with 30 μM of atovaquone. For 3 representative samples, the percentage of annexin V⁺/propidium iodidepositive (PI⁺) is shown for VC and atovaquone-treated wells at each time point to demonstrate the difference between spontaneous apoptosis and atovaquone-induced apoptosis. The panel to the right depicts means and standard errors of the mean (SEMs) for all 12 samples. Apoptosis rates were significantly different in primary samples exposed to atovaquone when compared with vehicle (P < .0001 by analysis of variance [ANOVA]). (B) In comparison, 3 normal bone marrow (NBM) samples supported with HS5 TW coculture demonstrated no significant increase in apoptosis over time when treated with 30 μM of atovaquone compared with vehicle. The panel to the right depicts means and SEMs for all 3 samples. (C) Representative normalized isobolograms demonstrating an additive effect of atovaquone in combination with either cytarabine, daunorubicin, or CPX-351 and a slightly antagonistic effect in combination with etoposide. For the depicted plots, median CI for atovaquone with cytarabine was 0.925; for atovaquone with daunorubicin, 1.1335; for atovaquone with CPX-351, 0.822; and for atovaquone with etoposide, 1.515. These combinations were tested using THP-1.ffluc cells, and plots are the product of 4 technical replicates. The table depicts median CI of 3 biologic replicates for each drug combination as well as the average median CI for each combination.