A mechanism for increased sensitivity of acute myeloid leukemia to mitotoxic drugs

Abstract

Mitochondria play a central and multifunctional role in the progression of tumorigenesis. Although many recent studies have demonstrated correlations between mitochondrial function and genetic makeup or originating tissue, it remains unclear why some cancers are more susceptible to mitocans (anticancer drugs that target mitochondrial function to mediate part or all of their effect). Moreover, fundamental questions of efficacy and mechanism of action in various tumor types stubbornly remain. Here we demonstrate that cancer type is a significant predictor of tumor response to mitocan treatment, and that acute myeloid leukemias (AML) show an increased sensitivity to these drugs. We determined that AML cells display particular defects in mitochondrial metabolism that underlie their sensitivity to mitocan treatment. Furthermore, we demonstrated that combinatorial treatment with a mitocan (CCCP) and a glycolytic inhibitor (2-deoxyglucose) has substantial synergy in AML cells, including primary cells from patients with AML. Our results show that mitocans, either alone or in combination with a glycolytic inhibitor, display anti-leukemia effects in doses much lower than needed to induce toxicity against normal blood cells, indicating that mitochondria may be an effective and selective therapeutic target.

Conflict of interest statement

The authors declare that they have no conflict of interest.

Figures

Fig. 1. Cell lines derived from AML are more susceptible to mitochondrial damage than cell lines derived from solid tumors.

aZ-scores of tumor cell lines from the NCI-60 panel. Sensitivity to non-mitocan drugs is shown on the x-axis, sensitivity to mitocans is shown on the y-axis. U251, a glioblastoma-derived cell line with high sensitivity to mitocans is shown in red, and ovarian cancer-derived cell lines, which show higher resistance, are shown in blue. b Density plot showing median rank for each possible permutation of six cell lines from the NCI-60 collection. The thick black line represents the median for leukemia cell lines (5.5) from the NCI-60 panel. c Sensitivity of AML (MV-4-11, THP-1, OCI-AML2, and MOLM-13), normal PBMCs, and solid tumor (U251, SKOV3, and OVCAR3) cell lines to mitocan treatment. Shown are LD50 values with 95% confidence intervals for MTX, DOX, CCCP, and ara-C based on results from 3–5 independent experiments. Comparisons of LD50 were done by the ratio test, the asterisk indicates significant difference compared with the next most sensitive cell line. d Fluorescence micrographs of MV-4-11 (top) or THP-1 (bottom) cells treated with either vehicle (left) or LD50 concentrations of CCCP. Cells were stained with acridine orange/propidium iodide. e The ratio of mitochondrial to genomic DNA was determined by quantitative PCR. Shown are mean values with SD. Statistical significance for comparison AML vs. healthy PBMCs was analyzed via the Student’s t-test. f Mitochondrial health in untreated MOLM-13 cells and normal PBMCs were compared, including mitochondrial mass (assessed via staining with MitoTracker Green), membrane potential (assessed via staining with JC-1), metabolic rate (assessed via Seahorse analysis of oxygen consumption rate and lactate production), steady-state ATP level, and protein level of the β subunit of ATP synthase. SKOV3, a mitocan-resistant cell line, is shown as a comparison. Shown is the mean of at least three independent experiments (in case of Seahorse data and ATP measurements dots represent all technical replicates), error bars are SD. Statistical analysis was performed using Student’s t-test with independent samples. ***p < 0.001; **p < 0.01; *p < 0.05

Fig. 2. AML-derived cells are more sensitive to mitocan treatment.

a Relative mtDNA content before and after 24 h exposure to doxorubicin treatment in AML cell lines (left) or in solid tumor cell lines (right). b Correlation between basal mtDNA content and LD50 of mitocans. Normalization of LD50 values was performed using min-max approach, linear functions were adjusted for the data using lm() method in R. Basal ATP level (c) and ATP-linked respiration (d) before and after exposure to doxorubicin for 24 h (c) or 4 h (d). e ATP synthase β subunit protein level before and after 24 h exposure to doxorubicin. β-actin was used as a loading control. f Changes in mitochondrial mass after 24 h doxorubicin treatment. Comparison of MOLM-13 vs. normal PBMCs is shown. g Mitochondrial membrane potential before and after 4 h doxorubicin treatment, as defined by JC-1 staining. Representative plots for red/green fluorescence (PE vs. FITC) after standard compensation for MOLM-13, healthy PBMCs, and SKOV3 cells are shown below for the untreated (blue) and doxorubicin-treated (red) cells. h Expression ratios of mitochondrial fusion/fission genes before and after 8 h doxorubicin treatment. The results of 3–6 independent experiments are presented as mean ± SD (in a, c, f, g, h) or mean ± SEM (in d, e). All technical replicates are shown in c. For statistical significance, either Student’s t-test with independent samples (a, c, e–h) or ANOVA with subsequent Dunn’s test (d) was used. ***p < 0.001; **p < 0.01; *p < 0.05; ns: p > 0.05

Fig. 3. Combination therapy exhibits synergy in AML cells.

FACI (combination index as a function of fraction affected) plots for combination treatment with CCCP and 2-DG in OCI-AML2, healthy PBMCs, and SKOV3 cells (a) or primary AML cells for one (upper) or multiple (n = 21, lower) patients (b). Non-linear regression was used to generate a function for combined FACI data for primary AML samples. c Drug combination landscapes for OCI-AML2 cells and normal PBMCs built using the Bioconductor package ‘synergyfinder’. d Viability of OCI-AML2 cells and normal PBMCs after treatment with CCCP/2-DG at LD25 concentrations corresponding to OCI-AML2 cell line. e ATP level before and after treatment with CCCP and 2-DG in AML cell lines, healthy PBMCs, and solid tumor cell lines (upper) or primary AML samples (lower). f Respiration (upper) and coupling efficiency (lower) before and after treatment with CCCP, 2-DG, or both. g Viability of OCI-AML2 and normal PBMC cells after treatment with ABT-199/2-DG. All graphs show results from at least three independent experiments (except b, where primary AML cells were limited by patient samples). FACI plots reflecting combination indices with 95%-confidence intervals were built using the Bioconductor package ‘drc’, * significant difference from combination index = 1 (additivity). Elsewhere, results are presented as mean ± SD (d, e, g) or mean ± SEM (f). Statistical significance was tested using Student’s t-test with independent samples (d, e, g) or by ANOVA with subsequent Dunn’s test (f). ***p < 0.001; **p < 0.01; *p < 0.05; ns: p > 0.05

Fig. 4. Bioenergetic profiling of AML cell lines, healthy PBMCs, and solid tumor cells.

a Oxygen consumption rate (OCR) measured in untreated cells using a Seahorse flux analyzer. Oligomycin (1), FCCP (2), and antimycin A (3) were added at the times indicated. b OCR measured in cells either untreated (blue) or treated with doxorubicin (red), CCCP (black), 2-DG (purple), or CCCP and 2-DG (orange). c ECAR in untreated AML, PBMCs, and SKOV3 cells. d Basal OCR, ECAR, ATP-linked OCR, proton leak, and coupling efficiency for AML cell lines, PBMCs, and a mitocan-resistant solid tumor cell line. (red—AML vs. PBMCs, blue—AML vs. SKOV3). e Intracellular ROS measured by dihydroethidium (DHE, 2.5 μg/ml) after 24 h treatment with CCCP. Mean fluorescence intensities (MFI) of DHE and Hoechst 33342 were quantified using ZEN software and presented as MFI ratios DHE/Hoechst 33342. All graphs show results as mean ± SEM from 3–6 independent experiments. Statistical testing was performed by Student's t-test with independent samples (e) or ANOVA with subsequent Dunn’s or Fisher’s LSD test (d). ***p < 0.001; **p < 0.01; *p < 0.05; ns: p > 0.05

Fig. 5. Mitocan treatment activates multiple cell death pathways.

Survival rates of AML cells and healthy PBMCs treated with mitoxantrone (a), doxorubicin (b), CCCP (c), or 2-DG (d) at LD50 doses. For each drug, cells were also tested with or without the following cell death pathway inhibitors: z-VAD-fmk (40 μM, pan-caspase inhibitor), 3-methyladenine (5 mM, autophagy inhibitor), or VX-765 (10 μM, caspase-1 inhibitor). Results show the mean ± SD from at least three replicates. Statistical testing was performed by Student’s t-test with independent samples. ***p < 0.001; **p < 0.01; *p < 0.05; ns: p > 0.05