Potent obatoclax cytotoxicity and activation of triple death mode killing across infant acute lymphoblastic leukemia

Abstract

Survival in infants younger than 1 year who have acute lymphoblastic leukemia (ALL) is inferior whether MLL is rearranged (R) or germline (G). MLL translocations confer chemotherapy resistance, and infants experience excess complications. We characterized in vitro sensitivity to the pan-antiapoptotic BCL-2 family inhibitor obatoclax mesylate in diagnostic leukemia cells from 54 infants with ALL/bilineal acute leukemia because of the role of prosurvival BCL-2 proteins in resistance, their imbalanced expression in infant ALL, and evidence of obatoclax activity with a favorable toxicity profile in early adult leukemia trials. Overall, half maximal effective concentrations (EC50s) were lower than 176 nM (the maximal plasma concentration [Cmax] with recommended adult dose) in 76% of samples, whether in MLL-AF4, MLL-ENL, or other MLL-R or MLL-G subsets, and regardless of patients' poor prognostic features. However, MLL status and partner genes correlated with EC50. Combined approaches including flow cytometry, Western blot, obatoclax treatment with death pathway inhibition, microarray analyses, and/or electron microscopy indicated a unique killing mechanism involving apoptosis, necroptosis, and autophagy in MLL-AF4 ALL cell lines and primary MLL-R and MLL-G infant ALL cells. This in vitro obatoclax activity and its multiple killing mechanisms across molecular cytogenetic subsets provide a rationale to incorporate a similarly acting compound into combination strategies to combat infant ALL.

Trial registration: ClinicalTrials.gov NCT00933985.

Figures

Figure 1

MTT assays showing broad-spectrum single-agent obatoclax cytotoxicity in primary infant ALL. (A) Surviving fraction plots of MTT assay data by MLL subsets in diagnostic leukemia samples from 54 cases of primary infant ALL (n = 52) or infant bilineal acute leukemia (n = 2). Cells were treated for 72 hours with increasing obatoclax concentrations. Experiments were performed 1-3 times per patient sample (3-6 replicates/condition per experiment). (B) Box and whisker plots of EC50 values ascertained by inhibitory sigmoid Emax models of MTT assay data in 54 cases shown in (A). Significant differences between EC50 values were determined by Wilcoxon’s test (Table 1). (C) Kaplan-Meier curves showing similar EFS among infants with pretreatment diagnostic ALL samples in obatoclax-sensitive and obatoclax-resistant categories defined using a 176-nM cutoff. (D) Surviving fraction plot of MTT assay data for PBMCs after 72 hours of treatment with increasing obatoclax concentrations; assay was performed twice (1-3 replicates/condition per experiment).

Figure 2

Synergy between obatoclax and chemotherapy against primary pediatric MLL-AF4 ALL. Cells were treated with obatoclax alone, indicated chemotherapy drug alone, or increasing concentrations of chemotherapy drug combined with obatoclax at fixed concentrations. Data from MTT assays performed 72 hours after treatment of each combination are shown in surviving fraction plots (first and third rows) and response surface models (second and fourth rows). Each experimental point represents average of 2 independent experiments (3 replicates/condition per experiment). An inhibitory sigmoid Emax model was used to determine the EC50 and Hill coefficient for single-agent dose responses (obatoclax: EC50 = 67.2 ± 13 nM, Hill coefficient = 1.91 ± 0.60; vincristine: EC50 = 150 ± 35 nM, Hill coefficient = 0.911 ± 0.160; L-aspariginase: EC50 = 958 ± 123 U/L, Hill coefficient = 0.785 ± 0.067; etoposide: EC50 = 1125 ± 210 nM, Hill coefficient = 0.76 ± 0.12; doxorubicin: EC50 = 47.4 ± 5.99 nM, Hill coefficient = 1.13 ± 0.13; cytosine arabinodise: EC50 = 693 ± 119 nM, Hill coefficient = 0.70 ± 0.10; dexamethasone: EC50 = 62.6 ± 8.37 nM, Hill coefficient = 0.64 ± 0.05). Three-dimensional scatter plots show Loewe additivity zero-interaction response surfaces for each obatoclax–chemotherapy combination (gray spheres) derived from single-agent experiments. If actual experimental combination effects (black spheres) are above zero-interaction response surface, combination is synergistic; if below, antagonistic; and if on the response surface, additive. Note synergistic interactions between obatoclax and each chemotherapy agent tested. ADR, doxorubicin; Ara-C, cytosine arabinodise; DEX, dexamethasone; L-ASP, L-aspariginase; VCR, vincristine; VP16, etoposide.

Figure 3

Flow cytometric assays showing induction of death, apoptosis, and cell cycle changes by single-agent obatoclax in MLL-AF4 ALL cell lines. (A) Surviving fraction plots of cell viability measured by MTT assays at 72 hours after obatoclax or ADR (positive control for apoptosis) exposure. Assays were performed at least 3 times (6 replicates/condition per experiment). Error bars indicate standard error. Results from MTT assays at 72 hours determined the low and high obatoclax (EC50; 3 × EC50) concentrations and the ADR concentration that were used in time-course flow cytometric assays (B-E). (B) Flow cytometric assays of cell death measured by PI fluorescence uptake in vehicle-, obatoclax-, or ADR-treated cell lines. Average percentage of dead (PI positive) cells from 5 independent experiments are plotted. (C) Flow cytometric assays of activated caspase 3 in vehicle-, obatoclax-, or ADR-treated cell lines. Activated caspase 3 positive cells were gated to determine percentages of apoptotic cells. Graphs represent average percentage of caspase 3 positive cells from 5 independent experiments. (D) FACS TUNEL assay of DNA fragmentation in vehicle-, obatoclax-, or ADR-treated cell lines. TUNEL-positive cells were gated to determine percentages. The average percentage of TUNEL positive cells from 4 independent experiments is plotted. Bars (B-D) show standard error. Asterisk indicates P ≤ .05 vs vehicle at individual points. (E) Representative cell cycle analyses in vehicle-, obatoclax-, or ADR-treated cell lines with cell cycle distribution determined using ModFit LT. (F) Bar graph plots of percentages of cells in G0/G1, S, and G2/M in vehicle-, obatoclax-, or ADR-treated SEM-K2 cells in 6 independent experiments.

Figure 4

Western blot analyses and MTT assays of obatoclax effects on apoptosis, autophagy, and necroptosis by itself or combined with cell death pathway inhibition in MLL-AF4 ALL cell lines. (A) Time-course Western blot analysis of PARP cleavage using whole-cell lysates from SEM-K2 and RS4:11 cells after exposure to vehicle, or obatoclax at concentrations approximating 72-hour EC50 or 3 × EC50, or ADR (positive control for apoptosis) for indicated times. Increases in cleaved PARP by 72 hours with obatoclax treatment indicate apoptosis. (B) Time-course Western blot analyses of LC3-I to LC3-II conversion and p62 protein levels after exposure to vehicle or obatoclax at concentrations approximating 72-hour EC10, EC25, EC50, or 3 × EC50 for indicated times. Note time- and dose-dependent increases in LC3-I to LC3-II conversion and lack of p62 accumulation with obatoclax treatment. (C) Unchanged obatoclax-induced cytotoxicity with genetic autophagy inhibition in SEM-K2 cells. Representative Western blot depicting BECN1 protein knock down by BECN1 siRNAs #1 and #2 during obatoclax treatment compared with cells transfected with control siRNA. Cell lysates were prepared after obatoclax treatment of transfected cells for indicated times (left). Surviving fraction plots of cell viability measured by MTT assays performed after 72-hour obatoclax exposure (6 replicates/condition per experiment) in cells transfected with control siRNA or BECN1 siRNA 2 (right). The assay was performed 3 times. Bars indicate standard error. Transfection with BECN1 siRNA #1 gave the same result (not shown). (D) Unchanged obatoclax-induced cytotoxicity by chemical autophagy inhibition in SEM-K2 (left) and RS4:11 (right) cells. Surviving fraction plots show cell viability measured by MTT assays after 72-hour exposure to increasing obatoclax concentrations with indicated fixed 3-MA concentrations. Assays were repeated 4-7 times (6 replicates/condition per experiment). Bars indicate standard error. (E) Attenuation of obatoclax-induced cell death in SEM-K2 (left) and RS4:11 (right) cells by chemical necroptosis inhibition ± apoptosis and/or autophagy inhibition. Surviving fraction plots show cell viability measured by MTT assays 72 hours after treatment with increasing obatoclax concentrations ± 50 μM Nec-1 ± 20 μM zVAD-fmk and/or 0.5 mM 3-MA. MTT assays were performed 3-6 times (6 replicates/condition per experiment). Bars show standard error. * P ≤ .05 vs obatoclax alone; # P ≤ .05 vs obatoclax + Nec-1; ▲, P ≤ .05 vs obatoclax + Nec-1 + zVAD-fmk. Data in (D) and (E) were normalized to cells only treated with the relevant inhibitors alone or combined (supplemental Figure 1A). (F) Western blot analysis of SEM-K2 cells demonstrating dose-dependent increases in cleaved PARP indicative of more apoptosis with obatoclax treatment + Nec-1 for 72 hours compared with obatoclax alone, and reduction in obatoclax-induced PARP cleavage as well as attenuation of Nec-1 effect on PARP cleavage by zVAD-fmk (top). Also note the increase in LC3-I to LC3-II conversion induced by obatoclax in the presence of Nec-1 ± zVAD-fmk compared with obatoclax alone (bottom). Obatoclax concentrations are approximate 72-hour EC50 (50 nM) or 3 × EC50 (150 nM). Western blots are representative of 3 independent experiments. (G) Representative Western blot analyses of LC-I to LC-II conversion in RS4:11 cells treated with obatoclax at approximate 72-hour EC50 (30 nM) or 3 × EC50 (90 nM) ± Nec-1. Note increased (F) or unchanged (G) LC-I to LC-II conversion with Nec-1, indicating that obatoclax-induced autophagy was necroptosis-independent.

Figure 5

Gene expression changes in SEM-K2 and RS4:11 cells and changes in cell morphology in SEM-K2 cells indicative of triple death mode killing by obatoclax. Cells were treated for 6 hours in duplicate with vehicle or obatoclax at approximate 72-hour EC50 or 3 × EC50 concentrations, and respective cDNAs were used for Affymetrix U133_Plus_2 microarray analysis. (A) Venn diagrams showing numbers of probesets up- or downregulated by obatoclax and their overlap between different obatoclax concentrations for each cell line. Significant changes were determined using differences between mean log2 expression levels in cells treated with obatoclax at EC50 or 3 × EC50 and mean log2 expression levels in vehicle-treated cells (analysis of variance, P ≤ .05 and >50% up-/downregulation in at least a single mean log2 expression level comparison of obatoclax at EC50 or 3 × EC50 vs vehicle in either cell line). For a complete list of significantly up- or downregulated probesets, see supplemental Table 1. (B) Mini heatmaps of autophagy- and necroptosis-related probesets showing upregulation (red) or downregulation (green) in expression at or near significance (P ≤ .05) with obatoclax treatment at EC50 or 3 × EC50 in either cell line. Expression levels for all 3 conditions (vehicle, obatoclax EC50, obatoclax 3 × EC50) and 2 biologic replicates per condition (6 samples) were drawn for each cell line; mean log2 expression levels of respective vehicle treatments were used as reference. (C) SEM-K2 cells treated with vehicle, obatoclax, or ADR (positive control for apoptosis) for indicated times were harvested, stained, and sectioned for electron microscopy. Indicated arrowheads mark condensed chromatin and/or nuclear fragmentation of apoptosis, autophagic structures, or swollen Golgi (obatoclax; 24 hours) or endoplasmic reticulum (obatoclax; 48 hours, 72 hours) of necroptosis. Magnified insets are in boxes. Note morphologic changes of all 3 modes of death in single cell (obatoclax, 48 hours; right middle panel).

Figure 6

Evidence for triple killing mechanism of obatoclax in primary infant ALL cells from flow cytometry, Western blot analyses, and MTT assays with cell death inhibitors. Molecular cytogenetic subtype and 72-hour obatoclax EC50s determined from surviving fraction plots in Figure 1 for each case are shown. In cases in (A-D), all 3 types of assays were performed; in 2 cases in (E), apoptosis and autophagy proteins were studied by Western blot. In the other MLL-R case in (C), AF4, AF9, and ENL partner genes were excluded.(A-D; left) show contour plots of FSC vs SSC in cells treated for indicated times at respective 72-hour obatoclax EC75 determined from Figure 1. Progressive decrease in FSC signal in all 4 cases indicates apoptosis. (A-D; middle) and (E) are Western blot analyses of apoptosis and autophagy. Note increases in cleaved PARP in all 6 cases in response to obatoclax consistent with apoptosis. Also note increases in LC3-II in all 6 cases in response to obatoclax (A-E); this occurs without increase in p62 in 4 cases in (A, D, E), indicating autophagy with autophagic flux. In MLL-AF4 case in (B), note increase in p62 at 24 hours but decrease by 48 hours after obatoclax treatment, suggesting that autophagy is not blocked. Also note increase in p62 at 48 hours in the other MLL-R case (C, middle panel), indicating either p62 induction or accumulation, the former of which is supported by gene expression profiling data in SEM-K2 (supplemental Table 2). (A-D; right) Surviving fraction plots for MTT assays 72 hours after obatoclax treatment at increasing concentrations by itself or with 50 μM Nec-1, 20 μM zVAD-fmk, or 0.5 mM 3-MA alone or altogether. Inhibitors were used at concentrations determined minimally toxic in cell lines as well as primary cases (supplemental Figure 1), rather than at case-by-case titrated concentrations, and were tested less extensively in obatoclax combinations. Data were normalized to respective primary samples only treated with relevant inhibitors alone or with each other (supplemental Figure 1B). Note inhibition of obatoclax-induced death by Nec-1 in both MLL-AF4 cases (A, B; right) and reduced obatoclax-induced death when all 3 pathways were inhibited (A; right).

Figure 7

Electron microscopy evidence of apoptosis, autophagy, and necroptosis in obatoclax-treated primary infant ALL cells. Cases are the same as in Figure 6A-D. Molecular cytogenetic subtype and 72-hour obatoclax EC50s from surviving fraction plots in Figure 1 are at left. Cells were treated for 48 hours with vehicle, obatoclax at respective 72-hour obatoclax EC75 or 50 nM ADR, and harvested, stained, and sectioned for electron microscopy. Different arrowheads mark condensed chromatin of apoptosis, autophagic structures, or swollen organelles of necroptosis. Magnified insets are in boxes. Note morphologic findings indicating all 3 modes of death, including examples in single cells in all cases. Also note chromatin condensation as well as less prominent autophagic structures in ADR control cells in all cases.