Selective BCL-2 inhibition by ABT-199 causes on-target cell death in acute myeloid leukemia

Abstract

B-cell leukemia/lymphoma 2 (BCL-2) prevents commitment to programmed cell death at the mitochondrion. It remains a challenge to identify those tumors that are best treated by inhibition of BCL-2. Here, we demonstrate that acute myeloid leukemia (AML) cell lines, primary patient samples, and murine primary xenografts are very sensitive to treatment with the selective BCL-2 antagonist ABT-199. In primary patient cells, the median IC50 was approximately 10 nmol/L, and cell death occurred within 2 hours. Our ex vivo sensitivity results compare favorably with those observed for chronic lymphocytic leukemia, a disease for which ABT-199 has demonstrated consistent activity in clinical trials. Moreover, mitochondrial studies using BH3 profiling demonstrate activity at the mitochondrion that correlates well with cytotoxicity, supporting an on-target mitochondrial mechanism of action. Our protein and BH3 profiling studies provide promising tools that can be tested as predictive biomarkers in any clinical trial of ABT-199 in AML.

Conflict of interest statement

Conflict of Interest

A. Letai is an advisor for AbbVie and M. Konopleva has a sponsored research agreement from AbbVie. Dr. Andreeff serves on the Scientific Advisory Board of Eutropics Pharmaceuticals which once had a license for BH3 profiling.

Figures

Figure 1. Selective inhibition of BCL-2 by ABT-199 kills AML cell lines quickly and effectively

A). AML cell lines were treated with ABT-199 or ABT-737 for 48 h. Calcusyn software was used to calculate the IC50 values based on the number of viable cells (i.e., Annexin V−/PI−) determined by FACS analysis. B). MOLM-13 AML cells were treated with indicated concentrations of ABT-199. Apoptosis induction was determined by Annexin V/PI flow cytometry. C). Viable (i.e., Annexin V−/PI−) cell counts were quantified by FACS analysis using CountBright counting beads. D). Serial bioluminescence images of mice bearing MOLM-13 tumors treated with the vehicle or ABT-199 (treatment started on day 4, administered by oral gavage at dose of 100 mg/kg). E). Kaplan-Meier survival curves for mice treated as described in E (n = 7 per arm). Statistical significance was calculated using Log-rank (Mantel-Cox) test (p < 0.0004 ). F). H&E staining of histological sections of liver, spleen, and bone marrow 15 d post leukemia cell injection. Age- and sex-matched mice without tumor were used as controls. Representative MOLM-13 cells are indicated by arrows. Representative engraftment areas are circled in green. All pictures were taken under the same magnification; scale bar equals 50 μm. G). Immunohistochemical staining of histological sections of liver, spleen, and bone marrow with human CD45 antibody 15 d post leukemia cell injection. Scale bar equals 50 μm.

Figure 2. Sensitivity to ABT-199 positively correlates with endogenous BCL-2 protein level and negatively correlates with BCL-XL protein level in AML cell lines

A) Western blot analysis of BCL-2 family proteins in untreated AML cells. The band intensity was quantified using Odyssey v2.0 software, and displayed numerically as a ratio of the band intensity detected in the OCI-AML3 cells. B) Significant correlations were observed between ABT-199 IC50 values and BCL-2/BCL-XL protein levels. The non-parametric one-tailed Spearman test was used to determine the correlation coefficient. The p values provided are nominal p values not corrected for multiple comparisons. C). MCL-1 knockdown by 85% was achieved by lentiviral shRNA. D). MCL-1 knockdown significantly sensitized OCI-AML3 cells to ABT-199. E). Western blot analysis showing HL-60 AML cells transfected to stably overexpress BCL-XL or BCL-2. F). Overexpression of BCL-XL or BCL-2 in HL-60 cells confers complete resistance to ABT-199-induced apoptosis.

Figure 3. ABT-199 functions selectively on BCL-2 dependent mitochondria in AML cell lines

A). The IC50 values of AML cell lines treated with ABT-737 from Figure 1A were correlated with the mitochondrial response of ABT-737 (1μM). Mitochondrial response was measured by JC1 based BH3 profiling. B). IC50 values of cell lines treated with ABT-199 from Figure 1A were correlated with the mitochondrial response of mitochondrial ABT-199 (0.1μM)C). IC50 values of AML cells treated with ABT-737 were correlated with the response to the BAD BH3 (80μM). The mitochondrial responses to the BAD BH3 peptide were measured by JC1 based BH3 profiling. D). IC50 values of AML cell treated with ABT-199 from Figure 1A were correlated with the mitochondrial response of the BAD BH3 (80uM) peptide. E). IC50 values of AML cells treated with ABT-737 were correlated with the response to the BAD BH3 (80μM) – HRK BH3 (80μM). The mitochondrial responses to the BAD and HRK BH3 peptides were measured by JC1 based BH3 profiling. F). IC50 values of AML cells treated with ABT-199 from Figure 1A were correlated with the mitochondrial response of the BAD BH3 (80uM) – HRK BH3 (80uM) peptide. Statistical correlation was performed using a one-tailed Spearman r using GraphPad Prism 6.

Figure 4. ABT-199 efficiently kills primary AML myeloblasts as a single agent

A). IC50 determination for ABT-199 and ABT-737 treatment of primary AML samples. Fresh mononuclear cells from AML patients were isolated from bone marrow or peripheral blood and treated with ABT-199 and ABT-737 for 48 h. The IC50 values were calculated based on viable (i.e., Annexin V−/PI−) cell numbers determined by FACS analysis. Samples with ABT-199 IC50 < 0.1 μM were defined as “sensitive”, while those with ABT-199 IC50 > 1 μM were defined as “resistant”. B). Frozen primary AML myeloblasts were thawed treated with ABT-199 and ABT-263 for 8 h in the absence of fetal bovine serum. Viability was assessed by Annexin-/PI- via FACS analysis and IC50 values were calculated using GraphPad Prism software. C). Thawed primary AML samples were treated for 2 h with 1–1000 nm of ABT-199 and viability was assessed by Annexin V-PI- by FACS analysis D). Nonparametric Spearman correlation analysis shows a significant (p = 0.017) negative correlation between ABT-199 IC50 values and BCL-2 protein levels. E) A non-significant (p = 0.069) positive correlation was observed between ABT-199 IC50 values and BCL-XL protein levels. F). Boxplots represent the quartiles and range of log2 values of mRNA expression for BCL-2 genes in different subgroups of AML and normal bone marrows. The median is indicated by the black line in each box. Numbers on top indicate number of patients in each specified subgroup. Differences in gene expression with P values ≤ 0.005 were considered statistically significant, as denoted by *. G). Patient AML samples treated with 100 nM ABT-199 for 24 hours were subjected to FACS analysis of specific apoptosis based on Annexin V staining in the bulk AML myleoblast and CD34+/CD38−/CD123+ LSC-containing population. P value determined via paired t-test.

Figure 5. BH3 profiling predicts AML myeloblast killing by ABT-199

A). Intracellular BH3 (iBH3) profiling was performed on thawed primary AML cells using the BAD BH3 (80 μM) and ABT-199 (1μM). The mitochondrial sensitivity to BAD BH3 and ABT-199 were positively correlated. B). There is no correlation between the IC50 of primary AML samples from Figure 4B with the BCL-XL specific BH3 peptide HRK (80 μM). C). The IC50 of primary AML samples from Figure 4B were correlated with the NOXA (80μM), a MCL-1 specific NOXA BH3 peptide. D). The ABT-199 IC50 of primary AML samples from Figure 4B were correlated with the BAD BH3 peptide (80uM). E). The ABT-199 IC50 from Figure 4B was correlated with the ABT-199 mitochondrial response (1μM). All correlations were tested using a one-tailed Spearman r correlation using GraphPad Prism software.

Figure 6. BH3 profiling predicts AML progression in a primary AML xenograft model

A–B). NSG mice were injected with primary AML cells as described under Methods. Mice were treated with ABT-199 100 mg/kg oral daily dose starting 3 weeks after AML cell injection, for two weeks. The graph represents % of human CD45+ leukemic cells in the murine bone marrow in mice sacrificed upon completion of the therapy. A non-parametric, unpaired, two-tailed t-test was used to evaluate the significance of mean difference. C). Intracellular BH3 profiling using the BAD BH3 (80 μM) and ABT-199 (10 μM) was performed on pre-treatment patient samples.