MCL-1-independent mechanisms of synergy between dual PI3K/mTOR and BCL-2 inhibition in diffuse large B cell lymphoma

Abstract

The PI3K/AKT/mTOR axis promotes survival and is a frequently mutated pathway in cancer. Yet, inhibitors targeting this pathway are insufficient to induce cancer cell death as single agents in some contexts, including diffuse large B cell lymphoma (DLBCL). In these situations, combinations with inhibitors targeting BCL-2 survival proteins (ABT-199 and ABT-263) may hold potential. Indeed, studies have demonstrated marked synergy in contexts where PI3K/mTOR inhibitors suppress expression of the pro-survival protein, MCL-1. In this study, we use BH3 profiling to confirm that BCL-2 and BCL-XL support survival following PI3K pathway inhibition, and that the dual PI3K/mTOR inhibitor BEZ235 strongly synergizes with BCL-2 antagonists in DLBCL. However, we identify an alternative mechanism of synergy between PI3K/mTOR and BCL-2 inhibitors, independent of MCL-1 down-regulation. Instead, we show that suppression of AKT activation by BEZ235 can induce the mitochondrial accumulation of pro-apoptotic BAD and BIM, and that expression of a constitutively active form of AKT prevents sensitization to BCL-2 antagonism. Thus, our work identifies an additional mechanism of synergy between PI3K pathway inhibitors and BCL-2 antagonists that strengthens the rationale for testing this combination in DLBCL.

Keywords: BCL-2; PI3K; apoptosis; lymphoma; mTOR.

Figure 1. PI3K pathway inhibition increases mitochondrial priming and enhances efficacy of ABT-263 in DLBCL cell lines

A., D. OCI-LY1 and SU-DHL4 cells were treated with PI3K pathway inhibitors for 16 hours prior to permeabilization and treatment with BIM peptide (0.3μM) for 60 minutes. Mitochondrial depolarization was quantified by loss of JC-1 aggregate fluorescence; data are normalized to DMSO treated cells (n = 3). B., E. OCI-LY1 and SU-DHL4 cells were treated with ABT-263 with or without PI3K pathway inhibitors for 48 hours. Viability was assessed using 7-AAD dye exclusion (n = 3). C., F. Correlation between ABT-263 sensitivity (IC₅₀) and MOMP induced by BIM peptide. IC₅₀ is the average of three independent ABT-263 titrations in OCI-LY1 and SU-DHL4 cells treated with half-log dilutions either with or without indicated PI3K pathway inhibitor; viability was assessed by 7-AAD dye exclusion. Correlation was calculated using Spearman r and is shown above with one-tailed P value. G. ABT-263 sensitivity of four DLBCL cell lines with or without co-treatment with BEZ235. IC₅₀ was obtained as described above (n = 3). H. Cells were treated with combinations of ABT-263 with BEZ235 with or without Q-VD-OPh (pan-caspase inhibitor) prior to assessing viability by 7-AAD dye exclusion (n = 3). All data are shown as mean ± SD. Significance was calculated using a paired one-tailed student's t test and is relative to untreated control unless otherwise indicated. *P < 0.05, **P < 0.005, ***P < 0.001.

Figure 2. BEZ235 does not enhance the toxicity of BH3 mimetics in normal human T cells

A. OCI-LY1 and SU-DHL4 cells were treated with ABT-263 (50 or 300 nM, respectively) or ABT-199 (5 or 50 nM, respectively) with or without BEZ235 or GDC-0980. Viability was measured using 7-AAD dye exclusion after 48 hours (n = 3). B. PBMCs were isolated from normal human blood donors and were treated with ABT-263 (30 nM) or ABT-199 (3 nM) with or without BEZ235 for 48 hours (n = 4). Cells were stained for CD4 and CD19 prior to assessing viability of lymphocyte subtypes by 7-AAD dye exclusion. Cells in the CD4⁻ CD19⁻ gate are mostly CD8⁺ T cells and natural killer cells. All data are shown as mean ± SD. Significance was calculated using a paired two-tailed student's t test. *P < 0.05, **P < 0.005, ***P < 0.001.

Figure 3. DLBCL cells over-expressing BCL-2 are resistant to a chemotherapeutic drug but remain sensitive to BEZ235 and ABT-199

A. Cells were treated with the indicated doxycycline doses for 24 hours. Densitometry values were normalized to GAPDH loading control, then normalized to untreated empty vector cells. Data are representative of three independent experiments. B. OCI-LY1 and SU-DHL4 cells expressing either empty vector or BCL-2 were treated with vincristine for 48 hours prior to assessing viability by 7-AAD dye exclusion. Cells were pre-treated with doxycycline (25 ng/ml or 1 μg/ml, respectively) for 24 hours to induce ectopic expression of BCL-2 (n = 3). C. OCI-LY1 and SU-DHL4 cells were pre-treated with doxycycline (25 ng/ml or 1 μg/ml, respectively) for 24 hours prior to treatment with ABT-199 (100 nM) with or without BEZ235 for 48 hours. Viability was assessed by 7-AAD dye exclusion (n = 3). All data are shown as mean ± SD. Significance was calculated using a paired one-tailed student's t test. *P < 0.05, **P < 0.005, ***P < 0.001.

Figure 4. BEZ235 does not affect MCL-1 expression in OCI-LY1 cells

A. OCI-LY1 (DLBCL), BV173 (B-cell acute lymphoblastic leukemia), and matched parental and ABT-199-resistant SU-DHL6 cells were treated with BEZ235 or PIK-75 for increasing time. Data are representative of three independent experiments. Western blots were probed with the antibodies indicated on the left. B. Cells stably transduced with shMCL-1 or shScramble control were treated with half-log dilutions of ABT-263 with or without BEZ235 to determine IC₅₀ using GraphPad Prism software (5.0c). Data represent replicates of independent knockdown populations for each hairpin (n = 4). C. Cells stably transduced with empty vector or a doxycycline-inducible MCL-1 expression vector were pre-treated with doxycycline (1 μg/ml) for 24 hours prior to determining IC₅₀ as noted above (n = 3). All data are shown as mean ± SD. Unless otherwise specified, significance was calculated using a paired one-tailed student's t test and is relative to untreated control. *P < 0.05, **P < 0.005, ***P < 0.001.

Figure 5. PI3K pathway inhibitors increase mitochondrial abundance of BAD and BIM

A. Immunoblot of mitochondrial and cytoplasmic fractions of OCI-LY1 cells treated with indicated PI3K pathway inhibitors for 16 hours. Densitometry values were normalized to COX IV (mitochondrial) or ERK (cytoplasmic) loading controls, this ratio was then normalized to untreated cells. Data are representative of three independent experiments. B. Average densitometry values of three replicates (n = 3) of panel A.. All data are shown as mean ± SD. Significance was determined using a two-tailed one-sample t test relative to normalized control. *P < 0.05, **P < 0.005. C. Immunoblot of immunoprecipitation of BCL-2 (upper) or whole cell lysates (lower) following 16 hour treatment with BEZ235, ABT-199 (100 nM), or the combination. Data are representative of three independent experiments. All cells were also treated with 10 μM Q-VD-OPh to prevent cleavage of BCL-2 family proteins by caspases.

Figure 6. AKT suppression is a critical component of synergy between BEZ235 and ABT-199

A. Immunoblot of OCI-LY1 cells expressing either empty vector or phospho-mimetic AKT (S473D). Cells were pre-treated with doxycycline (1 μg/ml) for 24 hours prior to treatment with indicated PI3K pathway inhibitors for an additional 3 hours. Data are representative of three independent experiments. B. Immunoblot of mitochondrial (M) and cytoplasmic C. fractions from OCI-LY1 cells expressing empty vector or phospho-mimetic AKT (S473D). Densitometry values were normalized to COX IV (mitochondrial) or ERK (cytoplasmic) loading controls, this ratio was then normalized to untreated cells. Data are representative of three independent experiments. C. Sensitivity of three DLBCL cell lines expressing AKT S473D to ABT-199 in the presence or absence of MK2206, BEZ235, or GDC-0980. Cells were pre-treated with doxycycline (1 μg/ml) for 24 hours prior to treatment with increasing concentrations of ABT-199 with or without BEZ235 (50 nM) for 48 hours. Viability was assessed by 7-AAD dye exclusion and IC₅₀ values were calculated using GraphPad Prism (5.0c) software (n = 3). All data are shown as mean ± SD. Significance was calculated using one-tailed student's t test *P < 0.05, **P < 0.005, ***P < 0.001.

Figure 7. Expression of exogenous murine Bad sensitizes OCI-LY1 cells to AKT inhibition

A. Immunoblot of mouse BAD (mBAD) induction following 24 hour treatment with indicated doses of doxycycline. Cells were also treated with 10 μM Q-VD-OPh to prevent caspase cleavage of BAD. # Indicates murine isoform, ## indicates human isoform. Data are representative of three independent experiments. B., C. OCI-LY1 cells transduced with empty vector, mBAD wild-type (WT), and phospho-null mBAD (S136A) were treated with increasing concentrations of doxycycline ± BEZ235 B. or MK2206 C. for 48 hours. Viability was assessed using 7-AAD dye exclusion (n = 3). All data are shown as mean ± SD. Significance was calculated using a paired one-tailed student's t test ***P < 0.001.

Figure 8. Model of synergy between BEZ235 and ABT-263 in DLBCL cell lines

Treatment with BEZ235 in DLBCL cell lines completely inhibits signaling through the PI3K and downstream effectors, AKT and mTORC1. Suppression of mTORC1 may reduce cap-dependent translation of pro-survival proteins (other than MCL-1) downstream of 4E-BPs. Loss of AKT activity promotes FOXO-mediated transcription of BIM and facilitates mitochondrial accumulation of dephosphorylated BAD. When combined with ABT-263 or ABT-199, the combination promotes MOMP through BIM-activated BAX and BAK oligomerization, leading to induction of apoptosis.