Characterization of midostaurin as a dual inhibitor of FLT3 and SYK and potentiation of FLT3 inhibition against FLT3-ITD-driven leukemia harboring activated SYK kinase

Oncogenic FLT3 kinase is a clinically validated target in acute myeloid leukemia (AML), and both multi-targeted and selective FLT3 inhibitors have been developed. Spleen tyrosine kinase (SYK) has been shown to be activated and increased in FLT3-ITD-positive AML patients, and has further been shown to be critical for transformation and maintenance of the leukemic clone in these patients. Further, over-expression of constitutively activated SYK causes resistance to highly selective FLT3 tyrosine kinase inhibitors (TKI). Up to now, the activity of the multi-targeted FLT3 inhibitor, midostaurin, against cells expressing activated SYK has not been explored in the context of leukemia, although SYK has been identified as a target of midostaurin in systemic mastocytosis. We compared the ability of midostaurin to inhibit activated SYK in mutant FLT3-positive AML cells with that of inhibitors displaying dual SYK/FLT3 inhibition, targeted SYK inhibition, and targeted FLT3 inhibition. Our findings suggest that dual FLT3/SYK inhibitors and FLT3-targeted drugs potently kill oncogenic FLT3-transformed cells, while SYK-targeted small molecule inhibition displays minimal activity. However, midostaurin and other dual FLT3/SYK inhibitors display superior anti-proliferative activity when compared to targeted FLT3 inhibitors, such as crenolanib and quizartinib, against cells co-expressing FLT3-ITD and constitutively activated SYK-TEL. Interestingly, additional SYK suppression potentiated the effects of dual FLT3/SYK inhibitors and targeted FLT3 inhibitors against FLT3-ITD-driven leukemia, both in the absence and presence of activated SYK. Taken together, our findings have important implications for the design of drug combination studies in mutant FLT3-positive patients and for the design of future generations of FLT3 inhibitors.


INTRODUCTION
Around 30% of patients with acute myeloid leukemia (AML) harbor activating mutations in FLT3 [1], a gene normally involved in regulating hematopoiesis. The most common type of FLT3 mutation results in internal tandem duplications (ITD) within the juxtamembrane domain, occurring in 20-25% of AML and strongly associated with decreased survival [2,3]. An additional 7% of patients have point mutations within the "activation loop" of FLT3 [4].
Numerous FLT3 kinase inhibitors, both multitargeted and selective, have been developed [5]. The N-indolocarbazole, midostaurin (PKC412; N-benzoylstaurosporine; Novartis Pharma AG) was shown to target oncogenic FLT3 in preclinical studies [6] was reported to significantly prolong survival of FLT3-mutated AML patients when combined with conventional induction and consolidation therapies in a randomized Phase III clinical trial [7]. Midostaurin was recently FDA approved for treatment of adult, newly diagnosed AML patients positive for oncogenic FLT3, in combination with standard chemotherapy. Other FLT3 inhibitors in clinical development include quizartinib (AC220), which exhibits high potency and selectivity against FLT3-ITD [8], and crenolanib besylate (CP-868596; AROG Pharmaceuticals, LLC) [9].
Spleen tyrosine kinase (SYK) is an obligatory signaling partner for FLT3 that is required for transformation to AML and necessary for myeloproliferative disease (MPD) development [10]. SYK (wild-type (wt)) is expressed in most hematopoietic cells [11,12] and belongs to the SYK/ZAP-70 family of non-receptor tyrosine kinases [13,14]. Oncogenic SYK has been identified as an important driver of different hematologic malignancies, including B-cell lymphoma, chronic lymphocytic leukemia (CLL) and mantle cell lymphoma [15][16][17], and was identified as a target in AML with SYK inhibition exhibiting anti-leukemia activity in mouse models of AML [18]. ITK-SYK, which results from the fusion between SYK and ITK (IL-2-inducible T-cell kinase), occurs as a recurrent translocation in 17% of patients with unspecified peripheral T-cell lymphomas [19,20]. TEL-SYK was originally detected in a patient with an atypical myelodysplastic syndrome with leukemic transformation [21,22]. TEL-SYK over-expression in murine pre B cells causes a B-acute lymphocytic leukemia (ALL)-like disease in mice [23] and cytoplasmic TEL-SYK fusion induces an acute panmyelosis with myelofibrosis-type acute myeloid leukemia (AML) in a bone marrow transplantation model [24].
Importantly, highly activated SYK has been found to be enriched in AML patients with a higher frequency in patients harboring the ITD mutant than wild-type (wt) FLT3 [10]. SYK has also been shown to be associated with resistance to FLT3-ITD-targeted therapy, and inhibition of FLT3-ITD with quizartinib combined with inhibition of SYK was shown to be more effective than FLT3 inhibition alone in FLT3-ITD-positive models of AML [10].
Given the reported significance of SYK in transformation and maintenance of AML, as well as FLT3 kinase inhibitor resistance, and implications of SYK as a potentially important target for AML treatment, we were interested in investigating the ability of midostaurin to inhibit SYK in FLT3 mutant-positive AML. SYK has previously been shown to be a target of midostaurin and its metabolites, CGP52421 and CGP62221, in advanced systemic mastocytosis, a hematopoietic neoplasm characterized by expansion and abnormal accumulation of mast cells [25]. Up to now, however, the activity of midostaurin, alone and combined with SYK inhibition, against cells expressing activated SYK has not been explored in the context of leukemia.
Here, we compare the SYK-targeting activity of midostaurin with different classes of kinase inhibitors, including those with SYK inhibitory activity and those without. We show that midostaurin, like the dual FLT3/SYK inhibitors, R406 (tamatinib), and R788 (fostamatinib) [26], inhibits SYK in cell-based models of FLT3-ITD-and activated SYK-driven leukemia to a greater extent than highly targeted inhibitors of FLT3. We also show that SYK inhibition alone is not sufficient to kill FLT3-ITD-positive cells and that the SYK inhibitory activity of midostaurin is insufficient to potently kill cells expressing activated SYK. However, additional targeted SYK inhibition or dual FLT3/SYK inhibition potentiates the effects of midostaurin and other inhibitors of FLT3 against both kinase inhibitor-sensitive-and -resistant FLT3-ITD-and activated SYK-positive leukemia.

SYK is a target of midostaurin
As a first assessment of the ability of midostaurin to inhibit SYK we tested its inhibitory activity of SYK in a purified enzyme assay and found the compound inhibits SYK with an IC 50 of 20.8 nM. In similar enzyme assays, the SYK inhibitor, R788, and its active metabolite, R406, have been determined to inhibit SYK with an IC 50 of 41 nM [26]. Both compounds are reported to be 5-fold less potent against FLT3 than Syk. Both R788 and R406 have been evaluated in clinical trials. In comparison, a highly targeted SYK inhibitor, PRT062607 (P505, BIIB057) has been reported in cell-free assays to inhibit SYK with an IC 50 of 1 nM and to inhibit FLT3 with an IC 50 of 139 nM [27].
Oncotarget 52028 www.impactjournals.com/oncotarget Effects of midostaurin, R406, and R788 on Ba/F3 cells expressing activated SYK In order to investigate the ability of midostaurin to inhibit SYK activity in cells, we utilized Ba/F3 cell lines that stably express constructs encoding fusion proteins comprised of SYK kinase coupled with a TEL moiety that constitutively activates SYK [10] (Supplementary Figure  1). Specifically, one Ba/F3 cell line was developed to overexpress constitutively active "TEL-SYK," the protein product of a fusion between a truncated form of SYK that is devoid of its SH2 domains, SH2 Nter + SH2 Cter, and a TEL sequence that takes the place of these two domains [22,10]. A second Ba/F3 cell line tested was engineered to over-express a constitutively and highly active variant of SYK, "SYK-TEL," which, unlike TEL-SYK, contains an intact SH2 domain (Supplementary Figure 1). Consistent with the order of potencies generated in cell-free assays, midostaurin inhibited the growth of Ba/F3-TEL-SYK cells with an IC 50 of 101.2 nM, a 2-to 3-fold higher potency than that of R406 (IC 50 = 196.8 nM) and R788 (IC 50 = 332.9 nM), however a lower potency than the IC 50 of 43.8 nM for PRT062607 against these cells (Figure 1A-1D and Table 1A). The effects of midostaurin, R406, and R788 against Ba/F3-TEL-SYK cells were modestly IL-3 rescued; the effects of PRT062607 were highly IL-3 rescue-able.
Similar results were observed with midostaurin treatment of Ba/F3-SYK-TEL cells, with a higher potency exhibited by midostaurin (IC 50 = 108.0 nM) as compared to R406 (IC 50 = 142.4 nM) and R788 (IC 50 = 237.7 nM), yet a lower potency as compared to PRT062607 (IC 50 = 30.0 nM) ( Figure 1E-1H and Table 1A). Midostaurin was observed to induce apoptosis, as measured by annexin/ PI staining, in cells expressing constitutively activated Oncotarget 52029 www.impactjournals.com/oncotarget SYK ( Figure 1I-1K). Consistent with this, midostaurin also more robustly increased the activities of caspase-3 and -7 in activated SYK-expressing cells as compared to growth factor-dependent Ba/F3 cells over-expressing wt FLT3 ( Figure 1L) and led to an increase in the subG0/ G1 fraction as determined by PI staining (Supplementary Figure 2).
For comparison with Ba/F3 cells expressing FLT3-ITD, we tested the effects of dual SYK/FLT3 and targeted SYK inhibitors against cells expressing FLT3-ITD and kinase inhibitor-resistant tyrosine kinase domain (TKD) point mutations. As expected, midostaurin, R406, and R788 treatment killed Ba/F3 cells expressing FLT3-ITD+FLT3 kinase inhibitor-resistant TKD point mutants to varying extents although each generally with less potency than toward FLT3-ITD, whereas PRT062607 was completely inactive against these cells ( Figure 3E-3H).
We next explored the ability of midostaurin, R406, and PRT062607 to inhibit the phosphorylation of FLT3 or SYK in Ba/F3 cells transformed by activated versions of each. Midostaurin and R406 were effective at inhibiting FLT3 phosphorylation in a concentration-dependent manner in Ba/F3-FLT3-ITD cells, which correlated with induction of cell death by these compounds (Figure 3I-3J). In Ba/F3-TEL-SYK cells, midostaurin, similar to PRT062607, inhibited TEL-SYK phosphorylation and activity of downstream effectors of TEL-SYK across a concentration range that led to inhibition of cell growth by the compounds ( Figure 4A). In Ba/F3-FLT3-ITD+TEL-  Oncotarget 52030 www.impactjournals.com/oncotarget SYK cells, both midostaurin and R406 inhibited phosphorylation of TEL-SYK, although to a lesser extent than PRT062607 ( Figure 4B-4D). Despite the observed inhibitory activity of PRT062607 against TEL-SYK, the fact that PRT062607 is a comparatively weak inhibitor of Ba/F3-FLT3-ITD+TEL-SYK proliferation again supports the notion that Ba/F3-FLT3-ITD+TEL-SYK cells are predominantly driven by FLT3-ITD.
It is important to note that concentrations of R406, and R788 were relatively ineffective against parental Ba/ F3 cells at up to 500-1000 nM following three days of drug treatment, however concentrations of midostaurin over 250 nM led to anti-proliferative effects against parental Ba/F3 cells suggesting that nonspecific toxicity could occur at those levels of drug ( Figure 5). In contrast, PRT062607 was ineffective against parental Ba/F3 cells at up to 1000 nM (Supplementary Figure 4).

Potentiation of the anti-leukemic effects of midostaurin by dual FLT3/SYK suppression and targeted SYK suppression
The ability of midostaurin to synergize with dual FLT3/SYK inhibitors, R406 and R788, or the highly targeted SYK inhibitor, PRT062607, was tested in a panel of Ba/F3 cell lines expressing constitutively active SYK in the presence or absence of FLT3-ITD, as well as Ba/F3-FLT3-ITD cells and human AML lines, such as MOLM13-luc+, MOLM14, and MV4,11, driven by FLT3-ITD. As described in detail below, general observations from these experiments include potentiation of the anti-proliferative effects of midostaurin by inhibitors targeting SYK, which supports the notion that additional SYK suppression is of benefit against FLT3-ITD-driven AML treated with a FLT3 kinase inhibitor.
We sought to confirm that potentiation of midostaurin by dual FLT3/SYK inhibition and targeted SYK inhibition could be advantageous in terms of overriding some forms of drug resistance. Midostaurinresistant MOLM13-luc+ cells were previously established via prolonged exposure to midostaurin, and exhibited increased levels of cell surface FLT3 protein and elevated levels of phospho-MAPK, however normal levels of phospho-STAT5 [28]. Midostaurin-resistant MOLM13-luc+ cells were tested with combinations of midostaurin and R406, R788, or PRT062607, and nearly additive effects were observed between midostaurin +/-R406 or R788 against the drug-resistant cells across a range of concentrations ( Figure 10A-10B and Table 2). Positive combination effects between midostaurin and PRT062607 against these cells were more striking, with combination indices indicative of synergy ( Figure 10C and Table  2). Midostaurin-resistant MOLM13-luc+ cells are less responsive to midostaurin than parental MOLM13-luc+ as evidenced by a rightward shift in the midostaurin doseresponse curve ( Figure 10D). Drug resistance is likely due to the over-expression of FLT3 protein in the midostaurinresistant cells, as previously reported and also confirmed in the present study ( Figure 10E). SYK protein expression, Oncotarget 52035 www.impactjournals.com/oncotarget however, remained the same between the two lines ( Figure  10E). We next evaluated the therapeutic potential of combining midostaurin with dual SYK/FLT3 or SYK inhibitors by testing the resulting growth inhibitory effect on FLT3-ITD-expressing AML primagraft cells from relapsed patients or patients that were refractory to chemotherapy treatment. As normal controls, PBMC cells from a healthy donor were tested with the combination of midostaurin+R406, and little to no inhibitory effects of the agents alone or combined were observed ( Figure 11A). MOLM14 cells were tested in parallel with the normal PBMC cells as a positive control for drug activity ( Figure  11B). We then tested the ability of R406 to potentiate midostaurin ( Figure 11C). Whereas little to no single agent activity was observed for midostaurin at up to 20 nM or R406 at up to 100 nM against AML primagraft sample #1 (disease stage at time of sample acquisition: relapsed post-allogeneic HSCT, Supplementary Table 1), the combination of midostaurin and R406 led to a higher percentage of cell death as compared to either single agent across the full range of concentrations ( Figure 11C). A second AML primagraft sample (#2) (disease stage at time of sample acquisition: relapsed after multiple modalities, Supplementary Table 2) was tested with midostaurin+/-PRT062607 at concentrations previously tested against FLT3-ITD-expressing cell lines ( Figure 11D). Whereas little to no single agent activity was observed for midostaurin at up to 40 nM, PRT062607 treatment led to around 50% cell death at 1000 nM, and a slight leftward shift in the dose-response curve (indicating a positive drug combination effect) was observed for the  Figure 11D). Interestingly, treatment of a third AML primagraft sample (#3) (disease stage at time of sample acquisition: primary refractory postinduction, Supplementary Table 3) with midostaurin (up to 40 nM), alone and in combination with PRT062607 across a comparatively conservative concentration range (up to only 400 nM, a 10-fold lower concentration than was used against FLT3-ITD-positive cell lines), resulted in a higher percentage in the combination of agents as compared to either agent alone ( Figure 11E). SYK protein expression was detectable in all FLT3-ITD-positive AML primagrafts tested; SYK expression in bone marrow and spleen samples corresponding to the primagrafts is shown alongside normal PBMC cells ( Figure 11F).

Potentiation of the effects of crenolanib by targeted SYK suppression against cells expressing FLT3-ITD and activated SYK
Given that dual SYK/FLT3 inhibition and targeted SYK inhibition potentiated the effects of midostaurin, an established dual inhibitor of SYK and FLT3, against cells driven by FLT3-ITD, activated SYK, or a combination of both, we were interested in investigating the ability of additional SYK inhibition to similarly augment the effectiveness of targeted FLT3 inhibition. The combination of crenolanib with PRT062607 was nearly additive against Ba/F3-FLT3-ITD cells for the concentrations of inhibitors at which 25% and 50% of cell growth inhibition is observed (ED25-ED50) ( Figure 12A and Table 2). PRT062607 potentiated the anti-leukemic Oncotarget 52037 www.impactjournals.com/oncotarget effects of crenolanib against cells co-expressing FLT3-ITD+SYK-TEL, with nearly additive effects observed for concentrations of inhibitors at which 25%, 50%, 75%, and 90% of cell growth inhibition is observed (ED25-ED90) ( Figure 12B and Table 2).

DISCUSSION
SYK has been shown to be important in FLT3 mutant-positive AML in that it transactivates FLT3-ITD, and through over-expression has been shown to play a role in AML transformation and resistance to FLT3 inhibition [10]. In addition, the combined inhibition of FLT3 by a targeted FLT3 inhibitor (quizartinib) and SYK has been demonstrated to be more effective than FLT3 inhibition alone [10]. Midostaurin and its metabolites, CGP52421 and CGP62221, have been reported as inhibitors of SYK in advanced systemic mastocytosis [25]. However, the activity of midostaurin as an inhibitor of SYK has not yet been explored in FLT3-mutant-positive AML.
Midostaurin, which has recently been FDA approved, was shown in a large trial (RATIFY (CALGB 10603)) in newly diagnosed patients to increase survival when combined with standard chemotherapy [7]. This study supports the notion that inhibition of FLT3 is important in patients with mutations in the FLT3 gene, and also highlights the clinical effectiveness of midostaurin as a therapeutic for AML. As highly activated SYK has been found to occur at a higher frequency in AML patients harboring the ITD mutant than patients expressing wt Oncotarget 52038 www.impactjournals.com/oncotarget FLT3 [10], we were interested in exploring the activity of midostaurin against SYK using cell-based models of SYK-driven and SYK-and FLT3-driven leukemia. We found that while midostaurin is an excellent inhibitor of FLT3-ITD-expressing cells, it is a relatively weak inhibitor of activated SYK alone (SYK-TEL or TEL-SYK) and activated SYK (SYK-TEL) co-expressed with FLT3-ITD. Specifically, we found the potency of midostaurin against activated SYK-expressing cells to be 2.5 to 5-fold less than that of the targeted SYK inhibitor, PRT062607. Conversely, the potency of midostaurin against cells expressing FLT3-ITD or FLT3-ITD+TEL-SYK (which are believed to be driven predominantly by FLT3-ITD due to a lack of transactivation of FLT3 by TEL-SYK) was close to 100-fold higher than PRT062607. Thus, the benefit of dual suppression of FLT3 and SYK by midostaurin allows this agent to be more versatile than a targeted SYK inhibitor in that it is effective against cells expressing oncogenic FLT3, activated SYK, or cells expressing both oncoproteins.
Having established the extent by which SYK is inhibited by midostaurin in addition to FLT3 in comparison to the dual SYK/FLT3 inhibitors, R406 and R788, we were then interested in exploring whether or not additional SYK inhibition is able to further improve the efficacy of midostaurin. Potentiation of the effects of midostaurin by R406 or PRT062607, as measured in cell growth assays and evidenced by increased killing of constitutively activated SYK and/or mutant FLT3expressing cells, correlated with increased suppression of signaling molecules, including those involved in the phosphatidylinositol 3-kinase pathway and mitogenactivated protein kinase signaling, characterized as phosphorylated substrates and exerting effects downstream of either oncogenic FLT3 or activated SYK [29,30,5,24].  Our results are consistent with earlier findings showing that, in a SYK-TEL cooperative model with FLT3-ITD, the combination of PRT062607 and quizartinib increased survival of mice, which correlated with inhibition of FLT3-ITD and SYK activation [10]. Importantly, our results also suggest that the effects of midostaurin can be potentiated by dual FLT3/SYK suppression or more targeted SYK suppression in the context of midostaurinresistant leukemia and in FLT3-ITD-expressing primary cells.
Drug combination effects were observed against phosphorylated STAT5, S6, MAPK, and AKT by midostaurin+R406 or midostaurin+PRT062607 in cells expressing constitutively activated SYK. Significantly, STAT5 has been identified as a major downstream signaling mediator of constitutively activated SYK, both in vitro and in vivo, and targeted deletion of Stat5 in vivo was shown to fully abrogate AML caused by TEL-SYK [24]. These results support the notion of STAT5 as a key driver of SYK-induced cellular transformation, and shed light on the potential clinical importance of our observation of enhanced suppression of activated STAT5 by midostaurin combined with inhibitors of FLT3 and SYK or SYK alone.
Importantly, the efficacy of midostaurin, R406 and R788 toward cells expressing FLT3-ITD alone or FLT3-ITD with activated SYK was compared with the efficacy of targeted FLT3 inhibitors quizartinib and crenolanib, neither of which targets SYK. We observed that all of the inhibitors were highly potent toward FLT3-ITD-driven cells in the absence of activated SYK, however midostaurin, R406 and R788 showed relatively higher potency toward cells expressing activated SYK.
Whereas inhibition of SYK alone does not substitute for FLT3 inhibition and is incapable of inducing death of cells predominantly driven by FLT3-ITD, additional SYK inhibition was observed to potentiate the antiproliferative activity of both multi-targeted and selective FLT3 inhibitors against cells driven by activated SYK. Taken together, the ability of SYK and dual FLT3/SYK suppression to enhance the growth inhibitory effects of midostaurin and other FLT3 inhibitors in the context of both kinase inhibitor-sensitive and kinase inhibitorresistant disease warrants further investigation for the continued development and optimization of midostaurin as a treatment strategy for AML.
All cell lines used in this study were cultured with 5% CO 2 at 37 o C, at a concentration of 2X10 5 to 5X10 5 in RPMI (Mediatech, Inc., Herndon, VA) with 10% fetal bovine serum (FBS) and supplemented with 2% L-glutamine and 1% penicillin/streptomycin. Parental Ba/ F3 cells and Ba/F3-wt FLT3 cells were cultured in RPMI with 10% FBS and supplemented with 2% L-glutamine and 1% penicillin/streptomcyin, as well as 20% WEHI (as a source of IL-3).
Human cell lines were submitted for cell line authentication and were authenticated within 6 months of manuscript preparation through cell line short tandem repeat (STR) profiling (DDC Medical, Fairfield, OH and Molecular Diagnostics Laboratory, Dana-Farber Cancer Institute). All cell lines tested matched >80% with lines listed in the ATCC or DSMZ Cell Line Bank STR. All cell lines were confirmed to be virus-and Mycoplasma-free.

Immunoblotting and immunoprecipitation
Protein lysate preparation, immunoblotting, and immunoprecipitation were carried out as previously described [6]. Briefly, to obtain protein lysates for immunoprecipitation and immunoblotting, cells were lysed in lysis buffer (0.02 M Tris [pH 8.0], 0.15 M NaCl, 10% glycerol, 1% NP-40 (wt/vol), 0.1 M NaF, 1 mM phenylmethylsulfonyl fluoride (PhCH2SO2), 1 mM sodium orthovanadate (Na3VO4), and HALT TM protease inhibitor cocktail, EDTA-free (100X) (ThermoFisher Scientific, Waltham, MA). Protein lysates were incubated on ice for 25 min, vortexed at 5 min intervals, and then centrifuged for 15 min at 12,000 X g. Supernatants were saved, and the Bio-Rad Protein Assay was used to determine protein yields (Bio-Rad Laboratories, Hercules, CA). Equivalent amounts of protein were then loaded onto a gel for immunoblotting. For immunoprecipitation, cell lysates were incubated with antibody and protein A/G Sepharose overnight with rocking at 4 o C. After the incubation, immune complexes were washed twice with lysis buffer, twice with 1X PBS, and were then resuspended in Laemmeli's sample buffer and boiled for 5 min. For immunoblotting and immunoprecipitation, samples were resolved on a sodium dodecyl sulfate (SDS)-10% polyacrylamide gel. Then proteins were electrophoretically transferred to a Protran nitrocellulose transfer and immobilization membrane (Schleicher and Schuell, Dassel, Germany). The membrane was then blocked overnight at 4 o C with BSA or milk in 1X TBS (10mM Tris-HCl [pH 8.0], 150 mM NaCl) and then probed with antibody overnight at 4 o C in 1X TBST buffer (10 mM Tris-HCl [pH 8.0], 150 mM NaCl, 0.05% Tween20). After three washes with 1X TBST, membranes were incubated for 1 hr at 25 o C with anti-mouse immunoglobulin (horseradish peroxidase-linked whole antibody from sheep) (Amersham Life Science, Inc., Arlington Heights, IL). The membrane was washed 5X in 1X TBST buffer, with 5 min intervals between buffer changes, and bound antibody was detected with enhanced luminol and oxidizing reagent as specified by the manufacturer (NEN Life Science Products, Boston, MA).

Cell growth studies
The trypan blue exclusion assay has been previously described [6] and was used for quantification of cells prior to seeding for Cell Titer Glo assays. The CellTiter Glo assay (Promega, Madison, WI) was used for proliferation studies and carried out according to manufacturer instructions. Cell viability is reported as percentage of www.impactjournals.com/oncotarget control (untreated) cells, and error bars represent the standard deviation for each data point.

Apoptosis assays and cell cycle analysis
Programmed cell death of inhibitor-treated cells was determined using the Annexin-V-Fluos Staining Kit (Boehringer Mannheim, Indianapolis, IN), as previously described [6]. Briefly, cells were washed once with 1X PBS and pelleted by centrifugation for 5 minutes at 1500 rpm. Cells were resuspended in 100µl of 20% propidium iodide (PI) and 20% Annexin-V-fluorescein labeling reagent, either agent alone (as controls), or were left unstained by diluting only in 1X binding buffer (as a control). All samples were incubated for 10-15 minutes at room temperature, and then stained cells were diluted in 0.8 mL of 1X binding buffer. Cells were then analyzed by flow cytometry.
Cell cycle analysis was performed as previously described [6]. Briefly, around 500,000 cells were centrifuged at 1500 rpm for 5 min and washed in 1X PBS, and the pellet was resuspended in 500 μl of propidium iodide solution (50 μg/ml propidium iodide, 0.1% NP-40, 0.1% sodium citrate). The mixture was stored in the dark at 4°C for a minimum of 15 min, and then analyzed by flow cytometry.

Caspase 3 activation assay
For measurement of caspase-3 and -7 activity, the Apo-ONE Homogenous Caspase-3/7 Assay kit was used (Promega, Madison, WI). The assay was carried out according to manufacturer's instructions.

Drug combination studies
For drug combination studies, cell viability was first determined using the Trypan Blue exclusion assay to quantify cells for cell seeding, and Cell Titer Glo was then implemented for proliferation studies. Single agents were added simultaneously at fixed ratios to cells. Cell viability was expressed as the function of growth affected (FA) drug-treated versus control cells; data were analyzed by Calcusyn software (Biosoft, Ferguson, MO and Cambridge, UK), which was utilized for synergy measurement and based on isobologram generation and the method of Chou-Talalay (1984) [33]. This method utilizes the median effect principle to quantify the effects of drug combinations to determine whether they give greater effects together than expected from a simple summation of their individual effects. After determining the ED 50 or IC 50 of each drug, combinations are studied where the concentrations are multiples, or fractions, of the ED/IC 50 . Statistical analysis is automatically part of the computations. Combination indices, values generated by the Calcusyn software, which are less than one indicate synergy, whereas values greater than one indicate antagonism. Calcusyn combination indices can be interpreted as follows: CI <0. 1