Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Combined venetoclax and alvocidib in acute myeloid leukemia

Combined venetoclax and alvocidib in acute myeloid leukemia www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 63), pp: 107206-107222 Research Paper 1,* 2,* 1 1 James Bogenberger , Clifford Whatcott , Nanna Hansen , Devora Delman , 1 2 2 2 2 Chang-Xin Shi , Wontak Kim , Hillary Haws , Katherine Soh , Ye Sol Lee , Peter 2 2 2 1 2 Peterson , Adam Siddiqui-Jain , Steven Weitman , Keith Stewart , David Bearss , 1 2 1,3 Ruben Mesa , Steven Warner and Raoul Tibes Mayo Clinic, Scottsdale, AZ, USA Tolero Pharmaceuticals, Lehi, UT, USA NYU School of Medicine, New York, NY, USA These authors contributed equally to this work Correspondence to: Raoul Tibes, email: raoul.tibes@nyumc.org James Bogenberger, email: bogenberger.james@mayo.edu Keywords: acute myeloid leukemia (AML); alvocidib/flavopiridol; venetoclax/ABT-199; BCL-2 Received: July 22, 2017 Accepted: October 10, 2017 Published: November 03, 2017 Copyright: Bogenberger et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ABSTRACT More effective treatment options for elderly acute myeloid leukemia (AML) patients are needed as only 25–50% of patients respond to standard-of-care therapies, response duration is typically short, and disease progression is inevitable even with some novel therapies and ongoing clinical trials. Anti-apoptotic BCL-2 family inhibitors, such as venetoclax, are promising therapies for AML. Nonetheless, resistance is emerging. We demonstrate that venetoclax combined with cyclin-dependent kinase (CDK) inhibitor alvocidib is potently synergistic in venetoclax-sensitive and -resistant AML models in vitro, ex vivo and in vivo. Alvocidib decreased MCL-1, and/or increased pro-apoptotic proteins such as BIM or NOXA, often synergistically with venetoclax. Over-expression of BCL-XL diminished synergy, while knock-down of BIM almost entirely abrogated synergy, demonstrating that the synergistic interaction between alvocidib and venetoclax is primarily dependent on intrinsic apoptosis. CDK9 inhibition predominantly mediated venetoclax sensitization, while CDK4/6 inhibition with palbociclib did not potentiate venetoclax activity. Combined, venetoclax and alvocidib modulate the balance of BCL-2 family proteins through complementary, yet variable mechanisms favoring apoptosis, highlighting this combination as a promising therapy for AML or high-risk MDS with the capacity to overcome intrinsic apoptosis mechanisms of resistance. These results support clinical testing of combined venetoclax and alvocidib for the treatment of AML and advanced MDS. and apoptotic resistance [8–12]. Clinical activity of INTRODUCTION venetoclax monotherapy is modest in AML [13], which may partially be explained by the genetic heterogeneity Targeting of intrinsic apoptosis via inhibition of AML, and more importantly by the heterogeneous of B-Cell Lymphoma 2 (BCL-2) family proteins is and concurrent expression of multiple anti-apoptotic a promising therapeutic strategy. BCL-2-selective BCL-2 family proteins in AML, and the selectivity of inhibitor venetoclax (ABT-199) has demonstrated clinical venetoclax for BCL-2 [14]. Consequently, therapeutics activity in chronic lymphocytic leukemia (CLL) and targeting complementary BCL-2 family proteins may has recently received FDA-approval for some relapsed increase the efficacy of venetoclax. Preclinical data CLL patients [1]. Importantly, BCL-2 family proteins demonstrate that BCL-2 family inhibition synergizes with are also therapeutic targets in acute myeloid leukemia hypomethylating agents (HMAs) [14–16] and, although (AML) [2–7] and high-risk myelodysplastic syndrome the precise mechanism of synergy remains incompletely (MDS) where increases in anti-apoptotic BCL-2 characterized, HMAs may down-regulate MCL-1 [16]. family proteins are associated with disease progression www.impactjournals.com/oncotarget 107206 Oncotarget Initial clinical trials examining venetoclax with HMAs in of venetoclax -resistant cells from µM to low nM doses previously untreated elderly AML patients have reported (30 to 50 nM) similar to that of sensitive cells/primary impressive overall response rates of > 75% [17], although samples. Fold-sensitization at 80 nM alvocidib was 14.5 clinical resistance is emerging. Putative mechanisms of ± 0.8 (p = 0.028) and 10.1 ± 3.4 (p = 0.17) in MOLM- venetoclax resistance include compensatory up-regulation, 13 and MV4-11, respectively. At clinically-achievable increased stability, or altered function of MCL-1 [18–20]. plasma concentrations of 80 and 160 nM alvocidib, MCL-1 plays a role in intrinsic resistance to BCL-2 synergy (expressed as Combination Index (CI) values) inhibitors by alternatively sequestering BIM dissociated was observed with all clinically-achievable doses of from BCL-2 by BCL-2 inhibitors [21–23]. MCL-1 venetoclax tested, in all cell lines examined (Figure 1B). inhibition or genetic knock-down is known to potentiate To confirm that synergistic effects of combined venetoclax BCL-2 inhibitors [7, 14, 24]. The cyclin-dependent kinase and alvocidib culminate in increased apoptosis, as opposed (CDK) inhibitor alvocidib (flavopiridol) reduces MCL-1 in to only cytostatic effects from putatively inhibiting hematologic malignancies [25–27]. Alvocidib is a potent cell cycle CDKs, we analyzed Annexin V levels and inhibitor of CDK9, the core catalytic component of the propidium iodide permeability by flow cytometry. In all P-TEFb complex [28–30]. Transcriptional repression of cells examined, we observed an increase in early and late short half-life oncoproteins, such as MCL-1, can potently apoptotic cells in response to the combination beyond induce apoptosis; however, additional anti-apoptotic the additive effects of either single-agent (Figure 1C). In proteins, such as BCL-2, can functionally counter this parallel, we assessed cell cycle distributions and found that repression by sequestering pro-apoptotic BH3-only 80 nM alvocidib resulted only in a moderate proportional proteins or blocking BAX/BAK dimerization. Increased increase in G1, with corresponding decreases in S and G2, pro-apoptotic BH3-only proteins such as BIM have also in three of four cell lines analyzed (median increase 28 ± been observed in response to alvocidib treatment [31]. We 5%); however, 80 nM alvocidib did not significantly alter hypothesized that alvocidib would synergize with BCL- cell cycle distribution in OCI-AML3 (Supplementary 2 inhibitor venetoclax in AML through complementary Figure 1). mechanisms, namely decreased MCL-1 and increased BIM, to modulate the overall balance of anti- and pro- Correlation of BCL-2 family proteins with apoptotic BCL-2 family proteins in favor of apoptosis alvocidib/venetoclax activity induction. To determine whether anti-apoptotic BCL-2 family RESULTS members correlate with single-agent alvocidib anti- leukemic activity, we initially quantified baseline protein Alvocidib potentiates venetoclax anti-leukemic levels of BCL-2, BCL-XL and MCL-1 in untreated cells. activity in both venetoclax-sensitive and – MCL-1 protein was relatively homogenous, differing by a median of 1.7 ± 0.8-fold. In contrast, BCL-2 levels were resistant AML cells highly variable, differing by 158-fold between the lowest To assess the potential for synergy between and highest expressing cells. BCL-XL protein expression venetoclax and alvocidib in venetoclax -sensitive AML was also variable spanning an 11.6-fold range (Figure 2A). cells in vitro, we selected MOLM-13 and MV4-11 cell Relative protein levels were then plotted against single- lines, which have low single-agent venetoclax half- agent alvocidib EC values from cell viability assays. maximal effective concentration (EC ) values (in assays BCL-2 levels did not correlate with alvocidib activity, assessing relative cell number) of 9.0 ± 1.6 and 7.8 ± 2.1 while BCL-XL levels positively correlated, and MCL- nM, respectively. These low nM venetoclax EC values 1 levels negatively correlated with alvocidib activity are consistent with values for primary AML samples (Figure 2B). BCL-2, BCL-XL and MCL-1 protein levels sensitive to venetoclax [7]. THP-1 and OCI-AML3 did not significantly correlate with venetoclax single-agent were selected to model venetoclax -resistant AML cells, activity in this panel of AML cell lines (Figure 2C). exhibiting higher single-agent venetoclax EC values of In contrast to the potent alvocidib sensitization 0.9 ± 0.2 and 2.3 ± 0.4 µM, respectively, 100- to 295- resulting in clinically meaningful 30–50 nM venetoclax fold greater than venetoclax -sensitive cells. Addition of EC values observed in venetoclax-resistant cells with Low alvocidib resulted in potent dose-dependent reduction low levels of BCL-XL (BCL-XL ), venetoclax-resistant High of venetoclax EC values in both venetoclax -sensitive cells with high levels of BCL-XL (BCL-XL ) exhibited and –resistant cells (Figure 1A). venetoclax EC fold- EC values that were not clinically meaningful ranging 50 50 sensitization by alvocidib was greatest in venetoclax from 1.7 to 4.7 µM (median 3.4 ± 1.2 µM). ABT-737 -resistant cells, with 80 nM alvocidib resulting in an 18.3 (navitoclax), a drug that inhibits BCL-XL and BCL-W ± 0.1 (p = 0.0039) and 77.7 ± 1.2 (p = 0.0069) venetoclax in addition to BCL-2, was used to examine the role of EC fold-shift for THP-1 and OCI-AML3, respectively. these additional BCL-2 family members in alvocidib Importantly, alvocidib shifted the absolute EC values synergy. Combined with alvocidib, EC values for ABT- 50 50 www.impactjournals.com/oncotarget 107207 Oncotarget Figure 1: Alvocidib potentiates venetoclax anti-leukemic activity in both venetoclax -sensitive and –resistant AML cells. (A–B) combination drug dose response assays with venetoclax and alvocidib in vitro were assessed in duplicate biological experiments, each containing four technical replicate data points for every dose/dose combination analyzed. Data represent mean ± SEM. The indicated AML cell lines were dosed with venetoclax or alvocidib as single-agent, and in combination, and incubated for 96 hours before determining relative cell number with ATP-based reagent CellTiter Glo. (A) leftward shifts toward lower doses of venetoclax demonstrate dose- dependent venetoclax fold-sensitization by alvocidib. (B) Combination Index (CI) values were calculated with CalcuSyn Software, and are shown for distinct dose combinations of venetoclax and alvocidib. CalcuSyn R values corresponding to single-agent dose curves are shown for each cell line below CI value tables. (C) AML cell lines were treated for 24 hours with 80 nM alvocidib, and a low dose or high dose of venetoclax, each alone and in combination, prior to harvesting for flow cytometry quantification of annexin V and propidium iodide * † permeability as a measurement of apoptosis. For venetoclax -sensitive cells MOLM-13 and MV4-11, 2.5 and 10 nM were used, while * † for venetoclax -resistant cell lines THP-1 and OCI-AML3, 0.25 and 1 µM venetoclax were used. Quantification from a representative experiment is shown graphically, and apoptosis results were confirmed in biological replicate experiments using one venetoclax -sensitive cell line (MOLM-13) and one venetoclax -resistant cell line (THP-1). www.impactjournals.com/oncotarget 107208 Oncotarget High Low 737 were lower than that of venetoclax in BCL-XL 13) and venetoclax-resistant (THP-1) BCL-XL cells, cells (Supplementary Figure 2A, upper panel). EC which rendered these cells more resistant to single-agent fold-sensitization by alvocidib was greater for ABT-737 venetoclax (from nM to µM doses) (Supplementary Figure High High than venetoclax in BCL-XL cells (Supplementary 2E). Similar to BCL-XL cells, although sensitization Low Figure 2A, lower panel). In contrast, BCL-XL cells was observed in BCL-XL over-expressing cells, the (both venetoclax-sensitive and -resistant) were equally, resultant venetoclax EC values were no longer clinically or more, sensitive to single-agent venetoclax than ABT- meaningful (Supplementary Figure 2F). 737 (Supplementary Figure 2B). Knock-down of BCL- XL significantly reduced venetoclax EC values as BCL-2 family proteins altered by alvocidib compared to non-silencing (NS) siRNA (Supplementary Figure 2C–2D). BCL-XL was over-expressed with a To examine effects of alvocidib on the BCL-2 family lentiviral construct in both venetoclax-sensitive (MOLM- protein landscape, we treated venetoclax-sensitive and Figure 2: Correlation of BCL-2 family proteins with alvocidib and venetoclax activity. (A) lysates were prepared from untreated AML cell lines, and levels of the indicated anti-apoptotic BCL-2 family proteins measured by western blot. Image J densitometry software was used to quantify bands and values normalized to β-tubulin. (B and C), relative protein levels of MCL-1, BCL-XL and BCL-2 were plotted against alvocidib EC values (B) or venetoclax EC values (C) determined from duplicate biological experiments. Regression 50 50 analysis was used to determine R and P values. www.impactjournals.com/oncotarget 107209 Oncotarget –resistant AML cells with ascending doses and assessed MOMP. Thus, we sought to evaluate the functional role total protein. Consistent with previous reports in other of BIM, specifically with regard to synergy between hematologic malignancies [25–27], we observed a dose- alvocidib and venetoclax. Knock-down of BIM prior to dependent reduction of ~40 kD anti-apoptotic isoform alvocidib and venetoclax combination treatment potently MCL-1 in three of four AML cell lines. Increased abrogated synergy compared to NS siRNA. BIM knock- Long BIM was observed in all four cell lines in response to down abrogated 97% (p = 0.0066) and 85% (p = 0.0075) alvocidib, although the increase was more pronounced of the venetoclax fold-sensitization induced by 80 nM in venetoclax-sensitive cells. MOLM-13 was unique alvocidib in OCI-AML3 and THP-1 cells, respectively. in that we did not observe a reduction in MCL-1 , Mean combination venetoclax EC values were 1.8 and Long 50 but instead observed a dose-dependent induction of all 1.3 µM with BIM siRNA, versus 0.05 and 0.14 µM with BH3-only proteins measured (NBK, NOXA, PUMA and NS control siRNA, demonstrating that synergy between BIM). BCL-2 was not significantly changed by alvocidib venetoclax and alvocidib is almost entirely dependent on treatment. Conversely, BCL-XL was variably induced by the intrinsic apoptotic pathway (Figure 3F–3G). alvocidib in all four cell lines (Figure 3A). However, the Low induced level of BCL-XL in BCL-XL MOLM-13 cells Pharmacological dissection of CDK isoforms High was significantly lower than baseline levels in BCL-XL contributing to venetoclax sensitization cells (Figure 3B). MCL-1 protein stability is regulated via complex phosphorylation by multiple kinases including To discern the contributions of different CDK CDK1 and -2 [32, 33]. Signaling via ERK, JNK, GSK3 isoforms toward venetoclax sensitization, we examined and CDK2, converge on Thr-163 phosphorylation of CDK inhibitors with putatively distinct inhibitory profiles MCL-1, which increases MCL-1 stability and thereby (LDC067, palbociclib, LY2857785, seliciclib, NU6102, plays a role in resistance to BCL-2 inhibitors [34]. To and Ro-3306), as well as a BRD4 inhibitor (JQ1) to more investigate whether alvocidib affects MCL-1 protein specifically assess P-TEFb. We compared single-agent stability through putative CDK1/2 inhibition, in addition activity of these compounds in a panel of ten AML cell to MCL-1 transcriptional down-regulation by CDK9 lines with diverse molecular, cytogenetic and lineage inhibition, we measured MCL-1 Thr-163 phosphorylation backgrounds (MDS-L, HL-60, OCI-AML2, OCI-AML3, levels. Alvocidib did not significantly decrease MCL-1 MV4-11, MOLM-13, THP-1, TF-1, HEL, and SET-2). Thr-163 phosphorylation in two of three cell lines where Correlation of each compound with alvocidib was used decreases in total MCL-1 were observed (Figure 3A). to assess pharmacological similarity. We also combined Published studies suggest that the polycistronic microRNA each of these unique CDK inhibitors with venetoclax to (miRNA) miR-17–92 negatively regulates BIM [35, 36], assess interaction. thus we measured five miRNAs derived from the miR-17– In terms of single-agent EC correlation, CDK9- 92 cluster in response to alvocidib treatment. Relative to selective inhibitor LDC067 [37] showed the strongest control miRNAs, all five miRNAs decreased in a dose- and correlation with alvocidib. Palbociclib, a recently FDA- time-dependent manner (Figure 3C). We were surprised to approved CDK4/6-selective inhibitor, did not significantly observe miR-17–92 decreases in THP-1 and OCI-AML3, correlate with alvocidib. Seliciclib (roscovitine), the most where potent increases in BIM protein were not observed well-characterized CDK5 inhibitor which also inhibits at 24 hours. Thus, we performed western blots at 48 CDK2, -7 and -9 with > 100-fold selectivity over CDK4 hours in these two cell lines and observed a more delayed and -6 [38, 39], significantly correlated with alvocidib. increase of BIM (Figure 3D). Next we examined BCL-2 LY2857785 [40], which compared to alvocidib more family proteins after 24 hour combination treatment, as potently inhibits CDK8, shows less potency for CDK5 compared to single-agent and untreated controls. Three of and similarly inhibits CDK9, correlated with alvocidib. four cell lines showed increases in either BIM or NOXA NU6102, a CDK2 and -1 selective inhibitor [41] also with the combination treatment. Surprisingly, we observed significantly correlated with alvocidib. Ro-3306, a CDK1- variable increases in MCL-1, BCL-2, and BCL-XL with selective inhibitor with 10-fold greater selectivity over combination treatment in all cell lines. The observed CDK2 and 50-fold over CDK4 [42] (also reported to increase in MCL-1 corresponded to increased Thr-163 decrease BCL-2 protein in AML [43]) did not correlate phosphorylation, which could primarily be attributed to with alvocidib. BRD4 inhibition with JQ1 exhibited single-agent venetoclax treatment in venetoclax-resistant a significant correlation with alvocidib. Notably, only cells, but not in venetoclax-sensitive cells (Figure 3E). alvocidib, LY857785, and JQ1 exhibited low nM EC The functional importance of reducing/inhibiting values, while the remaining CDK inhibitors exhibited µM MCL-1 for the anti-leukemic activity of BCL-2 inhibitors EC values (Figure 4A). is well established, including a role in sequestration Regarding synergy with venetoclax, selective CDK9 of BIM liberated by venetoclax [22]. BIM is a critical inhibition with LDC067 only partially recapitulated effector protein of the intrinsic apoptotic pathway with the full magnitude of potentiation and dose-dependent a capacity to act as a dual sensitizer and activator of activity of alvocidib. LDC067 was significantly less www.impactjournals.com/oncotarget 107210 Oncotarget Figure 3: BCL-2 family proteins altered by alvocidib. (A) AML cell lines were treated with 20 or 80 nM single-agent alvocidib for 24 hours, and BCL-2 family proteins quantified by western blot of duplicate biological extracts. ( B) AML cell lines with low (MOLM- 13) or high (HEL and SET-2) BCL-XL levels were treated with 80 nM alvocidib and resolved on the same gel to facilitate quantitative comparison across these cell lines. (C) After treatment with the alvocidib doses indicated on the x-axis, RNA extracts were prepared at 3 hours (upper panel) or 24 hours (lower panel), and the microRNAs indicated were measured by qPCR. Values shown are mean ± SEM of technical replicates and are representative of three biological replicate experiments. (D) THP-1 and OCI-AML3 were treated with 20 or 80 nM single-agent alvocidib for 24 and 48 hours, and BIM quantified by western blot. ( E) AML cell lines were treated with 80 nM alvocidib * * and 2.5 nM for venetoclax-sensitive lines (MOLM-13 and MV4-11) or 0.25 µM for venetoclax -resistant lines (THP-1 and OCI-AML3), as single-agent and in combination and BCL-2 family proteins quantified by western blot of duplicate biological extracts. ( F) OCI-AML3 and THP-1 cell lysates prepared after treatment with non-silencing (NS) or BIM siRNA were resolved by gel electrophoresis to quantify the level of specific BIM knock-down. ( G) BIM siRNA sequence 5 (the only sequence tested that effectively knocked-down all BIM isoforms) was compared against NS siRNA in cells treated with combined venetoclax and alvocidib in duplicate biological experiments. venetoclax EC values are plotted on a log10 scale at increasing concentrations of alvocidib. Data represent average ± STD. P values were calculated * ** using a two-tailed Student’s T-test. P value < 0.05, P value < 0.01. A–B & D–F, densitometry values shown were calculated using Image J software and normalized to β-Tubulin. www.impactjournals.com/oncotarget 107211 Oncotarget synergistic than alvocidib in venetoclax-resistant cells. values < 0.7, 4 of 13 (31%) exhibited “moderate synergy” Palbociclib showed marginal venetoclax fold-sensitization defined by CI values of 0.7 to 0.85, while one sample had in venetoclax-sensitive cell lines and only at the highest a CI of 0.87 defined as “slight synergy,” and remission (3 µM) dose. Further, palbociclib exhibited divergent sample #11 had a CI of 0.98 indicative of additive activity. activity in venetoclax-resistant cells, with dose-dependent The combination was synergistic in samples from patients antagonism observed in THP-1, and low-dose antagonism clinically refractory to azacitidine (samples 7 & 13) and high-dose sensitization observed in OCI-AML3. (Supplementary Table 1). With seliciclib, venetoclax sensitization (ranging from We analyzed protein levels of selected BCL- 33- to 108-fold) was observed in all cells examined; 2 family members in untreated extracts from eleven however, venetoclax-sensitization occurred only at high available samples used for ex vivo synergy assessment. (≥ 17.5 µM) seliciclib doses. Comparing equipotent Consistent with our previous report, BCL-2 family doses, LY2857785 resulted in similar maximal venetoclax expression was overlapping within individual samples sensitization, although LY2857785 did not recapitulate and heterogeneous across samples [14]. MCL-1 resolved the dose-dependent activity of alvocidib. NU6102 dose- as two distinct molecular weight (MW) bands, with dependently sensitized venetoclax in all cells examined, some samples exhibiting both bands, while most samples with maximal sensitization ranging from 20- to 193-fold predominantly expressed the lower MW band. MCL- at variable NU6102 doses from 2.5 to 20 µM. Despite 1 phosphorylated at Thr163 was detectable only in the poor single-agent correlation, Ro-3306 resulted in dose- single ex vivo venetoclax-resistant sample (#2) (Figure dependent sensitization in all cells examined, with 5C). BCL-XL, BCL-2 and BIM were not detected in the maximal venetoclax EC sensitization ranging from 130- remission sample (#11). No individual BCL-2 family to 242-fold. JQ1 generally recapitulated CDK inhibitor member correlated with synergy, as determined using activity in potentiating venetoclax, exhibiting greater CI values from Figure 5B (Supplementary Figure 5). potentiation in venetoclax-sensitive lines (Figure 4B–4C Levels of BIM correlated most strongly with BCL-2, and and Supplementary Figure 3). We measured apoptosis also significantly correlated with BCL-XL, yet neither to determine whether venetoclax potentiation with total MCL-1 nor high MW MCL-1 correlated with BIM, NU6102 or seliciclib was cytostatic or apoptotic. Similar suggesting that this small sampling of primary AML to alvocidib, early and late apoptosis increased in the samples was primarily dependent on BCL-2 followed by combination with each of these CDK inhibitors beyond BCL-XL for survival (Supplementary Figure 6). For two either single-agent alone, albeit at higher µM doses samples (#6 and #10), sufficient sample quantity enabled (Supplementary Figure 4A–4B). treatments with venetoclax and/or alvocidib. Consistent with divergent in vitro results, we found that alvocidib decreased MCL-1 in sample #6, yet increased MCL-1 in Alvocidib synergizes with venetoclax in short- sample #10. Total BIM was not significantly changed by term ex vivo cultures of AML patient samples alvocidib in either primary sample, while both showed increases in BCL-2 in the combination, consistent with in We analyzed combined venetoclax and alvocidib vitro results (Figure 5D). We identified and treated two in patient samples cultured ex vivo (N = 14). Twelve of additional AML samples to expand upon these results fourteen samples were confirmed as overt AML, one (presented in Figure 5A–5B as #13 and #14). Sample #14 sample was derived from a patient with residual disease exhibited a decrease in MCL-1 with 80 nM alvocidib, (sample #4), and one sample was derived from an which was further decreased by the combination. Sample AML patient in remission (sample #11). Of the samples #13 showed a marginal decrease of MCL-1 only with the analyzed, only one was resistant to venetoclax (sample #2; combination. Neither sample #13 or #14 showed increased EC 1.3 ± 0.08 µM), while the remission sample was also BIM. MCL-1 Thr163 phosphorylation was undetectable resistant (EC > 5 µM) as expected. Median venetoclax or detected at only low levels, and was not significantly EC was 5 nM (excluding the two resistant samples). altered with any treatment (Figure 5E). Alvocidib single-agent activity was homogeneous (median EC 52 ± 19 nM), which for the remission sample was not unexpected as CDK inhibitors can cause cytopenias Alvocidib potentiates venetoclax activity in vivo (Figure 5A). All of the confirmed AML samples exhibited To investigate alvocidib potentiation of venetoclax synergy. Median maximal CI value was 0.52 ± 0.18, activity in vivo, we assessed venetoclax-resistant OCI- occurring at variable doses of venetoclax and alvocidib. AML3 cells in a mouse xenograft model using tumor For consistent comparison, we plotted CI values for volume as the primary endpoint. Animal body weight was median doses at which maximal synergy occurred (40 recorded throughout the study as a surrogate for treatment nM venetoclax, 80 nM alvocidib), alongside the fraction toxicity. Treatment of OCI-AML3 tumor-bearing mice of cells affected, expressed as Fraction Affected (FA) with 100 mg/kg venetoclax resulted in a 31.6% average (Figure 5B). At these low, clinically-achievable doses, 7 tumor growth inhibition (TGI) by day 14 relative to of 13 samples (54%) exhibited “synergy” as defined by CI www.impactjournals.com/oncotarget 107212 Oncotarget vehicle treated animals. Alvocidib treatment at 2.5 mg/kg Altogether, both in vitro and ex vivo results are resulted in a 9.7% TGI over the same period. However, consistent in showing divergent effects on BCL-2 family the combination of alvocidib and venetoclax resulted in proteins despite similar levels of synergy, suggesting a significant reduction in tumor volume, corresponding to that no effect is absolutely associated with synergy per an 87.9% TGI at day 14 (Figure 6A). Neither venetoclax, se, yet the overall balance of pro- and anti-apoptotic nor alvocidib, resulted in a significant reduction in BCL-2 family proteins is consistently shifted in favor of bodyweight. The combination was well-tolerated, with apoptosis. Previous reports of exogenous manipulation of an average 7.5% reduction in bodyweight at day 7, MCL-1, and dual targeting of MCL-1 and BCL-2, support comparable to that of venetoclax alone (Figure 6B). this concept as it relates to alvocidib and enhanced venetoclax activity [26, 46–48]. Importantly, abrogation of synergy with either BIM knock-down or BCL-XL over- DISCUSSION expression also supports this concept and provides strong functional evidence that the majority of synergy between The synergistic anti-leukemic activity of combined venetoclax and alvocidib is based on mechanism(s) that venetoclax and alvocidib observed in venetoclax-sensitive converge upon the intrinsic apoptotic pathway. More and –resistant models of AML ex vivo, in vivo, and in specifically, while BIM knock-down showed a trend vitro suggests that this combination has the potential to towards abrogating single-agent venetoclax activity, improve outcomes in patients regardless of sensitivity the change was not significant, suggesting that, at to single-agent venetoclax. To our knowledge, this is the least in the two AML cell lines analyzed, single-agent first report exploring the potential synergy of venetoclax venetoclax operates largely independent of BIM through and alvocidib in AML. Considering the clinical relevance the classic displacement model [49], whereby BCL-2 of alvocidib and venetoclax, as well as lack of effective directly interacts with BAX/BAK to negatively regulate treatment options for AML, we propose that the current MOMP. Conversely, BIM knock-down abrogated the report is distinct from prior reports evaluating BCL-2 majority of synergy between venetoclax and alvocidib, inhibitors with CDK inhibitors, e.g. [44, 45]. suggesting that the combination is at least partially In the current report, a decrease of anti-apoptotic dependent on the direct activation model [49], whereby MCL-1 was observed in most AML cells tested, while an BIM acts as sensitizer and/or activator of BAX/BAK. To increase of pro-apoptotic BH3-only proteins, predominantly our knowledge the effect of BIM knock-down has not BIM, was observed in other cell lines. MOLM-13, a cell previously been examined in the context of the proposed line derived from an MDS patient transformed to AML, was combination, thus abrogation of synergy by BIM knock- unique amongst cell lines examined in that alvocidib did down is an important insight into the mechanism of not decrease MCL-1, yet demonstrated strong synergy with apoptotic synergy between alvocidib and venetoclax. venetoclax. Importantly, increases of all pro-apoptotic BH3- Importantly, reduction of MCL-1 by alvocidib similarly only proteins measured, including BIM, PUMA, NOXA has the capacity to operate through the displacement and NBK, were observed in MOLM-13. This suggests model, by decreasing MCL-1 available to inhibit BAX/ that an increase of pro-apoptotic BH3-only proteins can BAK, as well as the direct activation model, by decreasing be functionally equivalent to a decrease of anti-apoptotic MCL-1 available to act as a reservoir for sequestration of MCL-1 with regard to synergy with venetoclax. BIM and/or other BH3-only proteins. MicroRNAs from the polycistron miR-17–92, Using multiple complementary approaches, we known to negatively regulate BIM [35, 36], were clearly demonstrate that BCL-XL counteracts venetoclax/ decreased in alvocidib-treated cells in vitro, although High alvocidib synergy in vitro. Results with BCL-XL cells further experiments are needed to determine if this is a are consistent with a model whereby BCL-XL acts in direct or indirect result of CDK9 inhibition. Combined parallel to BCL-2 to inhibit BAX/BAK and/or sequester venetoclax and alvocidib was found to increase pro- BIM/BH3-only proteins, and with a model whereby BCL- apoptotic BH3-only proteins beyond the increase XL can also counteract MCL-1 reduction and/or BH3- observed with each single-agent for some cell lines. only protein increases by alvocidib. Given the totality Thus, in addition to increased levels of “free” BH3- of in vitro data regarding BCL-XL, we were surprised to only protein(s) (i.e. unbound by anti-apoptotic proteins) observe increased BCL-XL protein in alvocidib treated resulting from BCL-2 inhibition by venetoclax, and/or samples in vitro. Importantly however, the levels of alvocidib reduction of MCL-1 protein, alvocidib may have Low increased BCL-XL observed in BCL-XL cells were the capacity to induce apoptosis by increasing total BIM still strikingly lower than baseline levels observed in protein levels and/or other BH3-only proteins, although High BCL-XL cells, demonstrating that absolute levels of this activity was not observed in our small sampling of BCL-XL may be an important determinant of response AML analyzed ex vivo (n = 4). MCL-1 was decreased by in vitro. BCL-XL levels did not correlate with venetoclax alvocidib and/or the combination in three of four AML single-agent activity or synergy with alvocidib in our patient samples assessed ex vivo, suggesting MCL-1 as a small sampling of AML patients ex vivo, suggesting that predominant mechanism in the primary cells analyzed. www.impactjournals.com/oncotarget 107213 Oncotarget BCL-XL may not represent a major limitation to clinical venetoclax activity, and was actually antagonistic in some efficacy of the combination in AML. However, several AML cells. While we cannot rule-out the possibility that questions regarding BCL-XL as an in vivo mechanism CDK4/6 inhibition contributes to venetoclax sensitization of resistance to the combination remain unanswered, under concurrent CDK9 inhibition, these results suggest such as differences in relative versus absolute levels of that palbociclib may not be an ideal candidate for BCL-XL in vitro versus in vivo, as well as potential cell combination with venetoclax in AML. Strong venetoclax lineage effects of BCL-XL function [50]. It remains to be potentiation was also observed with seliciclib, NU6102, determined clinically whether some cases of AML with and Ro-3306; however, while these CDK inhibitors do high BCL-XL expression (putatively erythroleukemia or not potently inhibit CDK9, they likely inhibit CDK9 at megakaryoblastic lineage expression of BCL-XL) may be the relatively high doses required to elicit the synergy more resistant to the proposed combination of venetoclax observed. In comparing equipotent doses of LY2857785 or and alvocidib, and thus benefit more from dual BCL-2 alvocidib combined with venetoclax, maximal venetoclax and BCL-XL inhibition (i.e. navitoclax/ABT-737/ABT- fold-sensitization was similar for both compounds, 263) alone or combined with alvocidib. Nonetheless, our although synergy at the next closest dose was lower observations are consistent with recent reports showing for LY2857785, especially for venetoclax -resistant that mechanisms of venetoclax resistance are not universal cells. BRD4 inhibitor JQ1 resulted in dose-dependent or mutually exclusive [51, 52]. venetoclax sensitization similar to alvocidib, suggesting Upon examination of BCL-2 family proteins after that inhibition of P-TEFb through BRD4 is functionally combined venetoclax and alvocidib treatment, we were similar to CDK9 inhibition. As JQ1 unlikely has off-target surprised to observe increased MCL-1, BCL-2 and BCL- effects on CDKs, this indirect evidence supports the notion XL. However, we observed a corresponding increase that alvocidib predominantly synergizes with venetoclax in BIM beyond the increase attributable to single-agent through CDK9 inhibition. alvocidib in most cells, while NOXA was observed to be The functional and mechanistic evidence increased in the combination, beyond either single-agent, presented herein provide strong pre-clinical precedence in the only cell line for which we did not observe strong for clinically investigating the efficacy of combined BIM up-regulation at the same time point. BIM and BCL- alvocidib and venetoclax in AML. Clinical studies 2 have the capacity to mutually regulate each other, thus have shown that alvocidib can be safely added to we speculate that the increase in anti-apoptotic BCL-2 induction therapy with acceptable toxicity [57], thus family proteins could be a response to increased BIM, or the combination may be clinically feasible. However, vice versa [53]. Nonetheless, the data presented herein the cytopenias observed with alvocidib, and recently is consistent with the hypothesis that the net balance of with combined venetoclax and azacitidine (including pro- and anti-apoptotic BCL2 family members determines unexpected thrombocytopenia), emphasize the intrinsic apoptotic response in general, as well as for the importance of using the lowest possible doses, and proposed combination. BH3 profiling, an evolving assay highlights the importance of low dose synergy we that functionally probes the intrinsic apoptotic pathway, consistently demonstrate herein. Development of a could therefore prove to be a useful predictive biomarker clinical trial testing combined alvocidib and venetoclax for this regimen [4, 14, 44, 54, 55]. in AML and high-risk MDS is ongoing. We found that CDK9-selective inhibition correlated strongly with venetoclax potentiation. While the MATERIALS AND METHODS magnitude of venetoclax sensitization with LDC067 was significantly lower, especially in venetoclax-resistant cells, Cells, culture conditions and reagents a strong correlation was observed. We speculate these distinctions are due to differences in potency for CDK9, Primary samples were obtained with informed although inhibition of other CDK isoforms cannot be consent in accordance with Mayo Clinic IRB-approved ruled-out. We did not test the CDK9 inhibitor dinaciclib research protocols and handled according to Good in this study, as previous studies in diffuse large B-cell Clinical Practice. Primary cells were Ficoll-gradient lymphoma (DLBCL) have been reported [56]. While separated, and immediately cultured, or viably frozen for a stronger single-agent potency of dinaciclib versus subsequent short-term culture. Cell lines were obtained alvocidib in DLBCL cell lines was reported, a clear and from ATCC or DSMZ. All cell lines tested negative for comparable potentiation venetoclax upon combined mycoplasma before cryopreservation, and all cell lines dinaciclib was also reported, regardless of the single- were confirmed free of cross-contamination using PCR- agent potency of either dinaciclib or venetoclax in the based DNA fingerprinting. Cells were cultured in RPMI- DLBCL cell lines used. These observations are consistent 1640 containing 10% FBS, 2 mM L-glutamine, 100 IU/ with our findings demonstrating robust synergy regardless mL penicillin, 100 µg/mL streptomycin (Invitrogen) at of venetoclax single-agent activity in AML. CDK4/6 37°C/5% CO2. Compounds were obtained as follows: inhibition with palbociclib did not generally potentiate Alvocidib (Tolero Pharmaceuticals), venetoclax/ABT-199 www.impactjournals.com/oncotarget 107214 Oncotarget (ChemieTek), ABT-737, LDC067, palbociclib, seliciclib 96 hours using a Cytation3 plate reader (BioTek). Nine and Ro-3306 (SelleckChem), LY2857785 (MedChem doses of venetoclax titrated specifically for venetoclax- Express), NU6102 and JQ1 (Cayman Chemical). resistant or -sensitive cells, or ABT-737, were combined with six doses of the second drug evaluated, yielding 54 possible combinations, each evaluated in quadruplicate Drug dose response assays and CalcuSyn analysis for every experiment. For primary cells, where venetoclax Cells were plated in 384-well plates (Greiner Bio- sensitivity was unknown a priori, venetoclax was serially One) at 1000 cells/well for cell lines, or 2000 cells/ diluted 5-fold (5.0 µM to 0.013 nM), which yielded at least well for primary samples, and dosed simultaneously for three doses for CalcuSyn analysis, version 2.1 (Biosoft) combinations. Relative cell number (expressed as % [58, 59]. Prism version 5.03 (Prism Software Corporation) viability) was measured with CellTiter-Glo (Promega) at was used to calculate EC values. Figure 4: Pharmacological dissection of CDK isoforms contributing to venetoclax sensitization. (A) single-agent drug dose response assays were assessed with duplicate biological experiments, each with four technical replicates, in a panel of eleven AML cell lines, and each EC value is plotted versus the EC value of alvocidib. Data represent average ± STD. Regression analysis was 50 50 used to determine R and P values. (B and C), biological duplicate combination drug dose response assays, each with four technical replicates for every dose/dose combination evaluated, were performed with venetoclax and each inhibitor, and maximal venetoclax EC fold-sensitization (B) and penultimate EC fold-sensitization (C) plotted for each inhibitor. Data represent average ± STD. www.impactjournals.com/oncotarget 107215 Oncotarget Aldrich) added to [5 µg/mL]. After incubating 15 minutes, Flow cytometry cells were analyzed on LSR Fortessa (BD Biosciences). Cells were seeded at 2E5 cells/mL and incubated 24 hours with appropriate CDK inhibitor (80 nM alvocidib, Western blots 17.5 µM seliciclib, or 2.5 or 20 µM NU6102) with or without 2.5 or 10 nM venetoclax for venetoclax-sensitive After treatment with alvocidib and/or venetoclax at cells, or 0.25 or 1 µM venetoclax for venetoclax -resistant indicated doses/times, protein was harvested with lysis buffer cells. Cells were washed with ice-cold PBS, suspended in (Cell Signaling; #9803) containing 1 mM PMSF. Lysates binding buffer at 1E6 cells/mL, Annexin V antibody added were quantified by BCA (Pierce/Thermo Fisher), 26–30 µg at [1:20] (BD Biosciences) and propidium iodide (Sigma total protein resolved by 4–15% SDS-PAGE, transferred www.impactjournals.com/oncotarget 107216 Oncotarget Figure 5: Alvocidib synergizes with venetoclax in short-term ex vivo cultures of AML patient samples and in a mouse model. (A) twelve primary samples from bone aspirates of AML patients were treated with venetoclax and alvocidib in 96 hour culture, single-agent venetoclax and alvocidib EC50 values are plotted for each sample. Data represent average ± STD. (B) Combination Index (CI) values as an expression of synergy and Fraction Affected (FA) as an expression of fractional cell number (1.0 = maximal effect, while 0 = no effect) are shown for the combination of 40 nM venetoclax and 80 nM alvocidib as determined with CalcuSyn software. Dashed lines demark CI ranges: Strong synergy (< 0.3), synergy (0.3 to 0.7), moderate synergy (0.7 to 0.85), slight synergy (0.85 to 0.9), and additive (0.9 to 1.0). (C) untreated cell pellets were available for eleven of the fourteen primary samples analyzed in drug response assays, from which lysates were prepared for quantification of BCL-2 family proteins by western blot. ( D) primary samples #6 and #10 after treatment with 20 nM venetoclax and 80 nM alvocidib, alone and in combination, for quantification of BIM, MCL-1 and BCL-2 by western blot. ( E) primary samples #13 and #14 after treatment with 20 nM venetoclax and 80 nM alvocidib, alone and in combination, for quantification of BIM, MCL-1 and pMCL-1-Thr163 by western blot. For 5C–5E densitometry values calculated with Image J software are normalized to β-actin load controls. For ex vivo studies, primary sample material availability facilitated a single drug dose response assay for each sample, and a single lysate preparation for the indicated samples. www.impactjournals.com/oncotarget 107217 Oncotarget to PVDF membranes (80V/90 minute wet-transfer), and Small-interfering RNA transfection blocked with 5% non-fat dry milk before primary antibody siRNA were purchased from Qiagen. For target incubation, 4°C/overnight. MCL-1 (#4527), pMCL-1-Thr163 protein knock-down validation, 0.04 nmol siRNA was mixed (#1476), BCL-2 (#2872), BCL-XL (#2762), BIM (#2819), with 750 µL RNAiMax transfection reagent (ThermoFisher) PUMA (#4976), β-tubulin (#2128) (Cell Signaling), NBK diluted in OPTI-MEM (Invitrogen) (dilutions optimized (sc-365625; Santa Cruz Biotechnology), NOXA (ab140129; per cell line and reagent batch), after 30 minutes 2.6 to Abcam), and β-actin (A00702; GenScript). Densitometry was 4.3E5 cells (optimized per cell line) were added in 1 mL performed with Image J software (www.imagej.nih.gov). unsupplemented RPMI-1640, spiked with FBS to [2.2%], and harvested at 48 hours. For siRNA drug dose response Lentivirus construction assays, siRNA were pre-printed on 384-well plates, OPTI- MEM-diluted transfection reagent was added at 20 µL/ BCL-XL (cDNA clone ID 2823498, GE well, after 30 minutes 1000–2000 cells/well were added Dharmacon) was amplified by PCR and cloned into in 20 µL of unsupplemented RPMI-1640, and spiked with plasmid pSC11CMVFlag, with a CMV promoter and FBS to [2.2%]. For BIM siDDRs, plates were incubated for 5′3Xflag tag. BCL-XL was then cloned into lentiviral 24 hours prior to simultaneous dosing of venetoclax and vector pWPI (Addgene plasmid #12254) resulting in alvocidib and read 48 hours later. For BCL-XL siDDRs, plasmid pWPIS6FBCL-XL. Plasmid pWPIS6FBCL- plates were incubated for 48 hours prior to simultaneous XL, together with packaging plasmid psPAX2 (plasmid addition of venetoclax or ABT-737 and alvocidib and read #12660) and pMG2.g (plasmid #12259) were transfected 48 hours later. into 293T cells and rescued into lentivirus. Mouse studies Quantitative RT-PCR Animal studies were reviewed and approved for MicroRNAs were isolated using Directzol RNA ethical consideration by an internal review committee mini-prep kit (Zymo Research). Reverse transcription was at Tolero Pharmaceuticals, Inc. OCI-AML3 cells were performed using TaqMan MicroRNA RT Kit, resulting transplanted into 6–8 week old, female, hairless outbred cDNA underwent pre-amplification with TaqMan PreAmp scid hr SCID mice (Crl:SHO-Prkdc Hr ) (18–26 g, strain Master Mix, and RTqPCR was performed with TaqMan code 474, Charles River Laboratories). Study animals MicroRNA assays and TaqMan Universal Master Mix II were housed under standard conditions, and given containing UNG (Life Technologies). Figure 6: Alvocidib + venetoclax is an active regimen in the venetoclax-resistant AML xenograft (OCI-AML3) model. Athymic nude mice were injected subcutaneously in the hind flank with OCI-AML3 cells. When tumors reached approximately 200 mm3, mice were randomized and then treated with vehicle, 100 mg/kg venetoclax, 2.5 mg/kg alvocidib, or a combination of the two drugs. Mice were dosed daily by oral gavage (venetoclax) or intraperitoneal injection (alvocidib). Tumor volumes (A) and body weights (B) were measured and recorded twice weekly. Tumor volumes and body weights shown are mean ± SEM of eight animals per treatment cohort. www.impactjournals.com/oncotarget 107218 Oncotarget food and water ad libitum. 1E7 OCI-AML3 cells were REFERENCES suspended in serum-free media, mixed 1:1 in matrigel (Corning), and injected subcutaneously into the hind 1. Roberts AW, Davids MS, Pagel JM, Kahl BS, Puvvada flank. Randomization and treatment was initiated once SD, Gerecitano JF, Kipps TJ, Anderson MA, Brown JR, tumor volumes had reached approximately 200 mm . Gressick L, Wong S, Dunbar M, Zhu M, et al. Targeting Alvocidib was formulated in saline, while venetoclax BCL2 with Venetoclax in Relapsed Chronic Lymphocytic was formulated in Phosal 50PG: PEG400: Ethanol. Leukemia. N Engl J Med. 2016; 374:311–22. https://doi. Mice were administered vehicle or drug treatment by org/10.1056/NEJMoa1513257. intraperitoneal injection, (qdx5)x3 with alvocidib (2.5 mg/ 2. Kaufmann SH, Karp JE, Svingen PA, Krajewski S, Burke kg), by oral gavage (qdx5)x3 with venetoclax (100 mg/ PJ, Gore SD, Reed JC. Elevated expression of the apoptotic kg), or both alvocidib and venetoclax. Tumor length and regulator Mcl-1 at the time of leukemic relapse. Blood. width measurements were made with digital calipers, and 1998; 91:991–1000. tumor volumes were calculated as follows: length x width 3. van Stijn A, van der Pol MA, Kok A, Bontje PM, x width/2, where length was the longest diameter. Tumor Roemen GM, Beelen RH, Ossenkoppele GJ, Schuurhuis volume and body weights were measured and recorded GJ. Differences between the CD34+ and CD34- blast twice weekly. compartments in apoptosis resistance in acute myeloid leukemia. Haematologica. 2003; 88:497–508. Statistics 4. Vo TT, Ryan J, Carrasco R, Neuberg D, Rossi DJ, Stone RM, Deangelo DJ, Frattini MG, Letai A. Relative mitochondrial P values were calculated using two-tailed Student’s priming of myeloblasts and normal HSCs determines t-test. For correlative comparisons of drug EC values, or chemotherapeutic success in AML. Cell. 2012; 151:344–55. drug activity with protein level, regression analysis was used 2 https://doi.org/10.1016/j.cell.2012.08.038. to calculate R and corresponding p values. R values shown 5. Lagadinou ED, Sach A, Callahan K, Rossi RM, Neering in Figure 1B were calculated with CalcuSyn software. SJ, Minhajuddin M, Ashton JM, Pei S, Grose V, O’Dwyer KM, Liesveld JL, Brookes PS, Becker MW, et al. BCL-2 Author contributions inhibition targets oxidative phosphorylation and selectively eradicates quiescent human leukemia stem cells. Cell J.M.B. performed research, designed experiments, Stem Cell. 2013; 12:329–41. https://doi.org/10.1016/j. analyzed and interpreted data, and wrote the manuscript, stem.2012.12.013. C.J.W. A.S.J, S.L.W., and R.T. performed research, 6. Beurlet S, Omidvar N, Gorombei P, Krief P, Le Pogam C, designed experiments, and analyzed and interpreted Setterblad N, de la Grange P, Leboeuf C, Janin A, Noguera data. C.X.S. and A.K.S. assisted with lentiviral construct ME, Hervatin F, Sarda-Mantel L, Konopleva M, et al. BCL- preparation. N.H. and D.D. performed experiments 2 inhibition with ABT-737 prolongs survival in an NRAS/ and analyzed data, W.K., H.H., K.K.S., Y.S.L., and P.P. BCL-2 mouse model of AML by targeting primitive LSK performed experiments. J.M.B., C.J.W., A.S-J., S.W., and progenitor cells. Blood. 2013; 122:2864–76. https://doi. D.J.B., R.A.M, and R.T. conceptualized the study. All org/10.1182/blood-2012-07-445635. authors read and approved the final manuscript. 7. Pan R, Hogdal LJ, Benito JM, Bucci D, Han L, Borthakur G, Cortes J, DeAngelo DJ, Debose L, Mu H, Dohner H, Gaidzik ACKNOWLEDGMENTS VI, Galinsky I, et al. Selective BCL-2 inhibition by ABT-199 causes on-target cell death in acute myeloid leukemia. Cancer Vector pWPI, psPAX2 and pMG2.g were generous Discov. 2014; 4:362–75. https://doi.org/10.1158/2159-8290. gifts from Didier Trono. CD-13-0609. 8. Raza A, Gezer S, Mundle S, Gao XZ, Alvi S, Borok R, Rifkin S, Iftikhar A, Shetty V, Parcharidou A. Apoptosis CONFLICTS OF INTEREST in bone marrow biopsy samples involving stromal and hematopoietic cells in 50 patients with myelodysplastic C.J.W., W.K., H.H., K.K.S., Y.L., P.P., A.S-J., syndromes. Blood. 1995; 86:268–76. S.W., D.J.B. and S.L.W. are employed by Tolero Pharmaceuticals. R.T. is a scientific advisor for Tolero 9. Parker JE, Mufti GJ, Rasool F, Mijovic A, Devereux S, Pharmaceuticals. The remaining authors declare no Pagliuca A. The role of apoptosis, proliferation, and the Bcl- competing interests. 2-related proteins in the myelodysplastic syndromes and acute myeloid leukemia secondary to MDS. Blood. 2000; 96:3932–8. FUNDING 10. Invernizzi R, Pecci A, Bellotti L, Ascari E. Expression This work was supported by Mayo Clinic of p53, bcl-2 and ras oncoproteins and apoptosis levels institutional support of JMB and RT. in acute leukaemias and myelodysplastic syndromes. www.impactjournals.com/oncotarget 107219 Oncotarget Leuk Lymphoma. 2001; 42:481–9. https://doi. access of MCL-1 to its E3 ligase. FEBS Lett. 2005; org/10.3109/10428190109064605. 579:5603–8. https://doi.org/S0014-5793(05)01135-X. 11. Albitar M, Manshouri T, Shen Y, Liu D, Beran M, 21. Jilg S, Reidel V, Muller-Thomas C, Konig J, Schauwecker Kantarjian HM, Rogers A, Jilani I, Lin CW, Pierce S, J, Hockendorf U, Huberle C, Gorka O, Schmidt B, Burgkart Freireich EJ, Estey EH. Myelodysplastic syndrome is not R, Ruland J, Kolb HJ, Peschel C, et al. Blockade of BCL-2 merely “preleukemia”. Blood. 2002; 100:791–8. proteins efficiently induces apoptosis in progenitor cells of high-risk myelodysplastic syndromes patients. Leukemia. 12. Slape CI, Saw J, Jowett JB, Aplan PD, Strasser A, Jane SM, 2016; 30:112–23. https://doi.org/10.1038/leu.2015.179. Curtis DJ. Inhibition of apoptosis by BCL2 prevents leukemic transformation of a murine myelodysplastic syndrome. Blood. 22. Niu X, Zhao J, Ma J, Xie C, Edwards H, Wang G, Caldwell 2012; 120:2475–83. https://doi.org/blood-2012-05-430736. JT, Xiang S, Zhang X, Chu R, Wang J, Lin H, Taub JW, et al. 13. Konopleva M, Pollyea DA, Potluri J, Chyla BJ, Busman Binding of released Bim to Mcl-1 is a mechanism of intrinsic T, McKeegan E, Salem A, Zhu M, Ricker JL, Blum W, resistance to ABT-199 which can be overcome by combination DiNardo CD, Dunbar M, Kirby R, et al. A Phase 2 Study of with daunorubicin or cytarabine in AML cells. Clin Cancer ABT-199 (GDC-0199) in Patients with Acute Myelogenous Res. 2016. https://doi.org/clincanres.3057.2015. Leukemia (AML). Blood. 2014; 124:118. 23. Bodo J, Zhao X, Durkin L, Souers AJ, Phillips DC, Smith 14. Bogenberger JM, Kornblau SM, Pierceall WE, Lena R, MR, Hsi ED. Acquired resistance to venetoclax (ABT-199) Chow D, Shi CX, Mantei J, Ahmann G, Gonzales IM, in t(14;18) positive lymphoma cells. Oncotarget. 2016; Choudhary A, Valdez R, Camoriano J, Fauble V, et al. 7:70000–10. https://doi.org/10.18632/oncotarget.12132. BCL-2 family proteins as 5-Azacytidine-sensitizing targets 24. Pan R, Ruvolo VR, Wei J, Konopleva M, Reed JC, Pellecchia and determinants of response in myeloid malignancies. M, Andreeff M, Ruvolo PP. Inhibition of Mcl-1 with the Leukemia. 2014; 28:1657–65. pan-Bcl-2 family inhibitor (-)BI97D6 overcomes ABT-737 15. Bogenberger JM, Delman D, Hansen N, Valdez R, Fauble V, resistance in acute myeloid leukemia. Blood. 2015; 126:363– Mesa RA, Tibes R. Ex vivo activity of BCL-2 family inhibitors 72. https://doi.org/10.1182/blood-2014-10-604975. ABT-199 and ABT-737 combined with 5-azacytidine in 25. Chen R, Keating MJ, Gandhi V, Plunkett W. Transcription myeloid malignancies. Leuk Lymphoma. 2015; 56:226–9. inhibition by flavopiridol: mechanism of chronic https://doi.org/10.3109/10428194.2014.910657. lymphocytic leukemia cell death. Blood. 2005; 106:2513–9. 16. Tsao T, Shi Y, Kornblau S, Lu H, Konoplev S, Antony https://doi.org/2005-04-1678. A, Ruvolo V, Qiu YH, Zhang N, Coombes KR, Andreeff 26. Gojo I, Zhang B, Fenton RG. The cyclin-dependent M, Kojima K, Konopleva M. Concomitant inhibition kinase inhibitor flavopiridol induces apoptosis in multiple of DNA methyltransferase and BCL-2 protein function myeloma cells through transcriptional repression and down- synergistically induce mitochondrial apoptosis in acute regulation of Mcl-1. Clin Cancer Res. 2002; 8:3527–38. myelogenous leukemia cells. Ann Hematol. 2012; 91:1861– 27. Kitada S, Zapata JM, Andreeff M, Reed JC. Protein 70. https://doi.org/10.1007/s00277-012-1537-8. kinase inhibitors flavopiridol and 7-hydroxy-staurosporine 17. DiNardo C, Pollyea D, Pratz K, Thirman MJ, Letai A, Frattini down-regulate antiapoptosis proteins in B-cell chronic M, Jonas B, Leverson J, Zhu M, Dunbar M, Falotico N, Kirby lymphocytic leukemia. Blood. 2000; 96:393–7. R, Agarwal S, et al. A Phase 1b Study of Venetoclax (ABT-199/ 28. Chao SH, Price DH. Flavopiridol inactivates P-TEFb and GDC-0199) in Combination with Decitabine or Azacitidine in blocks most RNA polymerase II transcription in vivo. J Treatment-Naive Patients with Acute Myelogenous Leukemia Biol Chem. 2001; 276:31793–9. https://doi.org/10.1074/ Who Are ≥ to 65 Years and Not Eligible for Standard Induction jbc.M102306200. Therapy. Blood. 2015; 126:327. 29. Wada T, Takagi T, Yamaguchi Y, Watanabe D, Handa H. 18. De Biasio A, Vrana JA, Zhou P, Qian L, Bieszczad CK, Evidence that P-TEFb alleviates the negative effect of DSIF Braley KE, Domina AM, Weintraub SJ, Neveu JM, Lane on RNA polymerase II-dependent transcription in vitro. WS, Craig RW. N-terminal truncation of antiapoptotic EMBO J. 1998; 17:7395–403. https://doi.org/10.1093/ MCL1, but not G2/M-induced phosphorylation, is emboj/17.24.7395. associated with stabilization and abundant expression in tumor cells. J Biol Chem. 2007; 282:23919–36. https://doi. 30. Pallis M, Burrows F, Whittall A, Boddy N, Seedhouse C, org/M700938200. Russell N. Efficacy of RNA polymerase II inhibitors in targeting dormant leukaemia cells. BMC Pharmacol Toxicol. 19. Perciavalle RM, Stewart DP, Koss B, Lynch J, Milasta S, 2013; 14:32. https://doi.org/10.1186/2050-6511-14-32. Bathina M, Temirov J, Cleland MM, Pelletier S, Schuetz JD, Youle RJ, Green DR, Opferman JT. Anti-apoptotic 31. Chen S, Dai Y, Pei XY, Myers J, Wang L, Kramer LB, MCL-1 localizes to the mitochondrial matrix and couples Garnett M, Schwartz DM, Su F, Simmons GL, Richey JD, mitochondrial fusion to respiration. Nat Cell Biol. 2012; Larsen DG, Dent P, et al. CDK inhibitors upregulate BH3- 14:575–83. https://doi.org/10.1038/ncb2488. ncb2488. only proteins to sensitize human myeloma cells to BH3 20. Warr MR, Acoca S, Liu Z, Germain M, Watson M, mimetic therapies. Cancer Res. 2012; 72:4225–37. https:// Blanchette M, Wing SS, Shore GC. BH3-ligand regulates doi.org/10.1158/0008-5472.CAN-12-1118. www.impactjournals.com/oncotarget 107220 Oncotarget 32. Harley ME, Allan LA, Sanderson HS, Clarke PR. 42. Vassilev LT, Tovar C, Chen S, Knezevic D, Zhao X, Sun Phosphorylation of Mcl-1 by CDK1-cyclin B1 initiates its H, Heimbrook DC, Chen L. Selective small-molecule Cdc20-dependent destruction during mitotic arrest. EMBO J. inhibitor reveals critical mitotic functions of human CDK1. 2010; 29:2407–20. https://doi.org/10.1038/emboj.2010.112. Proc Natl Acad Sci U S A. 2006; 103:10660–5. https://doi. org/0600447103. 33. Inuzuka H, Shaik S, Onoyama I, Gao D, Tseng A, Maser RS, Zhai B, Wan L, Gutierrez A, Lau AW, Xiao Y, Christie AL, 43. Kojima K, Shimanuki M, Shikami M, Andreeff M, Aster J, et al. SCF(FBW7) regulates cellular apoptosis by Nakakuma H. Cyclin-dependent kinase 1 inhibitor RO-3306 targeting MCL1 for ubiquitylation and destruction. Nature. enhances p53-mediated Bax activation and mitochondrial 2011; 471:104–9. https://doi.org/10.1038/nature09732. apoptosis in AML. Cancer Sci. 2009; 100:1128–36. https:// nature09732. doi.org/10.1111/j.1349-7006.2009.01150.x. 44. Pallis M, Burrows F, Ryan J, Grundy M, Seedhouse 34. Choudhary GS, Tat TT, Misra S, Hill BT, Smith MR, C, Abdul-Aziz A, Montero J, Letai A, Russell N. Almasan A, Mazumder S. Cyclin E/Cdk2-dependent Complementary dynamic BH3 profiles predict co- phosphorylation of Mcl-1 determines its stability and operativity between the multi-kinase inhibitor TG02 and the cellular sensitivity to BH3 mimetics. Oncotarget. 2015; BH3 mimetic ABT-199 in acute myeloid leukaemia cells. 6:16912–16925. https://doi.org/10.18632/oncotarget.4857. Oncotarget. 2017; 8:16220–32. https://doi.org/10.18632/ oncotarget.8742. 35. Ventura A, Young AG, Winslow MM, Lintault L, Meissner A, Erkeland SJ, Newman J, Bronson RT, Crowley D, Stone 45. Polier G, Giaisi M, Kohler R, Muller WW, Lutz C, Buss EC, JR, Jaenisch R, Sharp PA, Jacks T. Targeted deletion reveals Krammer PH, Li-Weber M. Targeting CDK9 by wogonin and essential and overlapping functions of the miR-17 through related natural flavones potentiates the anti-cancer efficacy 92 family of miRNA clusters. Cell. 2008; 132:875–86. of the Bcl-2 family inhibitor ABT-263. Int J Cancer. 2015; https://doi.org/10.1016/j.cell.2008.02.019. 136:688–98. https://doi.org/10.1002/ijc.29009. 36. Xu Z, Sharp PP, Yao Y, Segal D, Ang CH, Khaw SL, Aubrey 46. Ma Y, Cress WD, Haura EB. Flavopiridol-induced apoptosis BJ, Gong J, Kelly GL, Herold MJ, Strasser A, Roberts AW, is mediated through up-regulation of E2F1 and repression Alexander WS, et al. BET inhibition represses miR17–92 of Mcl-1. Mol Cancer Ther. 2003; 2:73–81. to drive BIM-initiated apoptosis of normal and transformed 47. Phillips DC, Xiao Y, Lam LT, Litvinovich E, Roberts- hematopoietic cells. Leukemia. 2016; 30:1531–41. https:// Rapp L, Souers AJ, Leverson JD. Loss in MCL-1 function doi.org/10.1038/leu.2016.52. sensitizes non-Hodgkin’s lymphoma cell lines to the BCL- 37. Albert TK, Rigault C, Eickhoff J, Baumgart K, Antrecht 2-selective inhibitor venetoclax (ABT-199). Blood Cancer C, Klebl B, Mittler G, Meisterernst M. Characterization of J. 2016; 6:e403. https://doi.org/10.1038/bcj.2016.12. molecular and cellular functions of the cyclin-dependent 48. Teh TC, Nguyen NY, Moujalled DM, Segal D, Pomilio G, kinase CDK9 using a novel specific inhibitor. Br J Pharmacol. Rijal S, Jabbour A, Cummins K, Lackovic K, Blombery 2014; 171:55–68. https://doi.org/10.1111/bph.12408. P, Thompson E, Ekert PG, Lessene G, et al. Enhancing 38. Karaman MW, Herrgard S, Treiber DK, Gallant P, Atteridge venetoclax activity in acute myeloid leukemia by co- CE, Campbell BT, Chan KW, Ciceri P, Davis MI, Edeen PT, targeting MCL1. Leukemia. 2017. https://doi.org/10.1038/ Faraoni R, Floyd M, Hunt JP, et al. A quantitative analysis leu.2017.243. of kinase inhibitor selectivity. Nat Biotechnol. 2008; 49. Shamas-Din A, Brahmbhatt H, Leber B, Andrews DW. 26:127–32. https://doi.org/10.1038/nbt1358. BH3-only proteins: Orchestrators of apoptosis. Biochim 39. Meijer L, Borgne A, Mulner O, Chong JP, Blow JJ, Inagaki Biophys Acta. 2011; 1813:508–20. https://doi.org/10.1016/j. N, Inagaki M, Delcros JG, Moulinoux JP. Biochemical bbamcr.2010.11.024. and cellular effects of roscovitine, a potent and selective 50. Chonghaile TN, Roderick JE, Glenfield C, Ryan J, Sallan inhibitor of the cyclin-dependent kinases cdc2, cdk2 and SE, Silverman LB, Loh ML, Hunger SP, Wood B, DeAngelo cdk5. Eur J Biochem. 1997; 243:527–36. DJ, Stone R, Harris M, Gutierrez A, et al. Maturation 40. Yin T, Lallena MJ, Kreklau EL, Fales KR, Carballares stage of T-cell acute lymphoblastic leukemia determines S, Torrres R, Wishart GN, Ajamie RT, Cronier DM, BCL-2 versus BCL-XL dependence and sensitivity to Iversen PW, Meier TI, Foreman RT, Zeckner D, et al. A ABT-199. Cancer Discov. 2014; 4:1074–87. https://doi. novel CDK9 inhibitor shows potent antitumor efficacy in org/10.1158/2159-8290.CD-14-0353. preclinical hematologic tumor models. Mol Cancer Ther. 51. Tahir SK, Smith ML, Hessler P, Rapp LR, Idler KB, Park CH, 2014; 13:1442–56. https://doi.org/10.1158/1535-7163. Leverson JD, Lam LT. Potential mechanisms of resistance MCT-13-0849. to venetoclax and strategies to circumvent it. BMC Cancer. 41. Hardcastle IR, Arris CE, Bentley J, Boyle FT, Chen Y, 2017; 17:399. https://doi.org/10.1186/s12885-017-3383-5. Curtin NJ, Endicott JA, Gibson AE, Golding BT, Griffin RJ, 52. Lin KH, Winter PS, Xie A, Roth C, Martz CA, Stein EM, Jewsbury P, Menyerol J, Mesguiche V, et al. N2-substituted Anderson GR, Tingley JP, Wood KC. Targeting MCL-1/ O6-cyclohexylmethylguanine derivatives: potent inhibitors BCL-XL Forestalls the Acquisition of Resistance to ABT- of cyclin-dependent kinases 1 and 2. J Med Chem. 2004; 199 in Acute Myeloid Leukemia. Sci Rep. 2016; 6:27696. 47:3710–22. https://doi.org/10.1021/jm0311442. https://doi.org/10.1038/srep27696. www.impactjournals.com/oncotarget 107221 Oncotarget 53. Jorgensen TN, McKee A, Wang M, Kushnir E, White J, loss of MCL1. Leukemia. 2015; 29:1702–12. https://doi. Refaeli Y, Kappler JW, Marrack P. Bim and Bcl-2 mutually org/10.1038/leu.2015.99. affect the expression of the other in T cells. J Immunol. 57. Zeidner JF, Foster MC, Blackford AL, Litzow MR, Morris 2007; 179:3417–24. LE, Strickland SA, Lancet JE, Bose P, Levy MY, Tibes 54. Del Gaizo Moore V, Letai A. BH3 profiling--measuring R, Gojo I, Gocke CD, Rosner GL, et al. Randomized integrated function of the mitochondrial apoptotic pathway multicenter phase II study of flavopiridol (alvocidib), to predict cell fate decisions. Cancer Lett. 2013; 332:202–5. cytarabine, and mitoxantrone (FLAM) versus cytarabine/ https://doi.org/10.1016/j.canlet.2011.12.021. daunorubicin (7+3) in newly diagnosed acute myeloid leukemia. Haematologica. 2015; 100:1172–9. https://doi. 55. Pierceall WE, Kornblau SM, Carlson NE, Huang X, org/10.3324/haematol.2015.125849. Blake N, Lena R, Elashoff M, Konopleva M, Cardone MH, Andreeff M. BH3 profiling discriminates response 58. Chou TC, Talalay P. Quantitative analysis of dose-effect to cytarabine-based treatment of acute myelogenous relationships: the combined effects of multiple drugs or leukemia. Mol Cancer Ther. 2013; 12:2940–9. https://doi. enzyme inhibitors. Adv Enzyme Regul. 1984; 22:27–55. org/10.1158/1535-7163.MCT-13-0692. 59. Reynolds CP, Maurer BJ. Evaluating response to antineoplastic drug combinations in tissue culture models. 56. Li L, Pongtornpipat P, Tiutan T, Kendrick SL, Park S, Persky Methods Mol Med. 2005; 110:173–83. https://doi.org/1- DO, Rimsza LM, Puvvada SD, Schatz JH. Synergistic 59259-869-2:173. induction of apoptosis in high-risk DLBCL by BCL2 inhibition with ABT-199 combined with pharmacologic www.impactjournals.com/oncotarget 107222 Oncotarget http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Oncotarget Pubmed Central

Loading next page...
 
/lp/pubmed-central/combined-venetoclax-and-alvocidib-in-acute-myeloid-leukemia-qLE9bRDMgk

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Pubmed Central
Copyright
Copyright: © 2017 Bogenberger et al.
ISSN
1949-2553
eISSN
1949-2553
DOI
10.18632/oncotarget.22284
Publisher site
See Article on Publisher Site

Abstract

www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 63), pp: 107206-107222 Research Paper 1,* 2,* 1 1 James Bogenberger , Clifford Whatcott , Nanna Hansen , Devora Delman , 1 2 2 2 2 Chang-Xin Shi , Wontak Kim , Hillary Haws , Katherine Soh , Ye Sol Lee , Peter 2 2 2 1 2 Peterson , Adam Siddiqui-Jain , Steven Weitman , Keith Stewart , David Bearss , 1 2 1,3 Ruben Mesa , Steven Warner and Raoul Tibes Mayo Clinic, Scottsdale, AZ, USA Tolero Pharmaceuticals, Lehi, UT, USA NYU School of Medicine, New York, NY, USA These authors contributed equally to this work Correspondence to: Raoul Tibes, email: raoul.tibes@nyumc.org James Bogenberger, email: bogenberger.james@mayo.edu Keywords: acute myeloid leukemia (AML); alvocidib/flavopiridol; venetoclax/ABT-199; BCL-2 Received: July 22, 2017 Accepted: October 10, 2017 Published: November 03, 2017 Copyright: Bogenberger et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ABSTRACT More effective treatment options for elderly acute myeloid leukemia (AML) patients are needed as only 25–50% of patients respond to standard-of-care therapies, response duration is typically short, and disease progression is inevitable even with some novel therapies and ongoing clinical trials. Anti-apoptotic BCL-2 family inhibitors, such as venetoclax, are promising therapies for AML. Nonetheless, resistance is emerging. We demonstrate that venetoclax combined with cyclin-dependent kinase (CDK) inhibitor alvocidib is potently synergistic in venetoclax-sensitive and -resistant AML models in vitro, ex vivo and in vivo. Alvocidib decreased MCL-1, and/or increased pro-apoptotic proteins such as BIM or NOXA, often synergistically with venetoclax. Over-expression of BCL-XL diminished synergy, while knock-down of BIM almost entirely abrogated synergy, demonstrating that the synergistic interaction between alvocidib and venetoclax is primarily dependent on intrinsic apoptosis. CDK9 inhibition predominantly mediated venetoclax sensitization, while CDK4/6 inhibition with palbociclib did not potentiate venetoclax activity. Combined, venetoclax and alvocidib modulate the balance of BCL-2 family proteins through complementary, yet variable mechanisms favoring apoptosis, highlighting this combination as a promising therapy for AML or high-risk MDS with the capacity to overcome intrinsic apoptosis mechanisms of resistance. These results support clinical testing of combined venetoclax and alvocidib for the treatment of AML and advanced MDS. and apoptotic resistance [8–12]. Clinical activity of INTRODUCTION venetoclax monotherapy is modest in AML [13], which may partially be explained by the genetic heterogeneity Targeting of intrinsic apoptosis via inhibition of AML, and more importantly by the heterogeneous of B-Cell Lymphoma 2 (BCL-2) family proteins is and concurrent expression of multiple anti-apoptotic a promising therapeutic strategy. BCL-2-selective BCL-2 family proteins in AML, and the selectivity of inhibitor venetoclax (ABT-199) has demonstrated clinical venetoclax for BCL-2 [14]. Consequently, therapeutics activity in chronic lymphocytic leukemia (CLL) and targeting complementary BCL-2 family proteins may has recently received FDA-approval for some relapsed increase the efficacy of venetoclax. Preclinical data CLL patients [1]. Importantly, BCL-2 family proteins demonstrate that BCL-2 family inhibition synergizes with are also therapeutic targets in acute myeloid leukemia hypomethylating agents (HMAs) [14–16] and, although (AML) [2–7] and high-risk myelodysplastic syndrome the precise mechanism of synergy remains incompletely (MDS) where increases in anti-apoptotic BCL-2 characterized, HMAs may down-regulate MCL-1 [16]. family proteins are associated with disease progression www.impactjournals.com/oncotarget 107206 Oncotarget Initial clinical trials examining venetoclax with HMAs in of venetoclax -resistant cells from µM to low nM doses previously untreated elderly AML patients have reported (30 to 50 nM) similar to that of sensitive cells/primary impressive overall response rates of > 75% [17], although samples. Fold-sensitization at 80 nM alvocidib was 14.5 clinical resistance is emerging. Putative mechanisms of ± 0.8 (p = 0.028) and 10.1 ± 3.4 (p = 0.17) in MOLM- venetoclax resistance include compensatory up-regulation, 13 and MV4-11, respectively. At clinically-achievable increased stability, or altered function of MCL-1 [18–20]. plasma concentrations of 80 and 160 nM alvocidib, MCL-1 plays a role in intrinsic resistance to BCL-2 synergy (expressed as Combination Index (CI) values) inhibitors by alternatively sequestering BIM dissociated was observed with all clinically-achievable doses of from BCL-2 by BCL-2 inhibitors [21–23]. MCL-1 venetoclax tested, in all cell lines examined (Figure 1B). inhibition or genetic knock-down is known to potentiate To confirm that synergistic effects of combined venetoclax BCL-2 inhibitors [7, 14, 24]. The cyclin-dependent kinase and alvocidib culminate in increased apoptosis, as opposed (CDK) inhibitor alvocidib (flavopiridol) reduces MCL-1 in to only cytostatic effects from putatively inhibiting hematologic malignancies [25–27]. Alvocidib is a potent cell cycle CDKs, we analyzed Annexin V levels and inhibitor of CDK9, the core catalytic component of the propidium iodide permeability by flow cytometry. In all P-TEFb complex [28–30]. Transcriptional repression of cells examined, we observed an increase in early and late short half-life oncoproteins, such as MCL-1, can potently apoptotic cells in response to the combination beyond induce apoptosis; however, additional anti-apoptotic the additive effects of either single-agent (Figure 1C). In proteins, such as BCL-2, can functionally counter this parallel, we assessed cell cycle distributions and found that repression by sequestering pro-apoptotic BH3-only 80 nM alvocidib resulted only in a moderate proportional proteins or blocking BAX/BAK dimerization. Increased increase in G1, with corresponding decreases in S and G2, pro-apoptotic BH3-only proteins such as BIM have also in three of four cell lines analyzed (median increase 28 ± been observed in response to alvocidib treatment [31]. We 5%); however, 80 nM alvocidib did not significantly alter hypothesized that alvocidib would synergize with BCL- cell cycle distribution in OCI-AML3 (Supplementary 2 inhibitor venetoclax in AML through complementary Figure 1). mechanisms, namely decreased MCL-1 and increased BIM, to modulate the overall balance of anti- and pro- Correlation of BCL-2 family proteins with apoptotic BCL-2 family proteins in favor of apoptosis alvocidib/venetoclax activity induction. To determine whether anti-apoptotic BCL-2 family RESULTS members correlate with single-agent alvocidib anti- leukemic activity, we initially quantified baseline protein Alvocidib potentiates venetoclax anti-leukemic levels of BCL-2, BCL-XL and MCL-1 in untreated cells. activity in both venetoclax-sensitive and – MCL-1 protein was relatively homogenous, differing by a median of 1.7 ± 0.8-fold. In contrast, BCL-2 levels were resistant AML cells highly variable, differing by 158-fold between the lowest To assess the potential for synergy between and highest expressing cells. BCL-XL protein expression venetoclax and alvocidib in venetoclax -sensitive AML was also variable spanning an 11.6-fold range (Figure 2A). cells in vitro, we selected MOLM-13 and MV4-11 cell Relative protein levels were then plotted against single- lines, which have low single-agent venetoclax half- agent alvocidib EC values from cell viability assays. maximal effective concentration (EC ) values (in assays BCL-2 levels did not correlate with alvocidib activity, assessing relative cell number) of 9.0 ± 1.6 and 7.8 ± 2.1 while BCL-XL levels positively correlated, and MCL- nM, respectively. These low nM venetoclax EC values 1 levels negatively correlated with alvocidib activity are consistent with values for primary AML samples (Figure 2B). BCL-2, BCL-XL and MCL-1 protein levels sensitive to venetoclax [7]. THP-1 and OCI-AML3 did not significantly correlate with venetoclax single-agent were selected to model venetoclax -resistant AML cells, activity in this panel of AML cell lines (Figure 2C). exhibiting higher single-agent venetoclax EC values of In contrast to the potent alvocidib sensitization 0.9 ± 0.2 and 2.3 ± 0.4 µM, respectively, 100- to 295- resulting in clinically meaningful 30–50 nM venetoclax fold greater than venetoclax -sensitive cells. Addition of EC values observed in venetoclax-resistant cells with Low alvocidib resulted in potent dose-dependent reduction low levels of BCL-XL (BCL-XL ), venetoclax-resistant High of venetoclax EC values in both venetoclax -sensitive cells with high levels of BCL-XL (BCL-XL ) exhibited and –resistant cells (Figure 1A). venetoclax EC fold- EC values that were not clinically meaningful ranging 50 50 sensitization by alvocidib was greatest in venetoclax from 1.7 to 4.7 µM (median 3.4 ± 1.2 µM). ABT-737 -resistant cells, with 80 nM alvocidib resulting in an 18.3 (navitoclax), a drug that inhibits BCL-XL and BCL-W ± 0.1 (p = 0.0039) and 77.7 ± 1.2 (p = 0.0069) venetoclax in addition to BCL-2, was used to examine the role of EC fold-shift for THP-1 and OCI-AML3, respectively. these additional BCL-2 family members in alvocidib Importantly, alvocidib shifted the absolute EC values synergy. Combined with alvocidib, EC values for ABT- 50 50 www.impactjournals.com/oncotarget 107207 Oncotarget Figure 1: Alvocidib potentiates venetoclax anti-leukemic activity in both venetoclax -sensitive and –resistant AML cells. (A–B) combination drug dose response assays with venetoclax and alvocidib in vitro were assessed in duplicate biological experiments, each containing four technical replicate data points for every dose/dose combination analyzed. Data represent mean ± SEM. The indicated AML cell lines were dosed with venetoclax or alvocidib as single-agent, and in combination, and incubated for 96 hours before determining relative cell number with ATP-based reagent CellTiter Glo. (A) leftward shifts toward lower doses of venetoclax demonstrate dose- dependent venetoclax fold-sensitization by alvocidib. (B) Combination Index (CI) values were calculated with CalcuSyn Software, and are shown for distinct dose combinations of venetoclax and alvocidib. CalcuSyn R values corresponding to single-agent dose curves are shown for each cell line below CI value tables. (C) AML cell lines were treated for 24 hours with 80 nM alvocidib, and a low dose or high dose of venetoclax, each alone and in combination, prior to harvesting for flow cytometry quantification of annexin V and propidium iodide * † permeability as a measurement of apoptosis. For venetoclax -sensitive cells MOLM-13 and MV4-11, 2.5 and 10 nM were used, while * † for venetoclax -resistant cell lines THP-1 and OCI-AML3, 0.25 and 1 µM venetoclax were used. Quantification from a representative experiment is shown graphically, and apoptosis results were confirmed in biological replicate experiments using one venetoclax -sensitive cell line (MOLM-13) and one venetoclax -resistant cell line (THP-1). www.impactjournals.com/oncotarget 107208 Oncotarget High Low 737 were lower than that of venetoclax in BCL-XL 13) and venetoclax-resistant (THP-1) BCL-XL cells, cells (Supplementary Figure 2A, upper panel). EC which rendered these cells more resistant to single-agent fold-sensitization by alvocidib was greater for ABT-737 venetoclax (from nM to µM doses) (Supplementary Figure High High than venetoclax in BCL-XL cells (Supplementary 2E). Similar to BCL-XL cells, although sensitization Low Figure 2A, lower panel). In contrast, BCL-XL cells was observed in BCL-XL over-expressing cells, the (both venetoclax-sensitive and -resistant) were equally, resultant venetoclax EC values were no longer clinically or more, sensitive to single-agent venetoclax than ABT- meaningful (Supplementary Figure 2F). 737 (Supplementary Figure 2B). Knock-down of BCL- XL significantly reduced venetoclax EC values as BCL-2 family proteins altered by alvocidib compared to non-silencing (NS) siRNA (Supplementary Figure 2C–2D). BCL-XL was over-expressed with a To examine effects of alvocidib on the BCL-2 family lentiviral construct in both venetoclax-sensitive (MOLM- protein landscape, we treated venetoclax-sensitive and Figure 2: Correlation of BCL-2 family proteins with alvocidib and venetoclax activity. (A) lysates were prepared from untreated AML cell lines, and levels of the indicated anti-apoptotic BCL-2 family proteins measured by western blot. Image J densitometry software was used to quantify bands and values normalized to β-tubulin. (B and C), relative protein levels of MCL-1, BCL-XL and BCL-2 were plotted against alvocidib EC values (B) or venetoclax EC values (C) determined from duplicate biological experiments. Regression 50 50 analysis was used to determine R and P values. www.impactjournals.com/oncotarget 107209 Oncotarget –resistant AML cells with ascending doses and assessed MOMP. Thus, we sought to evaluate the functional role total protein. Consistent with previous reports in other of BIM, specifically with regard to synergy between hematologic malignancies [25–27], we observed a dose- alvocidib and venetoclax. Knock-down of BIM prior to dependent reduction of ~40 kD anti-apoptotic isoform alvocidib and venetoclax combination treatment potently MCL-1 in three of four AML cell lines. Increased abrogated synergy compared to NS siRNA. BIM knock- Long BIM was observed in all four cell lines in response to down abrogated 97% (p = 0.0066) and 85% (p = 0.0075) alvocidib, although the increase was more pronounced of the venetoclax fold-sensitization induced by 80 nM in venetoclax-sensitive cells. MOLM-13 was unique alvocidib in OCI-AML3 and THP-1 cells, respectively. in that we did not observe a reduction in MCL-1 , Mean combination venetoclax EC values were 1.8 and Long 50 but instead observed a dose-dependent induction of all 1.3 µM with BIM siRNA, versus 0.05 and 0.14 µM with BH3-only proteins measured (NBK, NOXA, PUMA and NS control siRNA, demonstrating that synergy between BIM). BCL-2 was not significantly changed by alvocidib venetoclax and alvocidib is almost entirely dependent on treatment. Conversely, BCL-XL was variably induced by the intrinsic apoptotic pathway (Figure 3F–3G). alvocidib in all four cell lines (Figure 3A). However, the Low induced level of BCL-XL in BCL-XL MOLM-13 cells Pharmacological dissection of CDK isoforms High was significantly lower than baseline levels in BCL-XL contributing to venetoclax sensitization cells (Figure 3B). MCL-1 protein stability is regulated via complex phosphorylation by multiple kinases including To discern the contributions of different CDK CDK1 and -2 [32, 33]. Signaling via ERK, JNK, GSK3 isoforms toward venetoclax sensitization, we examined and CDK2, converge on Thr-163 phosphorylation of CDK inhibitors with putatively distinct inhibitory profiles MCL-1, which increases MCL-1 stability and thereby (LDC067, palbociclib, LY2857785, seliciclib, NU6102, plays a role in resistance to BCL-2 inhibitors [34]. To and Ro-3306), as well as a BRD4 inhibitor (JQ1) to more investigate whether alvocidib affects MCL-1 protein specifically assess P-TEFb. We compared single-agent stability through putative CDK1/2 inhibition, in addition activity of these compounds in a panel of ten AML cell to MCL-1 transcriptional down-regulation by CDK9 lines with diverse molecular, cytogenetic and lineage inhibition, we measured MCL-1 Thr-163 phosphorylation backgrounds (MDS-L, HL-60, OCI-AML2, OCI-AML3, levels. Alvocidib did not significantly decrease MCL-1 MV4-11, MOLM-13, THP-1, TF-1, HEL, and SET-2). Thr-163 phosphorylation in two of three cell lines where Correlation of each compound with alvocidib was used decreases in total MCL-1 were observed (Figure 3A). to assess pharmacological similarity. We also combined Published studies suggest that the polycistronic microRNA each of these unique CDK inhibitors with venetoclax to (miRNA) miR-17–92 negatively regulates BIM [35, 36], assess interaction. thus we measured five miRNAs derived from the miR-17– In terms of single-agent EC correlation, CDK9- 92 cluster in response to alvocidib treatment. Relative to selective inhibitor LDC067 [37] showed the strongest control miRNAs, all five miRNAs decreased in a dose- and correlation with alvocidib. Palbociclib, a recently FDA- time-dependent manner (Figure 3C). We were surprised to approved CDK4/6-selective inhibitor, did not significantly observe miR-17–92 decreases in THP-1 and OCI-AML3, correlate with alvocidib. Seliciclib (roscovitine), the most where potent increases in BIM protein were not observed well-characterized CDK5 inhibitor which also inhibits at 24 hours. Thus, we performed western blots at 48 CDK2, -7 and -9 with > 100-fold selectivity over CDK4 hours in these two cell lines and observed a more delayed and -6 [38, 39], significantly correlated with alvocidib. increase of BIM (Figure 3D). Next we examined BCL-2 LY2857785 [40], which compared to alvocidib more family proteins after 24 hour combination treatment, as potently inhibits CDK8, shows less potency for CDK5 compared to single-agent and untreated controls. Three of and similarly inhibits CDK9, correlated with alvocidib. four cell lines showed increases in either BIM or NOXA NU6102, a CDK2 and -1 selective inhibitor [41] also with the combination treatment. Surprisingly, we observed significantly correlated with alvocidib. Ro-3306, a CDK1- variable increases in MCL-1, BCL-2, and BCL-XL with selective inhibitor with 10-fold greater selectivity over combination treatment in all cell lines. The observed CDK2 and 50-fold over CDK4 [42] (also reported to increase in MCL-1 corresponded to increased Thr-163 decrease BCL-2 protein in AML [43]) did not correlate phosphorylation, which could primarily be attributed to with alvocidib. BRD4 inhibition with JQ1 exhibited single-agent venetoclax treatment in venetoclax-resistant a significant correlation with alvocidib. Notably, only cells, but not in venetoclax-sensitive cells (Figure 3E). alvocidib, LY857785, and JQ1 exhibited low nM EC The functional importance of reducing/inhibiting values, while the remaining CDK inhibitors exhibited µM MCL-1 for the anti-leukemic activity of BCL-2 inhibitors EC values (Figure 4A). is well established, including a role in sequestration Regarding synergy with venetoclax, selective CDK9 of BIM liberated by venetoclax [22]. BIM is a critical inhibition with LDC067 only partially recapitulated effector protein of the intrinsic apoptotic pathway with the full magnitude of potentiation and dose-dependent a capacity to act as a dual sensitizer and activator of activity of alvocidib. LDC067 was significantly less www.impactjournals.com/oncotarget 107210 Oncotarget Figure 3: BCL-2 family proteins altered by alvocidib. (A) AML cell lines were treated with 20 or 80 nM single-agent alvocidib for 24 hours, and BCL-2 family proteins quantified by western blot of duplicate biological extracts. ( B) AML cell lines with low (MOLM- 13) or high (HEL and SET-2) BCL-XL levels were treated with 80 nM alvocidib and resolved on the same gel to facilitate quantitative comparison across these cell lines. (C) After treatment with the alvocidib doses indicated on the x-axis, RNA extracts were prepared at 3 hours (upper panel) or 24 hours (lower panel), and the microRNAs indicated were measured by qPCR. Values shown are mean ± SEM of technical replicates and are representative of three biological replicate experiments. (D) THP-1 and OCI-AML3 were treated with 20 or 80 nM single-agent alvocidib for 24 and 48 hours, and BIM quantified by western blot. ( E) AML cell lines were treated with 80 nM alvocidib * * and 2.5 nM for venetoclax-sensitive lines (MOLM-13 and MV4-11) or 0.25 µM for venetoclax -resistant lines (THP-1 and OCI-AML3), as single-agent and in combination and BCL-2 family proteins quantified by western blot of duplicate biological extracts. ( F) OCI-AML3 and THP-1 cell lysates prepared after treatment with non-silencing (NS) or BIM siRNA were resolved by gel electrophoresis to quantify the level of specific BIM knock-down. ( G) BIM siRNA sequence 5 (the only sequence tested that effectively knocked-down all BIM isoforms) was compared against NS siRNA in cells treated with combined venetoclax and alvocidib in duplicate biological experiments. venetoclax EC values are plotted on a log10 scale at increasing concentrations of alvocidib. Data represent average ± STD. P values were calculated * ** using a two-tailed Student’s T-test. P value < 0.05, P value < 0.01. A–B & D–F, densitometry values shown were calculated using Image J software and normalized to β-Tubulin. www.impactjournals.com/oncotarget 107211 Oncotarget synergistic than alvocidib in venetoclax-resistant cells. values < 0.7, 4 of 13 (31%) exhibited “moderate synergy” Palbociclib showed marginal venetoclax fold-sensitization defined by CI values of 0.7 to 0.85, while one sample had in venetoclax-sensitive cell lines and only at the highest a CI of 0.87 defined as “slight synergy,” and remission (3 µM) dose. Further, palbociclib exhibited divergent sample #11 had a CI of 0.98 indicative of additive activity. activity in venetoclax-resistant cells, with dose-dependent The combination was synergistic in samples from patients antagonism observed in THP-1, and low-dose antagonism clinically refractory to azacitidine (samples 7 & 13) and high-dose sensitization observed in OCI-AML3. (Supplementary Table 1). With seliciclib, venetoclax sensitization (ranging from We analyzed protein levels of selected BCL- 33- to 108-fold) was observed in all cells examined; 2 family members in untreated extracts from eleven however, venetoclax-sensitization occurred only at high available samples used for ex vivo synergy assessment. (≥ 17.5 µM) seliciclib doses. Comparing equipotent Consistent with our previous report, BCL-2 family doses, LY2857785 resulted in similar maximal venetoclax expression was overlapping within individual samples sensitization, although LY2857785 did not recapitulate and heterogeneous across samples [14]. MCL-1 resolved the dose-dependent activity of alvocidib. NU6102 dose- as two distinct molecular weight (MW) bands, with dependently sensitized venetoclax in all cells examined, some samples exhibiting both bands, while most samples with maximal sensitization ranging from 20- to 193-fold predominantly expressed the lower MW band. MCL- at variable NU6102 doses from 2.5 to 20 µM. Despite 1 phosphorylated at Thr163 was detectable only in the poor single-agent correlation, Ro-3306 resulted in dose- single ex vivo venetoclax-resistant sample (#2) (Figure dependent sensitization in all cells examined, with 5C). BCL-XL, BCL-2 and BIM were not detected in the maximal venetoclax EC sensitization ranging from 130- remission sample (#11). No individual BCL-2 family to 242-fold. JQ1 generally recapitulated CDK inhibitor member correlated with synergy, as determined using activity in potentiating venetoclax, exhibiting greater CI values from Figure 5B (Supplementary Figure 5). potentiation in venetoclax-sensitive lines (Figure 4B–4C Levels of BIM correlated most strongly with BCL-2, and and Supplementary Figure 3). We measured apoptosis also significantly correlated with BCL-XL, yet neither to determine whether venetoclax potentiation with total MCL-1 nor high MW MCL-1 correlated with BIM, NU6102 or seliciclib was cytostatic or apoptotic. Similar suggesting that this small sampling of primary AML to alvocidib, early and late apoptosis increased in the samples was primarily dependent on BCL-2 followed by combination with each of these CDK inhibitors beyond BCL-XL for survival (Supplementary Figure 6). For two either single-agent alone, albeit at higher µM doses samples (#6 and #10), sufficient sample quantity enabled (Supplementary Figure 4A–4B). treatments with venetoclax and/or alvocidib. Consistent with divergent in vitro results, we found that alvocidib decreased MCL-1 in sample #6, yet increased MCL-1 in Alvocidib synergizes with venetoclax in short- sample #10. Total BIM was not significantly changed by term ex vivo cultures of AML patient samples alvocidib in either primary sample, while both showed increases in BCL-2 in the combination, consistent with in We analyzed combined venetoclax and alvocidib vitro results (Figure 5D). We identified and treated two in patient samples cultured ex vivo (N = 14). Twelve of additional AML samples to expand upon these results fourteen samples were confirmed as overt AML, one (presented in Figure 5A–5B as #13 and #14). Sample #14 sample was derived from a patient with residual disease exhibited a decrease in MCL-1 with 80 nM alvocidib, (sample #4), and one sample was derived from an which was further decreased by the combination. Sample AML patient in remission (sample #11). Of the samples #13 showed a marginal decrease of MCL-1 only with the analyzed, only one was resistant to venetoclax (sample #2; combination. Neither sample #13 or #14 showed increased EC 1.3 ± 0.08 µM), while the remission sample was also BIM. MCL-1 Thr163 phosphorylation was undetectable resistant (EC > 5 µM) as expected. Median venetoclax or detected at only low levels, and was not significantly EC was 5 nM (excluding the two resistant samples). altered with any treatment (Figure 5E). Alvocidib single-agent activity was homogeneous (median EC 52 ± 19 nM), which for the remission sample was not unexpected as CDK inhibitors can cause cytopenias Alvocidib potentiates venetoclax activity in vivo (Figure 5A). All of the confirmed AML samples exhibited To investigate alvocidib potentiation of venetoclax synergy. Median maximal CI value was 0.52 ± 0.18, activity in vivo, we assessed venetoclax-resistant OCI- occurring at variable doses of venetoclax and alvocidib. AML3 cells in a mouse xenograft model using tumor For consistent comparison, we plotted CI values for volume as the primary endpoint. Animal body weight was median doses at which maximal synergy occurred (40 recorded throughout the study as a surrogate for treatment nM venetoclax, 80 nM alvocidib), alongside the fraction toxicity. Treatment of OCI-AML3 tumor-bearing mice of cells affected, expressed as Fraction Affected (FA) with 100 mg/kg venetoclax resulted in a 31.6% average (Figure 5B). At these low, clinically-achievable doses, 7 tumor growth inhibition (TGI) by day 14 relative to of 13 samples (54%) exhibited “synergy” as defined by CI www.impactjournals.com/oncotarget 107212 Oncotarget vehicle treated animals. Alvocidib treatment at 2.5 mg/kg Altogether, both in vitro and ex vivo results are resulted in a 9.7% TGI over the same period. However, consistent in showing divergent effects on BCL-2 family the combination of alvocidib and venetoclax resulted in proteins despite similar levels of synergy, suggesting a significant reduction in tumor volume, corresponding to that no effect is absolutely associated with synergy per an 87.9% TGI at day 14 (Figure 6A). Neither venetoclax, se, yet the overall balance of pro- and anti-apoptotic nor alvocidib, resulted in a significant reduction in BCL-2 family proteins is consistently shifted in favor of bodyweight. The combination was well-tolerated, with apoptosis. Previous reports of exogenous manipulation of an average 7.5% reduction in bodyweight at day 7, MCL-1, and dual targeting of MCL-1 and BCL-2, support comparable to that of venetoclax alone (Figure 6B). this concept as it relates to alvocidib and enhanced venetoclax activity [26, 46–48]. Importantly, abrogation of synergy with either BIM knock-down or BCL-XL over- DISCUSSION expression also supports this concept and provides strong functional evidence that the majority of synergy between The synergistic anti-leukemic activity of combined venetoclax and alvocidib is based on mechanism(s) that venetoclax and alvocidib observed in venetoclax-sensitive converge upon the intrinsic apoptotic pathway. More and –resistant models of AML ex vivo, in vivo, and in specifically, while BIM knock-down showed a trend vitro suggests that this combination has the potential to towards abrogating single-agent venetoclax activity, improve outcomes in patients regardless of sensitivity the change was not significant, suggesting that, at to single-agent venetoclax. To our knowledge, this is the least in the two AML cell lines analyzed, single-agent first report exploring the potential synergy of venetoclax venetoclax operates largely independent of BIM through and alvocidib in AML. Considering the clinical relevance the classic displacement model [49], whereby BCL-2 of alvocidib and venetoclax, as well as lack of effective directly interacts with BAX/BAK to negatively regulate treatment options for AML, we propose that the current MOMP. Conversely, BIM knock-down abrogated the report is distinct from prior reports evaluating BCL-2 majority of synergy between venetoclax and alvocidib, inhibitors with CDK inhibitors, e.g. [44, 45]. suggesting that the combination is at least partially In the current report, a decrease of anti-apoptotic dependent on the direct activation model [49], whereby MCL-1 was observed in most AML cells tested, while an BIM acts as sensitizer and/or activator of BAX/BAK. To increase of pro-apoptotic BH3-only proteins, predominantly our knowledge the effect of BIM knock-down has not BIM, was observed in other cell lines. MOLM-13, a cell previously been examined in the context of the proposed line derived from an MDS patient transformed to AML, was combination, thus abrogation of synergy by BIM knock- unique amongst cell lines examined in that alvocidib did down is an important insight into the mechanism of not decrease MCL-1, yet demonstrated strong synergy with apoptotic synergy between alvocidib and venetoclax. venetoclax. Importantly, increases of all pro-apoptotic BH3- Importantly, reduction of MCL-1 by alvocidib similarly only proteins measured, including BIM, PUMA, NOXA has the capacity to operate through the displacement and NBK, were observed in MOLM-13. This suggests model, by decreasing MCL-1 available to inhibit BAX/ that an increase of pro-apoptotic BH3-only proteins can BAK, as well as the direct activation model, by decreasing be functionally equivalent to a decrease of anti-apoptotic MCL-1 available to act as a reservoir for sequestration of MCL-1 with regard to synergy with venetoclax. BIM and/or other BH3-only proteins. MicroRNAs from the polycistron miR-17–92, Using multiple complementary approaches, we known to negatively regulate BIM [35, 36], were clearly demonstrate that BCL-XL counteracts venetoclax/ decreased in alvocidib-treated cells in vitro, although High alvocidib synergy in vitro. Results with BCL-XL cells further experiments are needed to determine if this is a are consistent with a model whereby BCL-XL acts in direct or indirect result of CDK9 inhibition. Combined parallel to BCL-2 to inhibit BAX/BAK and/or sequester venetoclax and alvocidib was found to increase pro- BIM/BH3-only proteins, and with a model whereby BCL- apoptotic BH3-only proteins beyond the increase XL can also counteract MCL-1 reduction and/or BH3- observed with each single-agent for some cell lines. only protein increases by alvocidib. Given the totality Thus, in addition to increased levels of “free” BH3- of in vitro data regarding BCL-XL, we were surprised to only protein(s) (i.e. unbound by anti-apoptotic proteins) observe increased BCL-XL protein in alvocidib treated resulting from BCL-2 inhibition by venetoclax, and/or samples in vitro. Importantly however, the levels of alvocidib reduction of MCL-1 protein, alvocidib may have Low increased BCL-XL observed in BCL-XL cells were the capacity to induce apoptosis by increasing total BIM still strikingly lower than baseline levels observed in protein levels and/or other BH3-only proteins, although High BCL-XL cells, demonstrating that absolute levels of this activity was not observed in our small sampling of BCL-XL may be an important determinant of response AML analyzed ex vivo (n = 4). MCL-1 was decreased by in vitro. BCL-XL levels did not correlate with venetoclax alvocidib and/or the combination in three of four AML single-agent activity or synergy with alvocidib in our patient samples assessed ex vivo, suggesting MCL-1 as a small sampling of AML patients ex vivo, suggesting that predominant mechanism in the primary cells analyzed. www.impactjournals.com/oncotarget 107213 Oncotarget BCL-XL may not represent a major limitation to clinical venetoclax activity, and was actually antagonistic in some efficacy of the combination in AML. However, several AML cells. While we cannot rule-out the possibility that questions regarding BCL-XL as an in vivo mechanism CDK4/6 inhibition contributes to venetoclax sensitization of resistance to the combination remain unanswered, under concurrent CDK9 inhibition, these results suggest such as differences in relative versus absolute levels of that palbociclib may not be an ideal candidate for BCL-XL in vitro versus in vivo, as well as potential cell combination with venetoclax in AML. Strong venetoclax lineage effects of BCL-XL function [50]. It remains to be potentiation was also observed with seliciclib, NU6102, determined clinically whether some cases of AML with and Ro-3306; however, while these CDK inhibitors do high BCL-XL expression (putatively erythroleukemia or not potently inhibit CDK9, they likely inhibit CDK9 at megakaryoblastic lineage expression of BCL-XL) may be the relatively high doses required to elicit the synergy more resistant to the proposed combination of venetoclax observed. In comparing equipotent doses of LY2857785 or and alvocidib, and thus benefit more from dual BCL-2 alvocidib combined with venetoclax, maximal venetoclax and BCL-XL inhibition (i.e. navitoclax/ABT-737/ABT- fold-sensitization was similar for both compounds, 263) alone or combined with alvocidib. Nonetheless, our although synergy at the next closest dose was lower observations are consistent with recent reports showing for LY2857785, especially for venetoclax -resistant that mechanisms of venetoclax resistance are not universal cells. BRD4 inhibitor JQ1 resulted in dose-dependent or mutually exclusive [51, 52]. venetoclax sensitization similar to alvocidib, suggesting Upon examination of BCL-2 family proteins after that inhibition of P-TEFb through BRD4 is functionally combined venetoclax and alvocidib treatment, we were similar to CDK9 inhibition. As JQ1 unlikely has off-target surprised to observe increased MCL-1, BCL-2 and BCL- effects on CDKs, this indirect evidence supports the notion XL. However, we observed a corresponding increase that alvocidib predominantly synergizes with venetoclax in BIM beyond the increase attributable to single-agent through CDK9 inhibition. alvocidib in most cells, while NOXA was observed to be The functional and mechanistic evidence increased in the combination, beyond either single-agent, presented herein provide strong pre-clinical precedence in the only cell line for which we did not observe strong for clinically investigating the efficacy of combined BIM up-regulation at the same time point. BIM and BCL- alvocidib and venetoclax in AML. Clinical studies 2 have the capacity to mutually regulate each other, thus have shown that alvocidib can be safely added to we speculate that the increase in anti-apoptotic BCL-2 induction therapy with acceptable toxicity [57], thus family proteins could be a response to increased BIM, or the combination may be clinically feasible. However, vice versa [53]. Nonetheless, the data presented herein the cytopenias observed with alvocidib, and recently is consistent with the hypothesis that the net balance of with combined venetoclax and azacitidine (including pro- and anti-apoptotic BCL2 family members determines unexpected thrombocytopenia), emphasize the intrinsic apoptotic response in general, as well as for the importance of using the lowest possible doses, and proposed combination. BH3 profiling, an evolving assay highlights the importance of low dose synergy we that functionally probes the intrinsic apoptotic pathway, consistently demonstrate herein. Development of a could therefore prove to be a useful predictive biomarker clinical trial testing combined alvocidib and venetoclax for this regimen [4, 14, 44, 54, 55]. in AML and high-risk MDS is ongoing. We found that CDK9-selective inhibition correlated strongly with venetoclax potentiation. While the MATERIALS AND METHODS magnitude of venetoclax sensitization with LDC067 was significantly lower, especially in venetoclax-resistant cells, Cells, culture conditions and reagents a strong correlation was observed. We speculate these distinctions are due to differences in potency for CDK9, Primary samples were obtained with informed although inhibition of other CDK isoforms cannot be consent in accordance with Mayo Clinic IRB-approved ruled-out. We did not test the CDK9 inhibitor dinaciclib research protocols and handled according to Good in this study, as previous studies in diffuse large B-cell Clinical Practice. Primary cells were Ficoll-gradient lymphoma (DLBCL) have been reported [56]. While separated, and immediately cultured, or viably frozen for a stronger single-agent potency of dinaciclib versus subsequent short-term culture. Cell lines were obtained alvocidib in DLBCL cell lines was reported, a clear and from ATCC or DSMZ. All cell lines tested negative for comparable potentiation venetoclax upon combined mycoplasma before cryopreservation, and all cell lines dinaciclib was also reported, regardless of the single- were confirmed free of cross-contamination using PCR- agent potency of either dinaciclib or venetoclax in the based DNA fingerprinting. Cells were cultured in RPMI- DLBCL cell lines used. These observations are consistent 1640 containing 10% FBS, 2 mM L-glutamine, 100 IU/ with our findings demonstrating robust synergy regardless mL penicillin, 100 µg/mL streptomycin (Invitrogen) at of venetoclax single-agent activity in AML. CDK4/6 37°C/5% CO2. Compounds were obtained as follows: inhibition with palbociclib did not generally potentiate Alvocidib (Tolero Pharmaceuticals), venetoclax/ABT-199 www.impactjournals.com/oncotarget 107214 Oncotarget (ChemieTek), ABT-737, LDC067, palbociclib, seliciclib 96 hours using a Cytation3 plate reader (BioTek). Nine and Ro-3306 (SelleckChem), LY2857785 (MedChem doses of venetoclax titrated specifically for venetoclax- Express), NU6102 and JQ1 (Cayman Chemical). resistant or -sensitive cells, or ABT-737, were combined with six doses of the second drug evaluated, yielding 54 possible combinations, each evaluated in quadruplicate Drug dose response assays and CalcuSyn analysis for every experiment. For primary cells, where venetoclax Cells were plated in 384-well plates (Greiner Bio- sensitivity was unknown a priori, venetoclax was serially One) at 1000 cells/well for cell lines, or 2000 cells/ diluted 5-fold (5.0 µM to 0.013 nM), which yielded at least well for primary samples, and dosed simultaneously for three doses for CalcuSyn analysis, version 2.1 (Biosoft) combinations. Relative cell number (expressed as % [58, 59]. Prism version 5.03 (Prism Software Corporation) viability) was measured with CellTiter-Glo (Promega) at was used to calculate EC values. Figure 4: Pharmacological dissection of CDK isoforms contributing to venetoclax sensitization. (A) single-agent drug dose response assays were assessed with duplicate biological experiments, each with four technical replicates, in a panel of eleven AML cell lines, and each EC value is plotted versus the EC value of alvocidib. Data represent average ± STD. Regression analysis was 50 50 used to determine R and P values. (B and C), biological duplicate combination drug dose response assays, each with four technical replicates for every dose/dose combination evaluated, were performed with venetoclax and each inhibitor, and maximal venetoclax EC fold-sensitization (B) and penultimate EC fold-sensitization (C) plotted for each inhibitor. Data represent average ± STD. www.impactjournals.com/oncotarget 107215 Oncotarget Aldrich) added to [5 µg/mL]. After incubating 15 minutes, Flow cytometry cells were analyzed on LSR Fortessa (BD Biosciences). Cells were seeded at 2E5 cells/mL and incubated 24 hours with appropriate CDK inhibitor (80 nM alvocidib, Western blots 17.5 µM seliciclib, or 2.5 or 20 µM NU6102) with or without 2.5 or 10 nM venetoclax for venetoclax-sensitive After treatment with alvocidib and/or venetoclax at cells, or 0.25 or 1 µM venetoclax for venetoclax -resistant indicated doses/times, protein was harvested with lysis buffer cells. Cells were washed with ice-cold PBS, suspended in (Cell Signaling; #9803) containing 1 mM PMSF. Lysates binding buffer at 1E6 cells/mL, Annexin V antibody added were quantified by BCA (Pierce/Thermo Fisher), 26–30 µg at [1:20] (BD Biosciences) and propidium iodide (Sigma total protein resolved by 4–15% SDS-PAGE, transferred www.impactjournals.com/oncotarget 107216 Oncotarget Figure 5: Alvocidib synergizes with venetoclax in short-term ex vivo cultures of AML patient samples and in a mouse model. (A) twelve primary samples from bone aspirates of AML patients were treated with venetoclax and alvocidib in 96 hour culture, single-agent venetoclax and alvocidib EC50 values are plotted for each sample. Data represent average ± STD. (B) Combination Index (CI) values as an expression of synergy and Fraction Affected (FA) as an expression of fractional cell number (1.0 = maximal effect, while 0 = no effect) are shown for the combination of 40 nM venetoclax and 80 nM alvocidib as determined with CalcuSyn software. Dashed lines demark CI ranges: Strong synergy (< 0.3), synergy (0.3 to 0.7), moderate synergy (0.7 to 0.85), slight synergy (0.85 to 0.9), and additive (0.9 to 1.0). (C) untreated cell pellets were available for eleven of the fourteen primary samples analyzed in drug response assays, from which lysates were prepared for quantification of BCL-2 family proteins by western blot. ( D) primary samples #6 and #10 after treatment with 20 nM venetoclax and 80 nM alvocidib, alone and in combination, for quantification of BIM, MCL-1 and BCL-2 by western blot. ( E) primary samples #13 and #14 after treatment with 20 nM venetoclax and 80 nM alvocidib, alone and in combination, for quantification of BIM, MCL-1 and pMCL-1-Thr163 by western blot. For 5C–5E densitometry values calculated with Image J software are normalized to β-actin load controls. For ex vivo studies, primary sample material availability facilitated a single drug dose response assay for each sample, and a single lysate preparation for the indicated samples. www.impactjournals.com/oncotarget 107217 Oncotarget to PVDF membranes (80V/90 minute wet-transfer), and Small-interfering RNA transfection blocked with 5% non-fat dry milk before primary antibody siRNA were purchased from Qiagen. For target incubation, 4°C/overnight. MCL-1 (#4527), pMCL-1-Thr163 protein knock-down validation, 0.04 nmol siRNA was mixed (#1476), BCL-2 (#2872), BCL-XL (#2762), BIM (#2819), with 750 µL RNAiMax transfection reagent (ThermoFisher) PUMA (#4976), β-tubulin (#2128) (Cell Signaling), NBK diluted in OPTI-MEM (Invitrogen) (dilutions optimized (sc-365625; Santa Cruz Biotechnology), NOXA (ab140129; per cell line and reagent batch), after 30 minutes 2.6 to Abcam), and β-actin (A00702; GenScript). Densitometry was 4.3E5 cells (optimized per cell line) were added in 1 mL performed with Image J software (www.imagej.nih.gov). unsupplemented RPMI-1640, spiked with FBS to [2.2%], and harvested at 48 hours. For siRNA drug dose response Lentivirus construction assays, siRNA were pre-printed on 384-well plates, OPTI- MEM-diluted transfection reagent was added at 20 µL/ BCL-XL (cDNA clone ID 2823498, GE well, after 30 minutes 1000–2000 cells/well were added Dharmacon) was amplified by PCR and cloned into in 20 µL of unsupplemented RPMI-1640, and spiked with plasmid pSC11CMVFlag, with a CMV promoter and FBS to [2.2%]. For BIM siDDRs, plates were incubated for 5′3Xflag tag. BCL-XL was then cloned into lentiviral 24 hours prior to simultaneous dosing of venetoclax and vector pWPI (Addgene plasmid #12254) resulting in alvocidib and read 48 hours later. For BCL-XL siDDRs, plasmid pWPIS6FBCL-XL. Plasmid pWPIS6FBCL- plates were incubated for 48 hours prior to simultaneous XL, together with packaging plasmid psPAX2 (plasmid addition of venetoclax or ABT-737 and alvocidib and read #12660) and pMG2.g (plasmid #12259) were transfected 48 hours later. into 293T cells and rescued into lentivirus. Mouse studies Quantitative RT-PCR Animal studies were reviewed and approved for MicroRNAs were isolated using Directzol RNA ethical consideration by an internal review committee mini-prep kit (Zymo Research). Reverse transcription was at Tolero Pharmaceuticals, Inc. OCI-AML3 cells were performed using TaqMan MicroRNA RT Kit, resulting transplanted into 6–8 week old, female, hairless outbred cDNA underwent pre-amplification with TaqMan PreAmp scid hr SCID mice (Crl:SHO-Prkdc Hr ) (18–26 g, strain Master Mix, and RTqPCR was performed with TaqMan code 474, Charles River Laboratories). Study animals MicroRNA assays and TaqMan Universal Master Mix II were housed under standard conditions, and given containing UNG (Life Technologies). Figure 6: Alvocidib + venetoclax is an active regimen in the venetoclax-resistant AML xenograft (OCI-AML3) model. Athymic nude mice were injected subcutaneously in the hind flank with OCI-AML3 cells. When tumors reached approximately 200 mm3, mice were randomized and then treated with vehicle, 100 mg/kg venetoclax, 2.5 mg/kg alvocidib, or a combination of the two drugs. Mice were dosed daily by oral gavage (venetoclax) or intraperitoneal injection (alvocidib). Tumor volumes (A) and body weights (B) were measured and recorded twice weekly. Tumor volumes and body weights shown are mean ± SEM of eight animals per treatment cohort. www.impactjournals.com/oncotarget 107218 Oncotarget food and water ad libitum. 1E7 OCI-AML3 cells were REFERENCES suspended in serum-free media, mixed 1:1 in matrigel (Corning), and injected subcutaneously into the hind 1. Roberts AW, Davids MS, Pagel JM, Kahl BS, Puvvada flank. Randomization and treatment was initiated once SD, Gerecitano JF, Kipps TJ, Anderson MA, Brown JR, tumor volumes had reached approximately 200 mm . Gressick L, Wong S, Dunbar M, Zhu M, et al. Targeting Alvocidib was formulated in saline, while venetoclax BCL2 with Venetoclax in Relapsed Chronic Lymphocytic was formulated in Phosal 50PG: PEG400: Ethanol. Leukemia. N Engl J Med. 2016; 374:311–22. https://doi. Mice were administered vehicle or drug treatment by org/10.1056/NEJMoa1513257. intraperitoneal injection, (qdx5)x3 with alvocidib (2.5 mg/ 2. Kaufmann SH, Karp JE, Svingen PA, Krajewski S, Burke kg), by oral gavage (qdx5)x3 with venetoclax (100 mg/ PJ, Gore SD, Reed JC. Elevated expression of the apoptotic kg), or both alvocidib and venetoclax. Tumor length and regulator Mcl-1 at the time of leukemic relapse. Blood. width measurements were made with digital calipers, and 1998; 91:991–1000. tumor volumes were calculated as follows: length x width 3. van Stijn A, van der Pol MA, Kok A, Bontje PM, x width/2, where length was the longest diameter. Tumor Roemen GM, Beelen RH, Ossenkoppele GJ, Schuurhuis volume and body weights were measured and recorded GJ. Differences between the CD34+ and CD34- blast twice weekly. compartments in apoptosis resistance in acute myeloid leukemia. Haematologica. 2003; 88:497–508. Statistics 4. Vo TT, Ryan J, Carrasco R, Neuberg D, Rossi DJ, Stone RM, Deangelo DJ, Frattini MG, Letai A. Relative mitochondrial P values were calculated using two-tailed Student’s priming of myeloblasts and normal HSCs determines t-test. For correlative comparisons of drug EC values, or chemotherapeutic success in AML. Cell. 2012; 151:344–55. drug activity with protein level, regression analysis was used 2 https://doi.org/10.1016/j.cell.2012.08.038. to calculate R and corresponding p values. R values shown 5. Lagadinou ED, Sach A, Callahan K, Rossi RM, Neering in Figure 1B were calculated with CalcuSyn software. SJ, Minhajuddin M, Ashton JM, Pei S, Grose V, O’Dwyer KM, Liesveld JL, Brookes PS, Becker MW, et al. BCL-2 Author contributions inhibition targets oxidative phosphorylation and selectively eradicates quiescent human leukemia stem cells. Cell J.M.B. performed research, designed experiments, Stem Cell. 2013; 12:329–41. https://doi.org/10.1016/j. analyzed and interpreted data, and wrote the manuscript, stem.2012.12.013. C.J.W. A.S.J, S.L.W., and R.T. performed research, 6. Beurlet S, Omidvar N, Gorombei P, Krief P, Le Pogam C, designed experiments, and analyzed and interpreted Setterblad N, de la Grange P, Leboeuf C, Janin A, Noguera data. C.X.S. and A.K.S. assisted with lentiviral construct ME, Hervatin F, Sarda-Mantel L, Konopleva M, et al. BCL- preparation. N.H. and D.D. performed experiments 2 inhibition with ABT-737 prolongs survival in an NRAS/ and analyzed data, W.K., H.H., K.K.S., Y.S.L., and P.P. BCL-2 mouse model of AML by targeting primitive LSK performed experiments. J.M.B., C.J.W., A.S-J., S.W., and progenitor cells. Blood. 2013; 122:2864–76. https://doi. D.J.B., R.A.M, and R.T. conceptualized the study. All org/10.1182/blood-2012-07-445635. authors read and approved the final manuscript. 7. Pan R, Hogdal LJ, Benito JM, Bucci D, Han L, Borthakur G, Cortes J, DeAngelo DJ, Debose L, Mu H, Dohner H, Gaidzik ACKNOWLEDGMENTS VI, Galinsky I, et al. Selective BCL-2 inhibition by ABT-199 causes on-target cell death in acute myeloid leukemia. Cancer Vector pWPI, psPAX2 and pMG2.g were generous Discov. 2014; 4:362–75. https://doi.org/10.1158/2159-8290. gifts from Didier Trono. CD-13-0609. 8. Raza A, Gezer S, Mundle S, Gao XZ, Alvi S, Borok R, Rifkin S, Iftikhar A, Shetty V, Parcharidou A. Apoptosis CONFLICTS OF INTEREST in bone marrow biopsy samples involving stromal and hematopoietic cells in 50 patients with myelodysplastic C.J.W., W.K., H.H., K.K.S., Y.L., P.P., A.S-J., syndromes. Blood. 1995; 86:268–76. S.W., D.J.B. and S.L.W. are employed by Tolero Pharmaceuticals. R.T. is a scientific advisor for Tolero 9. Parker JE, Mufti GJ, Rasool F, Mijovic A, Devereux S, Pharmaceuticals. The remaining authors declare no Pagliuca A. The role of apoptosis, proliferation, and the Bcl- competing interests. 2-related proteins in the myelodysplastic syndromes and acute myeloid leukemia secondary to MDS. Blood. 2000; 96:3932–8. FUNDING 10. Invernizzi R, Pecci A, Bellotti L, Ascari E. Expression This work was supported by Mayo Clinic of p53, bcl-2 and ras oncoproteins and apoptosis levels institutional support of JMB and RT. in acute leukaemias and myelodysplastic syndromes. www.impactjournals.com/oncotarget 107219 Oncotarget Leuk Lymphoma. 2001; 42:481–9. https://doi. access of MCL-1 to its E3 ligase. FEBS Lett. 2005; org/10.3109/10428190109064605. 579:5603–8. https://doi.org/S0014-5793(05)01135-X. 11. Albitar M, Manshouri T, Shen Y, Liu D, Beran M, 21. Jilg S, Reidel V, Muller-Thomas C, Konig J, Schauwecker Kantarjian HM, Rogers A, Jilani I, Lin CW, Pierce S, J, Hockendorf U, Huberle C, Gorka O, Schmidt B, Burgkart Freireich EJ, Estey EH. Myelodysplastic syndrome is not R, Ruland J, Kolb HJ, Peschel C, et al. Blockade of BCL-2 merely “preleukemia”. Blood. 2002; 100:791–8. proteins efficiently induces apoptosis in progenitor cells of high-risk myelodysplastic syndromes patients. Leukemia. 12. Slape CI, Saw J, Jowett JB, Aplan PD, Strasser A, Jane SM, 2016; 30:112–23. https://doi.org/10.1038/leu.2015.179. Curtis DJ. Inhibition of apoptosis by BCL2 prevents leukemic transformation of a murine myelodysplastic syndrome. Blood. 22. Niu X, Zhao J, Ma J, Xie C, Edwards H, Wang G, Caldwell 2012; 120:2475–83. https://doi.org/blood-2012-05-430736. JT, Xiang S, Zhang X, Chu R, Wang J, Lin H, Taub JW, et al. 13. Konopleva M, Pollyea DA, Potluri J, Chyla BJ, Busman Binding of released Bim to Mcl-1 is a mechanism of intrinsic T, McKeegan E, Salem A, Zhu M, Ricker JL, Blum W, resistance to ABT-199 which can be overcome by combination DiNardo CD, Dunbar M, Kirby R, et al. A Phase 2 Study of with daunorubicin or cytarabine in AML cells. Clin Cancer ABT-199 (GDC-0199) in Patients with Acute Myelogenous Res. 2016. https://doi.org/clincanres.3057.2015. Leukemia (AML). Blood. 2014; 124:118. 23. Bodo J, Zhao X, Durkin L, Souers AJ, Phillips DC, Smith 14. Bogenberger JM, Kornblau SM, Pierceall WE, Lena R, MR, Hsi ED. Acquired resistance to venetoclax (ABT-199) Chow D, Shi CX, Mantei J, Ahmann G, Gonzales IM, in t(14;18) positive lymphoma cells. Oncotarget. 2016; Choudhary A, Valdez R, Camoriano J, Fauble V, et al. 7:70000–10. https://doi.org/10.18632/oncotarget.12132. BCL-2 family proteins as 5-Azacytidine-sensitizing targets 24. Pan R, Ruvolo VR, Wei J, Konopleva M, Reed JC, Pellecchia and determinants of response in myeloid malignancies. M, Andreeff M, Ruvolo PP. Inhibition of Mcl-1 with the Leukemia. 2014; 28:1657–65. pan-Bcl-2 family inhibitor (-)BI97D6 overcomes ABT-737 15. Bogenberger JM, Delman D, Hansen N, Valdez R, Fauble V, resistance in acute myeloid leukemia. Blood. 2015; 126:363– Mesa RA, Tibes R. Ex vivo activity of BCL-2 family inhibitors 72. https://doi.org/10.1182/blood-2014-10-604975. ABT-199 and ABT-737 combined with 5-azacytidine in 25. Chen R, Keating MJ, Gandhi V, Plunkett W. Transcription myeloid malignancies. Leuk Lymphoma. 2015; 56:226–9. inhibition by flavopiridol: mechanism of chronic https://doi.org/10.3109/10428194.2014.910657. lymphocytic leukemia cell death. Blood. 2005; 106:2513–9. 16. Tsao T, Shi Y, Kornblau S, Lu H, Konoplev S, Antony https://doi.org/2005-04-1678. A, Ruvolo V, Qiu YH, Zhang N, Coombes KR, Andreeff 26. Gojo I, Zhang B, Fenton RG. The cyclin-dependent M, Kojima K, Konopleva M. Concomitant inhibition kinase inhibitor flavopiridol induces apoptosis in multiple of DNA methyltransferase and BCL-2 protein function myeloma cells through transcriptional repression and down- synergistically induce mitochondrial apoptosis in acute regulation of Mcl-1. Clin Cancer Res. 2002; 8:3527–38. myelogenous leukemia cells. Ann Hematol. 2012; 91:1861– 27. Kitada S, Zapata JM, Andreeff M, Reed JC. Protein 70. https://doi.org/10.1007/s00277-012-1537-8. kinase inhibitors flavopiridol and 7-hydroxy-staurosporine 17. DiNardo C, Pollyea D, Pratz K, Thirman MJ, Letai A, Frattini down-regulate antiapoptosis proteins in B-cell chronic M, Jonas B, Leverson J, Zhu M, Dunbar M, Falotico N, Kirby lymphocytic leukemia. Blood. 2000; 96:393–7. R, Agarwal S, et al. A Phase 1b Study of Venetoclax (ABT-199/ 28. Chao SH, Price DH. Flavopiridol inactivates P-TEFb and GDC-0199) in Combination with Decitabine or Azacitidine in blocks most RNA polymerase II transcription in vivo. J Treatment-Naive Patients with Acute Myelogenous Leukemia Biol Chem. 2001; 276:31793–9. https://doi.org/10.1074/ Who Are ≥ to 65 Years and Not Eligible for Standard Induction jbc.M102306200. Therapy. Blood. 2015; 126:327. 29. Wada T, Takagi T, Yamaguchi Y, Watanabe D, Handa H. 18. De Biasio A, Vrana JA, Zhou P, Qian L, Bieszczad CK, Evidence that P-TEFb alleviates the negative effect of DSIF Braley KE, Domina AM, Weintraub SJ, Neveu JM, Lane on RNA polymerase II-dependent transcription in vitro. WS, Craig RW. N-terminal truncation of antiapoptotic EMBO J. 1998; 17:7395–403. https://doi.org/10.1093/ MCL1, but not G2/M-induced phosphorylation, is emboj/17.24.7395. associated with stabilization and abundant expression in tumor cells. J Biol Chem. 2007; 282:23919–36. https://doi. 30. Pallis M, Burrows F, Whittall A, Boddy N, Seedhouse C, org/M700938200. Russell N. Efficacy of RNA polymerase II inhibitors in targeting dormant leukaemia cells. BMC Pharmacol Toxicol. 19. Perciavalle RM, Stewart DP, Koss B, Lynch J, Milasta S, 2013; 14:32. https://doi.org/10.1186/2050-6511-14-32. Bathina M, Temirov J, Cleland MM, Pelletier S, Schuetz JD, Youle RJ, Green DR, Opferman JT. Anti-apoptotic 31. Chen S, Dai Y, Pei XY, Myers J, Wang L, Kramer LB, MCL-1 localizes to the mitochondrial matrix and couples Garnett M, Schwartz DM, Su F, Simmons GL, Richey JD, mitochondrial fusion to respiration. Nat Cell Biol. 2012; Larsen DG, Dent P, et al. CDK inhibitors upregulate BH3- 14:575–83. https://doi.org/10.1038/ncb2488. ncb2488. only proteins to sensitize human myeloma cells to BH3 20. Warr MR, Acoca S, Liu Z, Germain M, Watson M, mimetic therapies. Cancer Res. 2012; 72:4225–37. https:// Blanchette M, Wing SS, Shore GC. BH3-ligand regulates doi.org/10.1158/0008-5472.CAN-12-1118. www.impactjournals.com/oncotarget 107220 Oncotarget 32. Harley ME, Allan LA, Sanderson HS, Clarke PR. 42. Vassilev LT, Tovar C, Chen S, Knezevic D, Zhao X, Sun Phosphorylation of Mcl-1 by CDK1-cyclin B1 initiates its H, Heimbrook DC, Chen L. Selective small-molecule Cdc20-dependent destruction during mitotic arrest. EMBO J. inhibitor reveals critical mitotic functions of human CDK1. 2010; 29:2407–20. https://doi.org/10.1038/emboj.2010.112. Proc Natl Acad Sci U S A. 2006; 103:10660–5. https://doi. org/0600447103. 33. Inuzuka H, Shaik S, Onoyama I, Gao D, Tseng A, Maser RS, Zhai B, Wan L, Gutierrez A, Lau AW, Xiao Y, Christie AL, 43. Kojima K, Shimanuki M, Shikami M, Andreeff M, Aster J, et al. SCF(FBW7) regulates cellular apoptosis by Nakakuma H. Cyclin-dependent kinase 1 inhibitor RO-3306 targeting MCL1 for ubiquitylation and destruction. Nature. enhances p53-mediated Bax activation and mitochondrial 2011; 471:104–9. https://doi.org/10.1038/nature09732. apoptosis in AML. Cancer Sci. 2009; 100:1128–36. https:// nature09732. doi.org/10.1111/j.1349-7006.2009.01150.x. 44. Pallis M, Burrows F, Ryan J, Grundy M, Seedhouse 34. Choudhary GS, Tat TT, Misra S, Hill BT, Smith MR, C, Abdul-Aziz A, Montero J, Letai A, Russell N. Almasan A, Mazumder S. Cyclin E/Cdk2-dependent Complementary dynamic BH3 profiles predict co- phosphorylation of Mcl-1 determines its stability and operativity between the multi-kinase inhibitor TG02 and the cellular sensitivity to BH3 mimetics. Oncotarget. 2015; BH3 mimetic ABT-199 in acute myeloid leukaemia cells. 6:16912–16925. https://doi.org/10.18632/oncotarget.4857. Oncotarget. 2017; 8:16220–32. https://doi.org/10.18632/ oncotarget.8742. 35. Ventura A, Young AG, Winslow MM, Lintault L, Meissner A, Erkeland SJ, Newman J, Bronson RT, Crowley D, Stone 45. Polier G, Giaisi M, Kohler R, Muller WW, Lutz C, Buss EC, JR, Jaenisch R, Sharp PA, Jacks T. Targeted deletion reveals Krammer PH, Li-Weber M. Targeting CDK9 by wogonin and essential and overlapping functions of the miR-17 through related natural flavones potentiates the anti-cancer efficacy 92 family of miRNA clusters. Cell. 2008; 132:875–86. of the Bcl-2 family inhibitor ABT-263. Int J Cancer. 2015; https://doi.org/10.1016/j.cell.2008.02.019. 136:688–98. https://doi.org/10.1002/ijc.29009. 36. Xu Z, Sharp PP, Yao Y, Segal D, Ang CH, Khaw SL, Aubrey 46. Ma Y, Cress WD, Haura EB. Flavopiridol-induced apoptosis BJ, Gong J, Kelly GL, Herold MJ, Strasser A, Roberts AW, is mediated through up-regulation of E2F1 and repression Alexander WS, et al. BET inhibition represses miR17–92 of Mcl-1. Mol Cancer Ther. 2003; 2:73–81. to drive BIM-initiated apoptosis of normal and transformed 47. Phillips DC, Xiao Y, Lam LT, Litvinovich E, Roberts- hematopoietic cells. Leukemia. 2016; 30:1531–41. https:// Rapp L, Souers AJ, Leverson JD. Loss in MCL-1 function doi.org/10.1038/leu.2016.52. sensitizes non-Hodgkin’s lymphoma cell lines to the BCL- 37. Albert TK, Rigault C, Eickhoff J, Baumgart K, Antrecht 2-selective inhibitor venetoclax (ABT-199). Blood Cancer C, Klebl B, Mittler G, Meisterernst M. Characterization of J. 2016; 6:e403. https://doi.org/10.1038/bcj.2016.12. molecular and cellular functions of the cyclin-dependent 48. Teh TC, Nguyen NY, Moujalled DM, Segal D, Pomilio G, kinase CDK9 using a novel specific inhibitor. Br J Pharmacol. Rijal S, Jabbour A, Cummins K, Lackovic K, Blombery 2014; 171:55–68. https://doi.org/10.1111/bph.12408. P, Thompson E, Ekert PG, Lessene G, et al. Enhancing 38. Karaman MW, Herrgard S, Treiber DK, Gallant P, Atteridge venetoclax activity in acute myeloid leukemia by co- CE, Campbell BT, Chan KW, Ciceri P, Davis MI, Edeen PT, targeting MCL1. Leukemia. 2017. https://doi.org/10.1038/ Faraoni R, Floyd M, Hunt JP, et al. A quantitative analysis leu.2017.243. of kinase inhibitor selectivity. Nat Biotechnol. 2008; 49. Shamas-Din A, Brahmbhatt H, Leber B, Andrews DW. 26:127–32. https://doi.org/10.1038/nbt1358. BH3-only proteins: Orchestrators of apoptosis. Biochim 39. Meijer L, Borgne A, Mulner O, Chong JP, Blow JJ, Inagaki Biophys Acta. 2011; 1813:508–20. https://doi.org/10.1016/j. N, Inagaki M, Delcros JG, Moulinoux JP. Biochemical bbamcr.2010.11.024. and cellular effects of roscovitine, a potent and selective 50. Chonghaile TN, Roderick JE, Glenfield C, Ryan J, Sallan inhibitor of the cyclin-dependent kinases cdc2, cdk2 and SE, Silverman LB, Loh ML, Hunger SP, Wood B, DeAngelo cdk5. Eur J Biochem. 1997; 243:527–36. DJ, Stone R, Harris M, Gutierrez A, et al. Maturation 40. Yin T, Lallena MJ, Kreklau EL, Fales KR, Carballares stage of T-cell acute lymphoblastic leukemia determines S, Torrres R, Wishart GN, Ajamie RT, Cronier DM, BCL-2 versus BCL-XL dependence and sensitivity to Iversen PW, Meier TI, Foreman RT, Zeckner D, et al. A ABT-199. Cancer Discov. 2014; 4:1074–87. https://doi. novel CDK9 inhibitor shows potent antitumor efficacy in org/10.1158/2159-8290.CD-14-0353. preclinical hematologic tumor models. Mol Cancer Ther. 51. Tahir SK, Smith ML, Hessler P, Rapp LR, Idler KB, Park CH, 2014; 13:1442–56. https://doi.org/10.1158/1535-7163. Leverson JD, Lam LT. Potential mechanisms of resistance MCT-13-0849. to venetoclax and strategies to circumvent it. BMC Cancer. 41. Hardcastle IR, Arris CE, Bentley J, Boyle FT, Chen Y, 2017; 17:399. https://doi.org/10.1186/s12885-017-3383-5. Curtin NJ, Endicott JA, Gibson AE, Golding BT, Griffin RJ, 52. Lin KH, Winter PS, Xie A, Roth C, Martz CA, Stein EM, Jewsbury P, Menyerol J, Mesguiche V, et al. N2-substituted Anderson GR, Tingley JP, Wood KC. Targeting MCL-1/ O6-cyclohexylmethylguanine derivatives: potent inhibitors BCL-XL Forestalls the Acquisition of Resistance to ABT- of cyclin-dependent kinases 1 and 2. J Med Chem. 2004; 199 in Acute Myeloid Leukemia. Sci Rep. 2016; 6:27696. 47:3710–22. https://doi.org/10.1021/jm0311442. https://doi.org/10.1038/srep27696. www.impactjournals.com/oncotarget 107221 Oncotarget 53. Jorgensen TN, McKee A, Wang M, Kushnir E, White J, loss of MCL1. Leukemia. 2015; 29:1702–12. https://doi. Refaeli Y, Kappler JW, Marrack P. Bim and Bcl-2 mutually org/10.1038/leu.2015.99. affect the expression of the other in T cells. J Immunol. 57. Zeidner JF, Foster MC, Blackford AL, Litzow MR, Morris 2007; 179:3417–24. LE, Strickland SA, Lancet JE, Bose P, Levy MY, Tibes 54. Del Gaizo Moore V, Letai A. BH3 profiling--measuring R, Gojo I, Gocke CD, Rosner GL, et al. Randomized integrated function of the mitochondrial apoptotic pathway multicenter phase II study of flavopiridol (alvocidib), to predict cell fate decisions. Cancer Lett. 2013; 332:202–5. cytarabine, and mitoxantrone (FLAM) versus cytarabine/ https://doi.org/10.1016/j.canlet.2011.12.021. daunorubicin (7+3) in newly diagnosed acute myeloid leukemia. Haematologica. 2015; 100:1172–9. https://doi. 55. Pierceall WE, Kornblau SM, Carlson NE, Huang X, org/10.3324/haematol.2015.125849. Blake N, Lena R, Elashoff M, Konopleva M, Cardone MH, Andreeff M. BH3 profiling discriminates response 58. Chou TC, Talalay P. Quantitative analysis of dose-effect to cytarabine-based treatment of acute myelogenous relationships: the combined effects of multiple drugs or leukemia. Mol Cancer Ther. 2013; 12:2940–9. https://doi. enzyme inhibitors. Adv Enzyme Regul. 1984; 22:27–55. org/10.1158/1535-7163.MCT-13-0692. 59. Reynolds CP, Maurer BJ. Evaluating response to antineoplastic drug combinations in tissue culture models. 56. Li L, Pongtornpipat P, Tiutan T, Kendrick SL, Park S, Persky Methods Mol Med. 2005; 110:173–83. https://doi.org/1- DO, Rimsza LM, Puvvada SD, Schatz JH. Synergistic 59259-869-2:173. induction of apoptosis in high-risk DLBCL by BCL2 inhibition with ABT-199 combined with pharmacologic www.impactjournals.com/oncotarget 107222 Oncotarget

Journal

OncotargetPubmed Central

Published: Nov 3, 2017

References