Highly Active Combination of BRD4 Antagonist and Histone Deacetylase Inhibitor … · Small Molecule Therapeutics See related article by Liu et al., p. 1194 Highly Active Combination

Small Molecule TherapeuticsSee related article by Liu et al., p. 1194

Highly Active Combination of BRD4 Antagonist and HistoneDeacetylase Inhibitor against Human Acute MyelogenousLeukemia Cells

Warren Fiskus1, Sunil Sharma2, Jun Qi3, John A. Valenta1, Leasha J. Schaub1, Bhavin Shah1,Karissa Peth1, Bryce P. Portier1, Melissa Rodriguez1, Santhana G.T. Devaraj1, Ming Zhan1,Jianting Sheng1, Swaminathan P. Iyer1, James E. Bradner3, and Kapil N. Bhalla1

AbstractThe bromodomain and extra-terminal (BET) protein family members, including BRD4, bind to acetylated

lysines on histones and regulate the expression of important oncogenes, for example, c-MYC and BCL2. Here,

we demonstrate the sensitizing effects of the histone hyperacetylation-inducing pan–histone deacetylase

(HDAC) inhibitor panobinostat on human acute myelogenous leukemia (AML) blast progenitor cells (BPC) to

the BET protein antagonist JQ1. Treatment with JQ1, but not its inactive enantiomer (R-JQ1), was highly lethal

against AMLBPCs expressingmutantNPM1cþwith orwithout coexpression of FLT3-ITD orAML expressing

mixed lineage leukemia fusion oncoprotein. JQ1 treatment reducedbinding ofBRD4andRNApolymerase II to

the DNA of c-MYC and BCL2 and reduced their levels in the AML cells. Cotreatment with JQ1 and the HDAC

inhibitor panobinostat synergistically induced apoptosis of the AML BPCs, but not of normal CD34þ

hematopoietic progenitor cells. This was associated with greater attenuation of c-MYC and BCL2, while

increasing p21, BIM, and cleaved PARP levels in the AML BPCs. Cotreatment with JQ1 and panobinostat

significantly improved the survival of the NOD/SCID mice engrafted with OCI-AML3 or MOLM13 cells

(P < 0.01). These findings highlight cotreatmentwith a BRD4 antagonist and anHDAC inhibitor as a potentially

efficacious therapy of AML. Mol Cancer Ther; 13(5); 1142–54. �2014 AACR.

IntroductionAcetylation–deacetylation is among the several post-

translational modifications of the histones involved inregulating gene expression (1). The resulting histonestates or "marks" are recognized by the "reader" proteins,which assemble a complex of coregulatory proteins at theenhancers or promoters that initiate and regulate genetranscription (2, 3). Among these "reader" proteins is thefamily of bromodomain and extra-terminal (BET) pro-teins, includingBRD2 (bromodomain 2), BRD3, andBRD4(4). Structurally, BET proteins contain the N-terminaldouble, tandem, 110 amino acid-long bromodomains,which bind to the acetylated lysines on the nucleosomalhistones (5). BET proteins also contain an extra terminalprotein-interacting domain in the C-terminus, through

which they interact and recruit coactivators and corepres-sor complexes containing chromatin modifying enzymes,chromatin remodeling factors, and themediator elementsto the chromatin for regulating gene transcription (4, 5).The C-terminal domain (CTD) of BRD4 also includes aproline- and glutamine-rich unstructured region, similarto the highly phosphorylatable CTD of the RNA poly-merase II (5, 6). This region interacts with pTEFb (positivetranscription elongation factor b), the heterodimer com-posed of CDK9 and cyclin T,which phosphorylates serine2 on the CTD of RNA pol II for mRNA transcript elon-gation (5, 6). Thus, BRD4 and the other BET proteins havebeen shown to couple histone acetylation to transcriptelongation, especially at the promoters and enhancers ofimportant cell growth and survival genes such as c-MYC,cyclin D1, BCL2, and FOSL1 (7–9). The essential role ofBRD4 in mammalian cells is further supported by the factthat nullmutation of BRD4 is early embryonic lethal (4, 5).Recently, pertinent for therapy, an RNA interferencescreen identified BRD4 as an effective and promisingtarget in human acute myelogenous leukemia (AML;ref. 10). In addition, prompted by this, several structure/activity-based BET protein small-molecule inhibitorshave been developed, including JQ1 and I-BET151 (IB),which displace the BET proteins, along with the asso-ciated transcript initiation and elongation factors, fromthe chromatin (11–13). This results in transcriptional

Authors' Affiliations: 1Houston Methodist Research Institute, Houston,Texas; 2HuntsmanCancer Institute,University ofUtah,Salt LakeCity,Utah;and 3Dana-Farber Cancer Institute, Boston, Massachusetts

Note: Supplementary data for this article are available at Molecular CancerTherapeutics Online (http://mct.aacrjournals.org/).

Corresponding Author: Kapil N. Bhalla, Houston Methodist ResearchInstitute, 6670 Bertner Ave., R9–113, Houston, TX 77030. Phone: 713-441-9113; Fax: 713-441-5349; E-mail: [email protected]

doi: 10.1158/1535-7163.MCT-13-0770

�2014 American Association for Cancer Research.

MolecularCancer

Therapeutics

Mol Cancer Ther; 13(5) May 20141142

on July 18, 2020. © 2014 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst January 16, 2014; DOI: 10.1158/1535-7163.MCT-13-0770

http://mct.aacrjournals.org/

repression of BCL2, c-MYC, and CDK6 as well as inducesgrowth arrest and apoptosis of AML cells (14, 15). BETprotein inhibitors have been shown to be especially activeagainst AML carrying the mixed lineage leukemia (MLL)containing fusion oncoproteins (16).In previous reports, we demonstrated that treatment

with hydroxamic acid analog pan–histone deacetylase(HDAC) inhibitors, such as panobinostat (LBH589),induces hyperacetylation of histones as well as mediatesgrowth arrest and apoptosis of cultured and primaryAML cells (17, 18). Concomitantly, panobinostat treat-ment attenuated the levels of progrowth and prosurvi-val proteins, for example, BCL2 and c-MYC, whilesimultaneously inducing the levels of proapoptotic pro-tein BIM (17–19). On the basis of these observations, wehypothesized that the lysine hyperacetylation inducedby treatment with panobinostat would increase thedependency of AML cells on BET protein–regulatedtranscription of the oncogenes c-MYC and BCL2, suchthat this would make the AML cells especially suscep-tible to the activity of the BET (BRD4/2) protein antag-onist JQ1. Therefore, in the present studies, we deter-mined whether cotreatment with panobinostat wouldaugment the in vitro and in vivo effects of JQ1 againstcultured and primary AML cells. Our findings demon-strate that combined treatment with panobinostat andJQ1 is synergistically active against human AML blastprogenitor cells (BPC), including those with mutantNPM1cþ or MLL fusion oncoprotein with coexpressionof FLT3-ITD. In addition, we discovered that cotreat-ment with panobinostat and JQ is more effective thaneach agent alone in significantly improving the survivalof nonobese diabetic/severe combined immunode-ficient (NOD/SCID) mice engrafted with AML cellsexpressing mutant nucleophosmin 1 (NPM1cþ) or MLLfusion oncoprotein with FLT3-ITD.

Materials and MethodsReagents(S)-JQ1 (active enantiomer, hereafter referred to as

JQ1) and its inactive enantiomer (R)-JQ1 were devel-oped as previously described (11). IBET-151 wasobtained from Xcess Biologicals. Panobinostat waskindly provided by Novartis Pharmaceuticals Inc.Chemical structures for the molecules are provided asSupplementary Fig. S1. All compounds were preparedas 10 mmol/L stocks in 100% dimethyl sulfoxide(DMSO) and frozen at �20�C in 10 mL aliquots to allowfor single use, thus avoidingmultiple freeze-thaw cyclesthat could result in compound decomposition and lossof activity. Anti-BRD4 antibody for chromatin immu-noprecipitation (ChIP) and Western blot analysis wasobtained from Bethyl Laboratories. Anti p-SER2 POL IIantibody and RNA POL II antibody for ChIP wereobtained from Millipore. Anti–cleaved PARP, anti–c-MYC, and anti-BIM antibodies were obtained from CellSignaling Technology. Anti-BCL2 and anti-CDK6 anti-

bodies were obtained from Santa Cruz Biotechnologies.Anti-p21WAF antibody was obtained from Neomar-kers. Anti-p27KIP antibody and anti-FLT3-PE–conju-gated antibody were obtained from BD Biosciences.Anti-NPM1, anti–b-actin antibody, and lentiviral shorthairpin RNAs (shRNA) targeting BRD4 or nontargetingshRNA (sh-NT) were obtained from Sigma Aldrich.

Cell cultureOCI-AML2, OCI-AML3, and MOLM13 cells were

obtained from the DSMZ. HL-60, U937, and MV4-11 cellswere obtained from American Type Culture Collection(ATCC). All experiments with cell lines were performedwithin 6months after thawing or obtaining fromATCCorDSMZ. Cell line authentication was performed by ATCCor DSMZ. The ATCC and DSMZ use short tandem repeatprofiling for characterization and authentication of celllines.

Primary normal progenitor and AML BPCsPrimary peripheral blood and/or bone marrow aspi-

rate AML samples were obtained and prepared for thestudies below, as previously described (17, 19, 20).Banked, delinked, and deidentified, normal or AMLCD34þ or AML CD34þCD38-LIN� bone marrow progen-itor/stem cells were purified, as previously described(18). The clinical presentation of the patient and mutationstatus of theprimaryAMLsamplesused in these studies isprovided in Supplementary Table S1.

Cell-cycle analysisFollowing the designated treatments with JQ1, cell-

cycle status was analyzed on a BD Accuri C6 flow cyt-ometer (BD Biosciences), as previously described (19).

Assessment of apoptosis by Annexin-V stainingUntreated or drug-treated cells were stained with

Annexin-V (Pharmingen) and TO-PRO-3 iodide and thepercentages of apoptotic cells were determined by flowcytometry, as previously described (18, 19). The com-bination index (CI) for each drug combination and theevaluation of the synergistic interactions were calculat-ed by median dose effect analyses (assuming mutualexclusivity) using the commercially available softwareCalcusyn (Biosoft; ref. 21). CI values of less than 1.0represent a synergistic interaction of the two drugs inthe combination.

Assessment of percentage nonviable cellsFollowing designated treatments, cells were washed

with 1� PBS, stained with propidium iodide, and ana-lyzed by flow cytometry, as previously described (19, 20).

Colony growth assayCultured AML cells were treated with JQ1 and/or

panobinostat for 48 hours. At the end of treatment, cellswere washed free of the drugs and 500 cells per conditionwere plated in methylcellulose and incubated at 37�C.

BET Protein Antagonist and HDAC Inhibitor against AML

www.aacrjournals.org Mol Cancer Ther; 13(5) May 2014 1143




Colony formationwasmeasured 7 to 10 days after plating(19, 20).

ChIP and real-time PCROCI-AML3, OCI-AML2, MOLM13, and primary AML

cells were treated with JQ1 for 16 hours. Following drugexposure, crosslinking, cell lysis, sonication, and ChIP forBRD4 or POL II were performed according to the manu-facturer’s protocol (Millipore). For quantitative assess-ment of binding of BRD4 or RNA POL II to the c-MYC,BCL2, and CDK6 promoter in the chromatin immunopre-cipitates, a SYBR Green PCR Mastermix from AppliedBiosystems was used. Relative enrichment of the promot-er DNA in the chromatin immunoprecipitates was nor-malized against the amount of c-MYC, BCL2, and CDK6promoter DNA in the input samples (19).

shRNA to BRD4Lentiviral shRNAs targeting BRD4 or nontargeting

shRNA (sh-NT) were transduced into OCI-AML3 cells,as previously described (20). Forty-eight hours after trans-duction, the cells were washed with complete media andplated with or without panobinostat for 48 hours forassessing apoptosis.

RNA isolation and quantitative PCRAfter the designated treatments with JQ1, total RNA

was isolated from cultured and primary AML cells with aHigh Pure RNA Isolation Kit (Roche Diagnostics) andreverse transcribed. Quantitative real-time PCR analysisfor the expression of c-MYC, BCL2, CDK6, and p21 wasperformed on cDNA using TaqMan probes fromAppliedBiosystems (19, 20). Relative mRNA expression was nor-malized to the expression of glyceraldehyde-3-phosphatedehydrogenase (GAPDH).

Gene expression microarray analysisTotal RNA from OCI-AML3 cells treated with JQ1 for

8 hours was extracted using the RNeasy Plus Mini Kit(Qiagen). One microgram of RNA was used for thegeneration of labeled cRNA and hybridization of thelabeled cRNA fragments, washing, staining, and scan-ning of the arrays were performed according to themanufacturer’s instructions. Labeled cRNAs were pro-filed using the Affymetrix Human Genome-U133-Plus2.0 microarray (14). Array data were imported intoAffymetrix Expression Console (Affymetrix), and nor-malized with robust multiarray average (RMA) method.Relative fold change analysis was calculated using Par-tek Genomic Suite. Genes were selected for P < 0.01 and2 fold changes of treated versus untreated. A one-wayheatmap for the selected genes was generated from JMP8 (SAS Inc.). The microarray data are from the analysesof data from an experiment performed in duplicate andare representative of two separate experiments. Allmicroarray data used in this article are deposited inthe Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under the accession number GSE51950.

Ingenuity pathway analysisSignificantly perturbed gene lists acquired from the

microarray analysis of JQ1-treated OCI-AML3 cells wereimported into the Ingenuity Pathway Analysis Tool (IPATool; Ingenuity Systems Inc.; http://www.ingenuity.com) for assignment of biologic function and identifica-tion of differentially altered genetic networks. Up- anddownregulated identifiers were defined as value para-meters for the analysis. Within the IPA software, a coreanalysis was performed to identify the signaling andmetabolic pathways as well as the molecular networksand biologic processes that were significantly perturbedin the differentially expressed gene (DEG) dataset. TheDEGs were mapped to genetic networks in the databaseand rankedby score. The significance of themolecular andcellular functions overrepresented in the gene set, as wellas the signaling pathways and biologic networks towhichthey belong was tested by Fisher exact test P value. Thecreated biologic networks were ranked according to thenumber of significantly differentially expressed genesthat they contained.

Cell lysis, protein quantitation, and immunoblotanalyses

Untreated or drug-treated cells were centrifuged andthe cell pelletswere lysedand theproteinquantitation andimmunoblot analyses were performed, as previouslydescribed (19, 20). Immunoblot analyses were performedat least twice. Representative immunoblots were sub-jected to densitometry analysis. Densitometry was per-formed using ImageQuant 5.2 (GE Healthcare).

In vivo model of AMLAll in vivo studies were approved by and conducted in

accordancewith the guidelines of the InstitutionalAnimalCare and Use Committee at Houston Methodist ResearchInstitute (Houston, TX). Female NOD/SCID mice wereexposed to 2.5 Gy of radiation. The following day, 5million OCI-AML3 or MOLM13 cells were injected intothe lateral tail vein of the mice and the mice were mon-itored for 4 to 7 days. Following treatments were admin-istered in cohorts of 8 mice for each treatment: vehiclealone, 50mg/kg JQ1, 5mg/kgpanobinostat, and JQ1 pluspanobinostat. Treatments were initiated on day 7 for OCI-AML3 and on day 4 forMOLM13 cells. JQ1 [formulated in10% 2-hydroxypropyl-b-cyclodextrin (CAS 128446-35-5)]was administered daily for 5 days per week (Monday–Friday) intraperitoneally for 3 weeks, and then discon-tinued. Panobinostat (formulated in 5% DMSO/95% nor-mal saline) was administered by intraperitoneal injection3 days per week (Monday, Wednesday, and Friday) for 3weeks and discontinued. The survival of mice from bothin vivomodels is represented by a Kaplan–Meier survivalplot. The doses of JQ1 and panobinostat used in thesestudies were determined to be safe and effective throughpreviously reported studies (11, 19). A separate in vivoexperiment was conducted for analysis of biomarkersusing OCI-AML3 cells. Following engraftment of the

Fiskus et al.

Mol Cancer Ther; 13(5) May 2014 Molecular Cancer Therapeutics1144




AML cells, mice were treated with JQ1 and/or panobino-stat, as described above, for 1week. Six hours after the lastdoseofpanobinostat, themicewerehumanely euthanizedand bone marrow was collected from both femurs forimmunoblot analyses.

Statistical analysisSignificant differences between values obtained in a

population of AML cells treated with different experi-mental conditions were determined using a two-tailed,paired t test or a one-way ANOVA analysis within ananalysis packageofMicrosoft Excel 2010 software orusingGraphPad Prism (GraphPad Software). P values of lessthan 0.05 were assigned significance. Statistical differ-ences in the survival of the mice treated with JQ1, pano-binostat, or JQ þ panobinostat were determined by log-rank (Mantel–Cox) test. P values of less than 0.05 wereassigned significance.

ResultsJQ1 exerts growth inhibitory and lethal effects incultured and primary AML BPCsWe first determined the lethal activity of the BET pro-

tein inhibitor JQ1 against several cultured human AMLcell lines, including those that express mutant NPM1cþ(OCI-AML3) or expressed MLL fusion oncoprotein withor without FLT3-ITD (MV4-11 andMOLM13; refs. 17, 20).Treatment with JQ1 dose dependently increased the per-centage of cells in the G1 phase while reducing the per-centage of S-phase cells, as well as concomitantly inducedapoptosis in the AML cell types, as has been previouslyreported (Fig. 1A and B; refs. 10, 15, 16). Among these,OCI-AML3 cells exhibited greater sensitivity to JQ1 thanMOLM13 and MV4-11 cells (Fig. 1A). The IC50 values forinducing apoptosis after 48-hour exposure to JQ1 were165 nmol/L for OCI-AML3 versus 280 nmol/L and 1,480nmol/L for MV4-11 andMOLM13 cells, respectively. JQ1treatment also inhibited the clonogenic survival of OCI-AML3 cells more than ofMOLM13, OCI-AML2, or HL-60cells (Fig. 1C). Because cultured AML cells coexpressingendogenous mutant NPM1cþ and FLT3-ITD, a common-ly encountered normal karyotype AML cell type in theclinic (22), have not been isolated and hence are unavail-able, we created OCI-AML3 cells with ectopic expressionof FLT3-ITD (OCI-AML3/FI cells) for determining theactivity of JQ1 against these cells. As compared with thecontrol OCI-AML3, OCI-AML3/FI cells exhibit higherexpression of FLT3 protein, as determined both by flowcytometry and immunoblot analysis (Fig. 1D and E).However, JQ1was equally effective in inducing apoptosisof OCI-AML3 and OCI-AML3/FI cells (Fig. 1E). Treat-ment with the inactive enantiomer of JQ1, that is, R-JQ1,neither perturbed the cell cycle nor induced apoptosis inOCI-AML3 or MOLM13 cells (Supplementary Fig. S2A–S2C; ref. 11). We also determined the lethal effects of JQ1against 10 separate samples of patient-derived primaryAML cells with normal karyotype expressing mutant

NPM1 and/or FLT3-ITD. Similar to the effects observedin the cultured AML cells lines, treatment with JQ1 alsodose dependently exerted lethal antileukemia effectsagainst primary CD34þ AML cells (Fig. 1F). Although,the loss of viability in the primary AML cells was lowerthan in the cultured AML cell types, we observed nosignificant difference in JQ1-induced loss of viability inAML cells with mutant NPM1cþ alone versus those thatcoexpressed FLT3-ITD (P > 0.05; Fig. 1F).

JQ1 inhibits the binding of BRD4 and RNA POL IIand attenuates the mRNA expression of c-MYC andBCL2 in AML cells

Disruption of binding of the bromodomain of BET pro-tein to acetylated histones has been shown to deplete theBET protein occupancy on the chromatin associated withthe promoters of BET protein target genes (14, 15). Consis-tent with this, ChIP analyses showed that treatment withJQ1 reduced the BRD4 occupancy at the promoters of c-MYC, BCL2, and CDK6 in OCI-AML3 and MOLM13 cells(Fig. 2A and Supplementary Fig. S3A). We also observedthat JQ1 reduced the occupancy of BRD4 at the samepromoters in NPM1cþ-expressing primary AML cells, aspresented in the data in Fig. 2B, representative of twoprimary AML samples. BRD4 is known to regulate thetranscriptional elongation of these genes through recruit-ment of P-TEFb, which phosphorylates and activates RNAPOL II (5, 6). Accordingly, we also found that treatmentwith JQ1 reduced the binding of RNA POL II to thepromoters of c-MYC, BCL2, and CDK6 genes in the OCI-AML3 as well as in the primary AML cells (Fig. 2C and D).Data in Fig. 2D are also representative of twoprimaryAMLsamples. Similar effects were also observed in JQ1-treatedcultured AML OCI-AML2 cells that express wild-typeNPM1 (Supplementary Fig. S3B and S3C). QuantitativePCR (qPCR) analyses of the gene expressions showed thatJQ1 treatment attenuated themRNA expression of c-MYC,BCL2, and CDK6 in the OCI-AML3, MOLM13, and MV4-11, as well as in primary AML cells expressing NPM1cþ(Fig. 3A and B; Supplementary Figs. S4A, S4B, and S5A). Incontrast, in these cell types, JQ1 treatment concomitantlyupregulated the mRNA and protein expression of p21(Supplementary Figs. S4C, S5B, and S5C). In tumor cells,high levels of c-MYC cause transcriptional amplification ofthe gene expression program involving a larger number ofgenes (23). Therefore,we also determined the effects of JQ1treatment on gene expression microarray profile in OCI-AML3cells. Figure3Cshowsaheatmapof thegreatest geneexpression changes following treatment with JQ1 for 8hours.As shown, JQ1 treatmentdownregulated themRNAexpression ofmore genes, as comparedwith the number ofgeneswhosemRNAexpressionwas upregulated (Fig. 3C).The fold changes in the most altered mRNA gene expres-sions are shown in Supplementary Table S2. Data sets ofgenes with altered expression profile derived from micro-array analyses were imported into the Ingenuity PathwayAnalysis (IPA) Tool (Ingenuity H Systems; http://www.ingenuity.com).Within the gene list, IPA identified the top






fivemost perturbednetworks inOCI-AML3 cells followingtreatment with JQ1 and assigned a score for these associ-ated network functions (Supplementary Table S3). Thescore (i.e., a score of 36) assigned by IPA indicates theprobability (1 in 1036) that the focus genes in the dataset aregrouped together in a perturbed network due to randomchance alone.Next, total RNA from the untreated and JQ1-treated cells used for the microarray analysis was alsoreverse transcribed and the resulting cDNA was used forqPCR analysis using c-MYC and BCL2-specific TaqMan

real-time PCR probes. This confirmed that JQ1 treatmentmarkedly decreased the mRNA expression of the c-MYCand BCL2 genes (Fig. 3D).

JQ1 treatment depletes p-Serine 2 RNA POL II, c-MYC, and BCL2 but induces p21 and BIM proteinlevels

We next compared the effects of JQ1 treatment on theprotein expression of BRD4, c-MYC, BCL2, CDK6, andpSer2 RNA POL II in OCI-AML3 versus MOLM13 cells.

B

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Control 100 250 500

% o

f cells

nmol/L, JQ1, 24 h

MOLM13

G2–M S G0–G1

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Control 50 100 250

% o

f cells

nmol/L, JQ1, 24 h

OCI-AML3

G2–M S G0–G1

C

0

20

40

60

80

100

120

Control 100 250

% C

olo

ny g

row

th

(re

lati

ve t

o c

on

tro

l)

nmol/L, JQ1, 48 h

nmol/L, JQ1, 48 h

HL-60

OCI-AML2

MOLM13

OCI-AML3

MV4-11F

0

10

20

30

40

50

60

70

80

90

100

Control 100 250 500

% A

po

pto

sis

nmol/L JQ1, 48 h

HL-60

U937

OCI-AML2

MOLM13

MV4-11

OCI-AML3

A

P = 0.045

P = 0.019

P = 0.027

D

0

10

20

30

40

50

60

70

80

90

Control 100 250 500

% A

po

pto

sis

nmol/L, JQ1, 48 h

OCI-AML3OCI-AML3/FI

Ve

cto

r

FL

T-I

TD

b-Actin

FLT3

EIsotype controlOCI-AML3

OCI-AML3/FLT3

0200

400

600

800

101 102 103 104 105 106 107.2

Co

un

t

FL-2 (FLT3 surface expression)

CD34+ primary AML cells (n = 10)

% N

on

via

ble

ce

lls

Figure 1. Treatment with the BET protein antagonist JQ1 induces cell-cycle growth arrest and lethal effects in cultured and primary AML BPCs. A,HL-60, U937, OCI-AML2, MOLM13, MV4-11, and OCI-AML3 cells were treated with the indicated concentrations of JQ1 for 48 hours. Following this, thepercentages of Annexin-V–positive, apoptotic cells were determined by flow cytometry. Columns, mean of three independent experiments; bars, SEM. B,OCI-AML3 and MOLM13 cells were treated with the indicated concentrations of JQ1 for 24 hours. Cell-cycle status was determined by flow cytometry.Columns, mean of three independent experiments; bars, SEM. C, HL-60, OCI-AML2, MOLM13, OCI-AML3, and MV4-11 cells were treated with theindicated concentrations of JQ1 for 48 hours. At the end of treatment, cells were washed free of the drug and plated in methylcellulose for 7 to 10 days. Thenumber of colonies in each conditionwas counted and thepercentage colony growthwasdetermined relative to the untreated control cells. Columns,meanofthree independent experiments; bars, SEM. D, OCI-AML3 cells were transduced with lentivirus expressing FLT3-ITD. The expression of FLT3 in thecells was detected by flow cytometry using anti-FLT3 antibody. E, OCI-AML3 cells with or without ectopic expression of FLT3-ITD were treated with theindicated concentrations of JQ1 for 48 hours. At the end of treatment, cells were stained with Annexin-V and TO-PRO-3 iodide, and the percentages ofAnnexin-V–positive, apoptotic cells were determined by flow cytometry. Columns, mean of three experiments; bars, SEM. IC50 values for JQ1 werecalculated for both cell lines. The IC50 value for OCI-AML3 and OCI-AML3/FI is 165 nmol/L and 188.4 nmol/L, respectively. Inset shows the FLT3 proteinexpression in the two cell lines. Expression of b-actin in the lysates served as the loading control. F, primary CD34þ AML cells (n ¼ 10) were exposed to theindicated concentrations of JQ1 for 48 hours. Then, cells were stained with propidium iodide (PI) and the percentages of PI-positive, nonviable cellswere determined by flow cytometry. The plot shows the individual values for each sample. The horizontal line indicates themedian loss of viability for all of thesamples. Bar, SEM.

Fiskus et al.





Asshown in the immunoblot analyses in Fig. 4A, althoughit had no effect on BRD4 and NPM1 (not shown), JQ1treatment dose dependently depleted the protein levelsof c-MYC, BCL2, CDK6, and pSer2 RNAPOL II, as well asinduced the levels of p21, p27, BIM, and cleaved PARP toa similar extent in both OCI-AML3 and MOLM13 cells.Exposure to (R)-JQ1 did not alter the levels of these pro-teins in either the OCI-AML3 or MOLM13 cells (Fig. 4Band data not shown). In a representative primary AMLBPC sample that expressed mutant NPM1cþ and FLT3-ITD, JQ1 treatment also caused a marked decline inc-MYC, BCL2, and p-Ser2 RNA POL II levels, with a con-comitant increase in the levels of p21, p27, and BIM, aswellas exerted no effect on NPM1 levels (Fig. 4C). Figure 4Dshows the mean � SEM for the decline in c-MYC, andinduction of BIM and p21 protein levels in three primaryAML samples (Fig. 4D). Collectively, these data showthat, in addition to its known activity in AML cells expres-sing MLL fusion oncoprotein, JQ1 treatment is effectiveagainst cultured and primary AML cells expressing mu-tant NPM1cþ irrespective of coexpressed FLT3-ITD.

Cotreatment with JQ1 and panobinostat issynergistically active against cultured AMLcells expressing mutant NPM1cþ or MLLfusion oncoproteinWe next determined whether cotreatment with the

potent pan-HDAC inhibitor such as panobinostat, knownto induce in vitro and in vivo lysine acetylation of histones,

would increase the dependency on BRD4-regulated pro-growth and prosurvival gene expressions and therebysensitizeAMLBPCs to JQ1-inducedapoptosis (19, 24).Fig-ure 5A and Supplementary Fig. S6A to S6C demonstratethat cotreatment with panobinostat and JQ1 synergisti-cally induced apoptosis of OCI-AML3, MOLM13, MV4-11, and HL-60 cells, with combination indices of less than1.0 by the isobologram analyses. The specific activity ofJQ1 in this synergistic interaction is underscored by thefact that cotreatment with (R)-JQ1 did not enhance pano-binostat-induced apoptosis of OCI-AML3 and MOLM13cells (Fig. 5B). Cotreatment with another BET proteinantagonist, I-BET151, and panobinostat also synergisti-cally induced apoptosis of AML cells (SupplementaryFig. S6D). As compared with each agent alone, combinedtreatment with JQ1 and panobinostat also induced grea-ter loss of the clonogenic survival of OCI-AML3 andMOLM13, as well as of the other AML cell types (Fig.5C andSupplementary Fig. S6E). Thiswas associatedwithmarkeddepletion in the protein levels of p-Ser2 RNAPOLII and c-MYC,with concomitant upregulation of the levelsof BIM protein isoforms (Fig. 5D and Supplementary Fig.S6F). Panobinostat and JQ1 combination did not alter thetotal NPM1 levels in OCI-AML3 cells (Fig. 5D). Takentogether, these findings show that cotreatmentwith pano-binostat sensitizes cultured AML cells to the anti-AMLactivity of a BRD4 antagonist.Wenext determinedwheth-er specific depletion of BRD4 by shRNA would pheno-copy the effects of JQ1 in increasing the anti-AML activity

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0 500

Fo

ld e

nri

ch

me

nt

(re

lati

ve

to

un

tre

ate

d)

nmol/L, JQ1, 16 h

OCI-AML3 BRD4 ChIP

c-MYC promoter

BCL2 promoter

CDK6 promoter

A

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0 250

Fo

ld e

nri

ch

me

nt

(re

lati

ve

to

un

tre

ate

d)

nmol/L, JQ1, 16 h

Primary AMLBRD4 ChIP

c-MYC promoter

BCL2 promoter

CDK6 promoter

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 250

Fo

ld e

nri

ch

me

nt

(re

lati

ve

to

un

tre

ate

d)

nmol/L, JQ1, 16 h

Primary AMLPOL II ChIP

c-MYC promoter

BCL2 promoter

CDK6 promoter

D

B

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 500

Fo

ld e

nri

ch

me

nt

(re

lati

ve

to

un

tre

ate

d)

nmol/L, JQ1, 16 h

OCI-AML3 POL II ChIP

c-MYC promoter

BCL2 promoter

CDK6 promoter

C

Figure 2. Treatment with JQ1reduces BRD4 and Pol IIoccupancy on the promoters of c-MYC, BCL2, and CDK6 in humanAMLcells. A andB,OCI-AML3 andprimary AML cells were treatedwith the indicated concentration ofJQ1 for 16 hours. Following this,ChIP was conducted with BRD4-specific antibody. The chromatinimmunoprecipitated DNA wassubjected to quantitative real-timePCR with primers for the promoterof c-MYC, BCL2, and CDK6. Thefold enrichment was calculatedusing theCt value of the ChIP DNAcompared with the Ct value of theinput DNA. C and D, OCI-AML3and primary AML cells weretreated with the indicatedconcentration of JQ1 for 16 hours.Following this, ChIP wasconducted with RNA Pol IIantibody. The chromatinimmunoprecipitated DNA wassubjected to quantitative real-timePCR with primers for the promoterof c-MYC, BCL2, and CDK6. Thefold enrichment was calculatedusing theCt value of the ChIP DNAcompared with the Ct value of theinput DNA.






of panobinostat. Figure 5E demonstrates that, as com-pared with treatment with the nontargeted controlshRNA, treatmentwith BRD4 shRNA reduced themRNAlevels of BRD4, c-MYC, and BCL2, while simultaneouslyincreasing the mRNA levels of p21. In the cells treatedwith BRD4 shRNA, but not the nontargeted shRNA,panobinostat treatment induced significantly more apo-ptosis of OCI-AML3 cells (P < 0.05; Fig. 5F). The IC50

values for panobinostat were significantly lower in OCI-AML3 cells treated with BRD4 shRNA versus those trea-ted with nontargeted shRNA (P ¼ 0.0165; Fig. 5G).

Cotreatment with JQ1 and panobinostat exertssynergistic lethal activity against primary AMLBPCsbut not normal hematopoietic progenitor cells

Wenext compared the lethal activity of the combinationof panobinostat and JQ1 against 9 samples of primaryAML versus CD34þ normal hematopoietic progenitor

cells. Figure 6A (and the inset) demonstrates that, ascompared with the treatment with each agent alone,cotreatment with panobinostat and JQ1 induced signifi-cantly more apoptosis with increased PARP cleavage inprimaryAMLBPCs that expressedmutantNPM1cþwithor without the coexpression of FLT3-ITD. Although treat-mentwith JQ1 alone (500 nmol/L) depleted the levels of c-MYC and BCL2 and induced p21 levels, cotreatment withpanobinostat and JQ1 caused more downregulation of c-MYC and BCL2 as well as induced more p21 and BIMlevels in the AML BPCs (Fig. 6B). Importantly, the com-bination of panobinostat and JQ1 exerted synergisticlethal activity in the subpopulation of CD34þCD38-Lin�

BPCs (Fig. 6C). In this combination, the levels of panobi-nostat used are clinically achievable and safe and havebeen demonstrated to induce in vivo histone acetylation incells of patientswithAML (24).Notably, cotreatmentwithpanobinostat and JQ1 did not exert significantly greater

A

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 500

Rel

ativ

e m

RN

A e

xpre

ssio

n

nmol/L, JQ1, 16 h

OCI-AML3 c-MYC

BCL2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 500R

elat

ive

mR

NA

exp

ress

ion

nmol/L, JQ1, 16 h

MOLM13 c-MYC

BCL2

B

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 500

Rel

ativ

e m

RN

A e

xpre

ssio

n

nmol/L, JQ1, 16 h

Primary AML c-MYC

BCL2

D

00.20.40.60.8

11.21.4

Control 500

Rel

ativ

e m

RN

A

expr

essi

on

nmol/L, JQ1, 8 h

OCI-AML3 BCL2

c-MYC

Control JQ1

c-MYC

C

Figure 3. Treatment with JQ1 depletes the mRNA expression of c-MYC and BCL2 in AML cells. A and B, OCI-AML3, MOLM13, and primary AML cells weretreated with the indicated concentrations of JQ1 for 16 hours. At the end of treatment, RNA was isolated and reverse transcribed. The resulting cDNAwas used for real-time quantitative PCR analysis of c-MYC and BCL2. The relative mRNA expression was normalized to GAPDH and compared with theuntreated cells. C, heatmap showing the 50 most up- and downregulated mRNA expression changes in OCI-AML3 cells treated with 500 nmol/L ofJQ1 for 8 hours.D, relativemRNAexpression changesofBCL2andc-MYC in themRNAused for expression analyseswerequantifiedbyquantitative real-timePCR. Relative expression changes were normalized to GAPDH and compared with the untreated cells.

Fiskus et al.





lethal activity against normal CD34þ hematopoietic pro-genitor cells (Fig. 6D). These findings highlight the anti-AML selectivity of the cotreatmentwith panobinostat andJQ1 against AML BPCs.

Combined treatment with panobinostat and JQ1exerts superior in vivo activity against theestablished AML xenografts in NOD/SCID miceWe next determined the in vivo anti-AML activity of

panobinostat and/or JQ1 against the OCI-AML3 AMLxenografts engrafted in the bone marrow of the NOD/SCID mice. Following the tail vein infusion and engraft-ment of OCI-AML3 cells in the bone marrow of the

NOD/SCID mice, the anti-AML effects and the survivalimprovement due to treatment with daily intraperito-neal JQ1 and/or intraperitoneal panobinostat (Monday,Wednesday, and Friday) for 3 weeks were comparedwith the effects of the treatment with the vehicle alone.The Kaplan–Meier plot depicting the survival of micedemonstrated that, as compared with treatment withthe vehicle alone, treatment with either JQ1 or panobi-nostat significantly improved the survival of the miceinfused with OCI-AML3 cells (P < 0.05; Fig. 7A). Nota-bly, combined treatment with panobinostat and JQ1further significantly improved survival of the mice, ascompared with treatment with JQ1 or panobinostat

c-MYC

p21

NPM1

p27

b-Actin

BRD4

Primary AML

0 100 500 nmol/L, JQ1, 24 h

BIMEL

BIMS

BIML

C

D

0

0.2

0.4

0.6

0.8

1

1.2

0 500

Re

lati

ve

pro

tein

ex

pre

ss

ion

nmol/L, JQ1, 24 h

Primary AML (n = 3)c-Myc

0

1

2

3

4

5

6

7

8

0 500

Re

lati

ve

pro

tein

ex

pre

ss

ion

nmol/L, JQ1, 24 h

Primary AML (n = 3)BIM

p21

**

*

B

c-MYC

b-Actin

BCL2

CDK6

pSer2 POL II

OCI-AML3

0 100 250 500 0 100 250 500 nmol/L, (R)-JQ1, 24 h

MOLM13

AOCI-AML3

0 100 500

pSer2 POL II

c-MYC

p21

p27

b-Actin

BRD4

MOLM13

0 100 500 nmol/L, JQ1, 24 h

BIMEL

BIMS

BIML

CL. PARP

CDK6

BCL2

p23

pSer2 POL II

BCL2

Figure 4. Treatment with JQ1 depletes p-Serine 2 RNA POL II, c-MYC, and BCL2 but induces p21, p27, and BIM protein levels in AML cells. A, OCI-AML3and MOLM13 cells were treated with the indicated concentrations of JQ1 for 24 hours. Following this, total cell lysates were prepared and immunoblotanalyses were conducted for the expression levels of BRD4, c-MYC, BCL2, CDK6, pSer2 POL II, p21, p27, BIM, cleaved PARP, and b-actin in the lysates.B, OCI-AML3 and MOLM13 cells were treated with the indicated concentrations of the inactive enantiomer of JQ1, (R)-JQ1, for 24 hours. At the end oftreatment, cell lysates were prepared and immunoblot analyses were conducted for the expression levels of c-MYC, BCL2, CDK6, pSer2 POL II, andb-actin in the lysates. C, primary AML cells were treated with the indicated concentrations of JQ1 for 24 hours. Then, total cell lysates were prepared andimmunoblot analyses were conducted for the expression levels of BRD4, c-MYC, BCL2, NPM1, pSer2 POL II, p21, p27, BIM, and b-actin in thelysates. D, primary AML cells (n ¼ 3) were treated with the indicated concentrations of JQ1 for 24 hours. After this, total cell lysates were prepared andimmunoblot analyses were conducted for the expression levels of c-MYC, p21, and BIM in the lysates. Densitometry analysis was performed and thepercentage depletion of c-MYC (left) or the induction of BIM and p21 (right) was calculated. Columns, mean depletion of c-MYC or inductionof p21 and BIM in the three AML samples; bars, SEM.






alone (P < 0.001). In cohorts of three mice treated withthe vehicle control versus treatment with JQ1 and/orpanobinostat for 5 days, bone marrow was harvested,and the cell lysates were analyzed for proteinexpression. Figure 7B demonstrates that as comparedwith treatment with each agent alone, cotreatment withpanobinostat and JQ1 was associated with the mostreduction in the levels of c-MYC, BCL2, and CDK6

proteins. We also determined the effects of JQ1 and/orpanobinostat on the expression levels of BIM. Asshown in Fig. 7B, although treatment with each drugalone increased the levels of BIM, cotreatment with JQ1and panobinostat did not exhibit further increase in thelevels of BIM in the mice. We also determined the in vivoanti-AML activity of panobinostat and/or JQ1 againstthe more aggressive MOLM13 xenograft model, in

A

0 0.2 0.4 0.6 0.8 1.00

0.5

1.0

1.5

Fractional effect

CI

OCI-AML3

JQ1 and PS

0 0.2 0.4 0.6 0.8 1.00

0.5

1.0

1.5

Fractional effect

CI

MOLM13

JQ1 and PS

B

0

10

20

30

40

50

60

70

80

90

100

% A

po

pto

sis

OCI-AML3

MOLM13*

*

C

0

20

40

60

80

100

120

% C

olo

ny g

row

th OCI-AML3

MOLM13

D

c-MYC

NPM1

OCI-AML3

BRD4

0 250 0 250 nmol/L, JQ1, 24 h

+ 10 nmol/L PS

BIMEL

BIMS

BIML

b-Actin

pSer2 POL II

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

NT shRNA BRD4 shRNA

Re

lati

ve

mR

NA

ex

pre

ss

ion OCI-AML3

BRD4c-MYCBCL2p21

E

F

0

20

40

60

80

100

120

0 5 10 20

% A

po

pto

sis

nmol/L, PS, 48 h

OCI-AML3

NT shRNA

BRD4 shRNA

*

*

G

IC5

0,n

mo

l/L

,PS

NT shRNA BRD4 shRNA0

5

10

15

20

OCI-AML3

P = 0.0165

*

100 nmol/L R-JQ1 + 10 nmol/L

PS

Control Contro

l

100 nmol/L JQ1

100 nmol/L JQ1

5 nmol/L PS

JQ1 + PS

100 nmol/L R-JQ1

10 nmol/L PS

100 nmol/L JQ1 + 10 nmol/L

PS

Figure 5. Cotreatment with JQ1 or shRNA-mediated knockdown of BRD4 synergistically enhances panobinostat-induced apoptosis of AML cells. A, OCI-AML3 and MOLM13 cells were treated with JQ1 and panobinostat (PS) for 48 hours. Then, the percentages of Annexin-V–positive, apoptotic cells weredetermined by flow cytometry. Median dose effect and isobologram analyses were performed using the commercially available software CalcuSyn,assuming mutual exclusivity. CI values less than 1.0 indicate a synergistic interaction of the two agents in the combination. B, OCI-AML3 and MOLM13cells were treated with the indicated concentrations of JQ1 or (R)-JQ1 and/or panobinostat for 48 hours. At the end of treatment, the percentages ofAnnexin-V–positive, apoptotic cells were determined by flow cytometry. Columns, mean of three independent experiments; bars, SEM. �, values significantlygreater in the combination than treatment with JQ1 or panobinostat alone (P < 0.05). C, OCI-AML3 and MOLM13 cells were treated with JQ1 and/orpanobinostat for 48 hours. Following this, cells were washed free of the drug, plated in methylcellulose, and cultured for 7 to 10 days. Colonies were countedand the colony growth for each cell line is reported as a percentage of the untreated cells. Columns,mean of three experiments; bars, SEM.D, OCI-AML3 cellswere treated with the indicated concentrations of JQ1 and/or panobinostat for 24 hours. Following this, cells were harvested and total cell lysates wereprepared. Immunoblot analyses were conducted for the expression levels of BRD4, p-Ser2 POL II, c-MYC, BIM, NPM1, and b-actin in the lysates. E,OCI-AML3 cells were transduced with nontargeting (NT) or BRD4-specific shRNA-containing lentivirus for 48 hours. Then, total RNA was harvested, reversetranscribed, and qPCR was conducted for the indicated targets. The relative expression of each mRNA was normalized to GAPDH. F, OCI-AML3 cellstransfected with NT shRNA or BRD4 shRNA for 48 hours were treated with the indicated concentrations of panobinostat for 48 hours. At the end oftreatment, cells were stained with Annexin-V and TO-PRO-3 iodide and the percentages of Annexin-V–positive, apoptotic cells were determined by flowcytometry. �, values significantly greater in BRD4 shRNA-transfected cells compared with NT shRNA-transfected cells. G, IC50 values for panobinostatin OCI-AML3 cells transduced with NT-shRNA or BRD4 shRNA for 48 hours, then treated with panobinostat for 48 hours.

Fiskus et al.





which all mice treated with vehicle control succumbedto AML in less than 25 days. As shown in Fig. 7C, thecombination regimen of JQ1 and panobinostat for 3weeks was superior to the treatment with JQ1 or pano-binostat alone in improving the median survival andoverall survival of the mice (P < 0.0001), which trans-lated into a plateau in the survival curve. This suggests apotentially curative impact of the combination on thesurvival of the mice (Fig. 7C).

DiscussionFindings presented here demonstrate for the first time

that the BET protein antagonist (S)-JQ1, but not its in-active enantiomer (R)-JQ1, exerts a high level of in vitroand in vivo activity against AML BPCs expressing themutant NPM1cþ, with or without the coexpression of

FLT3-ITD. This was observed not only in the culturedOCI-AML3 cells with ectopic expression of FLT3-ITD,but also in the primary AML BPCs. The BET proteinBRD4 binds and recruits pTEFb to the promoters oftranscriptionally active genes to phosphorylate CTD ofRNA POL II, which is necessary to cause the pauserelease of RNA POL II for mRNA transcript elongation(7, 23). This is especially true for the MYC-regulatedtranscriptome, in which c-MYC binds and recruitspTEFb to the core promoters of the actively transcribedgenes, causing overall transcriptional amplificationwhich may attenuate the rate-limiting constraints ontumor growth and proliferation (7, 9, 23). Inhibition ofBRD4 by JQ1 has been shown to downregulate c-MYC–dependent target genes in AML and other hematologicmalignancies (14, 15). Recently, JQ1 has also been dem-onstrated to deplete the binding of BRD4, mediator, and

C

B

0 0.2 0.4 0.6 0.8 1.00

0.2

0.4

0.6

0.8

1.0

Fractional effect

CI

Primary CD34+CD38-LIN– AML

A

D

0

2

4

6

8

10

12

14

% N

on

via

ble

cells

Primary Normal

CD34+ cells (n = 2)

*

0

10

20

30

40

50

60

70

80

90

100

% N

on

via

ble

ce

lls

Primary CD34+ AML cells (n = 9)

PARP

b-Actin

Cl. PARP

500 nmol/L JQ1

20 nmol/L PS

+

–

–

–

–

+

+

+

c-MYC

BIMEL

Primary AML

BCL2

0 500 0 500 nmol/L JQ1, 24 h

+ 20 nmol/L PS

b-Actin

NPM1

pSer2 POL II

p21

JQ1 and PS

JQ1 + PS

JQ1 + PS

Control

Control

500 nmol/L JQ1

250 nmol/L JQ1

20 nmol/L PS

10 nmol/L PS

Figure 6. Cotreatment with JQ1 and panobinostat (PS) exerts synergistic lethal activity against primary AML BPCs but not normal hematopoieticprogenitor cells. A, primary CD34þ AML cells (n¼ 9) were treated with the indicated concentrations of JQ1 and/or panobinostat for 48 hours. Following this,cells were stained with propidium iodide and the percentages of nonviable cells were determined by flow cytometry. Columns, mean loss of viabilityof the samples; bars, SEM. �, values significantly greater in the combination than treatment with either agent alone (P < 0.01). Inset shows a representativeWestern blot analysis of cleaved PARP following treatment with the indicated concentrations of JQ1 and/or panobinostat in primary AML cells. B,primary AML cells were treated with the indicated concentrations of JQ1 and/or panobinostat for 24 hours. Then, cells were harvested and total cell lysateswere prepared. Immunoblot analyseswere conducted for the expression levels of p-Ser2 POL II, c-MYC, BCL2, BIM, p21, NPM1, and b-actin in the lysates. C,primary CD34þCD38-Lin� AML cells were treated with JQ1 and/or panobinostat for 48 hours and the percentages of nonviable cells were determinedby flow cytometry. Median dose effect and isobologram analysis were performed using CalcuSyn. CI values less than 1.0 indicate a synergistic interaction ofthe two agents in the combination. D, primary CD34þ normal cells (n ¼ 2) were treated with the indicated concentrations of JQ1 and/or panobinostat for 48hours. At the end of treatment, cells were stained with propidium iodide and the percentages of nonviable cells were determined by flow cytometry.






pTEFb to the enhancers and, in the case of some onco-genes, such as c-MYC and BCL2, to the clusters ofenhancers called the superenhancers, thereby depletingthe transcript levels of c-MYC and BCL2 (7, 14). Accord-ingly, in our studies, treatment with JQ1 clearly inhib-ited the binding of BRD4 and RNA POL II to the c-MYC,BCL2, and CDK6 promoters, accompanied with theattenuation of mRNA levels of these genes in the cul-tured and primary AML cells expressing NPM1cþ. Incontrast, as c-MYC is known to repress p21, JQ1 treat-ment increased p21 levels in the AML cells (25). Thus,JQ1 perturbed the levels of not only c-MYC but also ofthe c-MYC–targeted genes. We also found that shRNA toBRD4 reduced c-MYC and BCL2 mRNA levels, whileinducing the p21 levels, along with inhibition of growthand induction of apoptosis of OCI-AML3 cells. In the

gene expression microarray analysis in the OCI-AML3cells, more genes were suppressed than induced aftertreatment with JQ1. Again, among the genes transcrip-tionally attenuated was c-MYC.

Gain-of-functionmutations in FLT3 are associatedwitha poor prognosis following the standard induction ther-apy of AML (22, 26, 27). Although exhibiting a promisingpreliminary activity, the use of FLT3 tyrosine kinaseinhibitors (TKI) alone or in conjunction with standardinduction chemotherapy has exposed novel mechanismsby which AML BPCs acquire resistance to FLT3 kinaseinhibitors (28, 29). Because these are type II FLT3 TKIs thatbind and inactivate the "DFG-out" inactive conformationof FLT3, the clinically relevant resistance mechanismsidentified so far have involved point substitutions inamino acid residue that disrupt the conformation and

A

P = 0.0002

Median survival Days25Vehicle3450 mg/kg JQ139.55 mg/kg PS51.5JQ1 + PS

OCI-AML3 xenografts

Days after implantation

Perc

en

t su

rviv

al

0 10 20 30 40 50 60 70 800

50

100

150 Vehicle50 mg/kg JQ15 mg/kg PS50 mg/kg JQ1 + 5 mg/kg PS

JQ1: M-FPS: M,W,F3 weeks

B

C

Median survival Days18.5Vehicle24.550 mg/kg JQ126.55 mg/kg PS39.5JQ1 + PS

MOLM13 xenografts

JQ1: M-FPS: M,W,F3 weeks

P < 0.0001 P = 0.0133, JQ1 vs. vehicle

Days after implantation100806040200

0

50

Perc

en

t su

rviv

al

100 Vehicle50 mg/kg JQ15 mg/kg PSJQ1 + PS

P = 0.0005, PS vs. vehicle

c-MYC

OCI-AML3

xenograft

50 mg/kg JQ1

5.0 mg/kg PS

––

+–

–+

++

BCL2

CDK6

b-Actin

BIMEL

BIMLBIMS

Figure 7. Compared with treatment with either agent alone, combined treatment with JQ1 and panobinostat (PS) significantly improves the survival of NOD/SCIDmicebearingOCI-AML3orMOLM13xenografts. A, after apreconditioningdoseof gamma radiation (2.5Gy),micewere injectedwith5millionOCI-AML3cells in the lateral tail vein. Seven days after implantation, mice were treated with vehicle (10% 2-hydroxypropyl-b-cyclodextrin), JQ1 (50mg/kg formulated in10% 2-hydroxypropyl-b-cyclodextrin intraperitoneal injection daily � 5 days), panobinostat (5 mg/kg intraperitoneal injection, 3 times per week), or thecombination of JQ1 and panobinostat for 3 weeks, and then treatment was stopped. Survival of the mice is represented by the Kaplan–Meier plot. A,treatment with JQ1 significantly improved the survival of the mice compared with vehicle-treated mice (P ¼ 0.009). In addition, mice treated with thecombination of JQ1 and panobinostat exhibited significantly improved survival compared with themice treated with JQ1 or panobinostat alone (P¼ 0.0002).B, NOD/SCID mice were irradiated and injected with OCI-AML3 cells as in A and monitored for 2 weeks. Then, mice were treated with vehicle (10% 2-hydroxypropyl-b-cyclodextrin), JQ1 (50 mg/kg intraperitoneal injection daily � 5 days), panobinostat (5 mg/kg intraperitoneal injection 3 times per week), orthe combination of JQ1 and panobinostat for 1 week, and then the treatment was stopped. Mice were sacrificed and bone marrow was collectedfrom the femurs of mice in each treatment group. Total cell lysates were prepared and immunoblot analyses were conducted for the expression levels ofc-MYC, BCL2, CDK6, BIM, and b-actin in the lysates. Representative immunoblots are shown. C, after a preconditioning dose of gamma radiation(2.5 Gy), mice were injected with 5 million MOLM13 cells in the lateral tail vein. Four days after implantation, mice were treated with vehicle (10% 2-hydroxypropyl-b-cyclodextrin), JQ1 (50mg/kg formulated in 10%2-hydroxypropyl-b-cyclodextrin by IP injection daily� 5 days), panobinostat (5mg/kg by IPinjection 3 times perweek), or the combination of JQ1andpanobinostat (JQ1was administered on alternating dayswith panobinostat for this combination) for3 weeks, and then treatment was stopped. Survival of the mice is represented by Kaplan-Meier plot. Treatment with JQ1 significantly improved the survivalof the mice compared with vehicle treated mice (P ¼ 0.0133). In addition, mice treated with the combination of JQ1 and panobinostat had significantlyimproved survival compared with the mice treated with JQ1 or PS alone (P < 0.0001).

Fiskus et al.





binding of the TKI to FLT3. Our findings show that JQ1,due to its unique mechanism, which is disparate from themechanism of activity of the TKIs, retains potent activityagainst cultured and primary AML BPCs coexpressingFLT3-ITD and NPM1cþ or MLL fusion proteins. Thishighlights the promise of including a BET protein antag-onist in combinationwith standard chemotherapy and/ora FLT3 kinase inhibitor in the treatment of high-risk AMLexpressing FLT3-ITD.In addition, our findings also support the rationale

that, by inducing hyperacetylation of lysine residues onthe histone proteins, panobinostat could be inducinggreater dependency on the BRD4-regulated transcriptionof oncoproteins, such that cotreatment with panobinostatand JQ1 synergistically leads to growth inhibition andapoptosis of the cultured and primary AML cells regard-less of the coexpression of FLT3-ITD. A molecular sche-ma demonstrating the enhanced activity of BET proteinantagonist and panobinostat is shown in Fig. 8. Indeed,consistent with this, the synergistic activity of panobino-stat and JQ1 was associated with marked depletion ofthe protein levels of c-MYC, BCL2, and CDK6, as waspreviously reported (22). Findings presented here alsodemonstrate that this synergy extended in vitro tothe immunophenotypically defined leukemia stem cells(LSC) that have been previously demonstrated to exhibitin vivo leukemia-initiating potential (30). Targeting LSCsis an important goal in achieving deeper remissions andovercoming treatment refractoriness in minimal residualAML, as LSC gene expression program confers a poorprognosis and influences the clinical outcome in AML(30, 31). Cotreatment with panobinostat and JQ1 was also

significantly superior to treatment with each agent alonein improving the survival of the mice engrafted withOCI-AML3 or MOLM13 cells, which express NPM1cþ orMLL-AF9 and FLT3-ITD, respectively. The improvementin survival due to panobinostat and JQ1 treatment of themice engrafted with the aggressive MOLM13 AML cellswas associated with a significantly higher plateau in thesurvival curve, which suggests the possibility of a pro-longed disease-free survival and cure in the mice treatedwith the combination. The regimen of panobinostatand JQ1, as used here for 3 weeks, did not induce anydiscernible toxicity in the mice. The superior in vivo anti-AML selectivity was also associated with a marked andcollective depletion of c-MYC, BCL2, and CDK6 in theengrafted bone marrow–derived OCI-AML3 cells, fol-lowing only 5 days of treatment with the combination.Attenuation of these proteins could very well be theexplanation for the superior antileukemia activity in thetreated mice. However, other possible mechanisms, notprobed here, may also be contributing toward the super-ior outcome. BRD4 has also been shown to bind to theacetylated lysine-310 of the RelA subunit of NF-kB andregulate its transcriptional activity (32, 33). By inducingRelA acetylation, panobinostat treatment may alsoincrease the BRD4 dependency of the NF-kB activity inAML cells. Therefore, suppression of this activity bycotreatment with panobinostat and JQ1 may contributetoward the superior anti-AML activity of the combina-tion. Collectively, the findings presented here provide astrong rationale for further in vivo testing of combinedtherapy with BET protein antagonists and HDAC inhib-itor against AML.

pTEFb

BRD4

K5 acK8 acK12 acK16 ac

H4

K9 acK14 ac

H3

Mediator RNAPol II

P

PP

BETprotein

antagonist

Low-dosePS

Transcription factor

complex (e.g. MYC)

HATc-MYC

BCL2BCL-xL

Cell growth

pTEFb

BRD4

K5 acK8 acK12 acK16 ac

H4

K9 acK14 ac

H3

MediatorRNAPol II

PIM1

Apoptosis+

Figure 8. Molecular schema demonstrating the superior activity of BET protein antagonist and pan-HDAC inhibitor against AML cells. A, in AML cells, histoneacetyltransferases (HAT) are bound to transcription factor complexes (TFC; e.g., c-MYC) at the promoters of target genes, which acetylates the lysine (K)residues on histone H3 and H4 and creates a permissive environment for transcription. Treatment with low doses of pan-HDAC inhibitors, such aspanobinostat (PS), inhibits HDAC activity which results in hyperacetylation of N-terminal tails on histone H3 and H4 and greater dependency on BET proteinfunction. BRD4 binds to acetylated histones and recruits pTEFb to the enhancer/promoter of target genes such as c-MYC, BCL2, and CDK6. Upon itsrecruitment by BRD4, pTEFb phosphorylates the CTD of RNA POL II resulting in transcription of target genes. Treatment with BET protein inhibitors JQ1or I-BET151 inhibits BRD4 activity causing depletion of RNA POL II phosphorylation at Ser 2, reduced binding of BRD4 and RNA POL II at target genepromoters, and decreased transcription of target genes. In addition, treatment with JQ1 inhibits cell growth and induces apoptosis of AML cells. Combinedtreatment with BET protein inhibitor and panobinostat results in greater depletion of target genes (e.g., c-MYC, BCL2, Bcl-xL, and PIM1). This leadsto a greater loss of cell growth and further induction of apoptosis than treatment with either agent alone in the AML cells.






Disclosure of Potential Conflicts of InterestS. Sharma has ownership interest (including patents) in Convergene,

Beta Cat Pharmaceuticals, and Salarius Pharma. J.E. Bradner has owner-ship interest (including patents) and is a consultant/advisory boardmember in Tensha Therapeutics. No potential conflicts of interest weredisclosed by the other authors.

Authors' ContributionsConception and design: K.N. BhallaDevelopment of methodology: W. Fiskus, S. SharmaAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): K. Peth, M. Rodriguez, S.P. IyerAnalysis and interpretation of data (e.g., statistical analysis, biostatis-tics, computational analysis): W. Fiskus, S. Sharma, J.A. Valenta, M.Rodriguez, M. Zhan, J. Sheng, S.P. Iyer, J.E. Bradner, K.N. Bhalla

Writing, review, and/or revision of the manuscript: W. Fiskus, B.P.Portier, S.P. Iyer, J.E. Bradner, K.N. BhallaAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): J. Qi, L.J. Schaub, B. Shah, S.G.T.Devaraj, J.E. Bradner

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received September 12, 2013; revised November 24, 2013; acceptedNovember 25, 2013; published OnlineFirst January 16, 2014.

References1. Bhalla KN. Epigenetic and chromatin modifiers as targeted therapy of

hematologic malignancies. J Clin Oncol 2005;23:3971–93.2. Shen H, Laird PW. Interplay between the cancer genome and epigen-

ome. Cell 2013;153:38–55.3. Arrowsmith CH, Bountra C, Fish PV, Lee K, Schapira M. Epigenetic

protein families: a new frontier for drug discovery. Nat Rev Drug Disc2012;11:384–400.

4. Wu SY, Chiang CM. The double bromodomain-containing chromatinadaptor Brd4 and transcriptional regulation. J Biol Chem 2007;282:13141–5.

5. Belkina AC, Denis GV. BET domain co-regulators in obesity, inflam-mation and cancer. Nat Rev Cancer 2012;12:465–77.

6. Patel MC, DebrosseM, Smith M, Dey A, HuynhW, Sarai N, et al. BRD4coordinates recruitment of pause release factor P-TEFb and thepausing complex NELF/DSIF to regulate transcription elongation ofinterferon-stimulated genes. Mol Cell Biol 2013;33:2497–507.

7. Loven J, Hoke HA, Lin CY, Lau A, Orlando DA, Vakoc CR, et al.Selective inhibition of tumor oncogenes by disruption of super-enhan-cers. Cell 2013;153:320–34.

8. Lockwood WW, Zejnullahu K, Bradner JE, Varmus H. Sensitivity ofhuman lung adenocarcinoma cell lines to targeted inhibition of BETepigenetic signaling proteins. Proc Natl Acad Sci U S A 2012;109:19408–13.

9. Lee TI, Young RA. Transcriptional regulation and its misregulation indisease. Cell 2013;152:1237–51.

10. Zuber J, Shi J, Wang E, Rappaport AR, Herrmann H, Sison EA, et al.RNAi screen identifies Brd4 as a therapeutic target in acute myeloidleukaemia. Nature 2011;478:524–8.

11. Filippakopoulos P, Qi J, Picaud S, Shen Y, SmithWB, FedorovO, et al.Selective inhibition of BET bromodomains. Nature 2010;468:1067–73.

12. Mirguet O, Lamotte Y, Donche F, Toum J, Gellibert F, Bouillot A, et al.From ApoA1 upregulation to BET family bromodomain inhibition:discovery of I-BET151. Bioorg Med Chem Lett 2012;22:2963–7.

13. Seal J, Lamotte Y, Donche F, Bouillot A, Mirguet O, Gellibert F, et al.Identification of a novel series of BET family bromodomain inhibitors:binding mode and profile of I-BET151 (GSK1210151A). Bioorg MedChem Lett 2012;22:2968–72.

14. Delmore JE, Issa GC, Lemieux ME, Rahl PB, Shi J, Jacobs HM, et al.BET bromodomain inhibition as a therapeutic strategy to target c-Myc.Cell 2011;146:904–17.

15. Mertz JA, Conery AR, Bryant BM, Sandy P, Balasubramanian S, MeleDA, et al. Targeting MYC dependence in cancer by inhibiting BETbromodomains. Proc Natl Acad Sci U S A 2011;108:16669–74.

16. Dawson MA, Prinjha RK, Dittmann A, Giotopoulos G, Bantscheff M,Chan WI, et al. Inhibition of BET recruitment to chromatin as aneffective treatment for MLL-fusion leukaemia. Nature 2011;478:529–33.

17. George P, Bali P, Annavarapu S, Scuto A, Fiskus W, Guo F, et al.Combination of the histone deacetylase inhibitor LBH589 and thehsp90 inhibitor 17-AAG is highly active against human CML-BC cells

and AML cells with activating mutation of FLT-3. Blood 2005;105:1768–76.

18. Fiskus W, Verstovsek S, Manshouri T, Smith JE, Peth K, Abhyankar S,et al. Dual PI3K/AKT/mTOR inhibitor BEZ235 synergistically enhancesthe activity of JAK2 inhibitor against cultured and primary humanmyeloproliferative neoplasm cells. Mol Cancer Ther 2013;12:577–88.

19. Fiskus W, Wang Y, Sreekumar A, Buckley KM, Shi H, Jillella A, et al.Combined epigenetic therapy with the histone methyltransferaseEZH2 inhibitor 3-deazaneplanocin A and the histone deacetylaseinhibitor panobinostat against human AML cells. Blood 2009;114:2733–43.

20. Balusu R, Fiskus W, Rao R, Chong DG, Nalluri S, Mudunuru U, et al.Targeting levels or oligomerization of nucleophosmin 1 induces dif-ferentiation and loss of survival of humanAMLcellswithmutantNPM1.Blood 2011;118:3096–106.

21. Chou TC, Talalay P. Quantitative analysis of dose-effect relationships:the combined effects of multiple drugs or enzyme inhibitors. AdvEnzyme Regul 1984;22:27–55.

22. Schnittger S, Bacher U, Kern W, Alpermann T, Haferlach C, HaferlachT. Prognostic impact of FLT3-ITD load inNPM1mutatedacutemyeloidleukemia. Leukemia 2011;25:1297–304.

23. Lin CY, Loven J, Rahl PB, Paranal RM, Burge CB, Bradner JE, et al.Transcriptional amplification in tumor cells with elevated c-Myc. Cell2012;151:56–67.

24. Deangelo DJ, Spencer A, Bhalla KN, Prince HM, Fischer T, Kindler T,et al. Phase Ia/II, two-arm, open-label, dose-escalation study of oralpanobinostat administered via two dosing schedules in patients withadvanced hematologic malignancies. Leukemia 2013;27:1628–36

25. Dang CV. MYC on the path to cancer. Cell 2012;149:22–35.26. Kindler T, Lipka DB, Fischer T. FLT3 as a therapeutic target in AML: still

challenging after all these years. Blood 2010;116:5089–102.27. Gale RE, Green C, Allen C, Mead AJ, Burnett AK, Hills RK, et al. The

impact of FLT3 internal tandem duplicationmutant level, number, size,and interaction with NPM1 mutations in a large cohort of young adultpatients with acute myeloid leukemia. Blood 2008;111:2776–84.

28. Levis M. FLT3/ITD AML and the law of unintended consequences.Blood 2011;117:6987–90.

29. Smith CC, Wang Q, Chin CS, Salerno S, Damon LE, Levis MJ, et al.Validation of ITD mutations in FLT3 as a therapeutic target in humanacute myeloid leukaemia. Nature 2012;485:260–3.

30. Eppert K, Takenaka K, Lechman ER, Waldron L, Nilsson B, van GalenP, et al. Stemcell geneexpressionprograms influenceclinical outcomein human leukemia. Nat Med 2011;17:1086–93.

31. BussEC,HoAD. Leukemia stemcells. Int JCancer 2011;129:2328–36.32. Zou Z, Huang B, Wu X, Zhang H, Qi J, Bradner J, et al. Brd4 maintains

constitutively active NF-kappaB in cancer cells by binding to acety-lated RelA. Oncogene. Epub 2013 May 20.

33. Zhang G, Liu R, Zhong Y, Plotnikov AN, ZhangW, Zeng L, et al. Down-regulation of NF-kappaB transcriptional activity in HIV-associatedkidney disease by BRD4 inhibition. J Biol Chem 2012;287:28840–51.


Fiskus et al.




2014;13:1142-1154. Published OnlineFirst January 16, 2014.Mol Cancer Ther Warren Fiskus, Sunil Sharma, Jun Qi, et al. Cells

LeukemiaDeacetylase Inhibitor against Human Acute Myelogenous Highly Active Combination of BRD4 Antagonist and Histone

Updated version

10.1158/1535-7163.MCT-13-0770doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2014/01/16/1535-7163.MCT-13-0770.DC1

Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/13/5/1142.full#ref-list-1

This article cites 32 articles, 13 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/13/5/1142.full#related-urls

This article has been cited by 21 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/13/5/1142To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-13-0770

http://mct.aacrjournals.org/content/suppl/2014/01/16/1535-7163.MCT-13-0770.DC1

http://mct.aacrjournals.org/content/13/5/1142.full#ref-list-1

http://mct.aacrjournals.org/content/13/5/1142.full#related-urls

http://mct.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mct.aacrjournals.org/content/13/5/1142


Documents

Highly Active Combination of BRD4 Antagonist and Histone Deacetylase Inhibitor … · Small Molecule Therapeutics See related article by Liu et al., p. 1194 Highly Active Combination