Full Text (PDF)

Published OnlineFirst August 29, 2012; DOI: 10.1158/1078-0432.CCR-12-0873

Clinical Cancer Cancer Therapy: Preclinical Research

Superior Efﬁcacy of a Combined Epigenetic Therapy against Human Mantle Cell Lymphoma Cells

Warren Fiskus1, Rekha Rao1, Ramesh Balusu1, Siddhartha Ganguly1, Jianguo Tao2, Eduardo Sotomayor2, Uma Mudunuru1, Jacqueline E. Smith1, Stacey L. Hembruff1, Peter Atadja3, Victor E. Marquez4, and Kapil Bhalla1

Abstract Purpose: A deregulated epigenome contributes to the transformed phenotype of mantle cell lymphoma (MCL). This involves activity of the polycomb repressive complex (PRC) 2, containing three core proteins, EZH2, SUZ12, and EED, in which the SET domain of EZH2 mediates the histone methyltransferase activity. We determined the effects of 3-deazaneplanocin A (DZNep), an S-adeno- sylhomocysteine hydrolase inhibitor, and/or pan-histone deacetylase inhibitor panobinostat (PS) on cultured and primary MCL cells. Experimental Design: Following treatment with DZNep and/or PS, apoptosis and the levels and activity of EZH2 and PRC2 proteins in cultured and primary MCL cells were determined. Results: Treatment with DZNep depleted EZH2, SUZ12, and 3MeK27H3 in the cultured human MCL cells. DZNep also increased expression of p21, p27, and FBXO32, whereas it depleted Cyclin D1 and Cyclin E1 levels in MCL cells. In addition, DZNep treatment induced cell-cycle arrest and apoptosis in cultured and primary MCL cells. Furthermore, as compared with treatment with each agent alone, cotreatment with DZNep and PS caused greater depletion of EZH2, SUZ12, 3MeK27H3, and Cyclin D1 levels, whereas it induced greater expression of FBXO32, p16, p21, and p27. Combined treatment with DZNep and PS synergistically induced apoptosis of cultured and primary MCL cells while relatively sparing normal CD34 þ cells. Cotreatment with DZNep and PS also caused signiﬁcantly greater inhibition of tumor growth of JeKo-1 xenografts in NOD/SCID mice. Conclusions: These preclinical in vitro and in vivo ﬁndings show that cotreatment with DZNep and PS is an active combined epigenetic therapy worthy of further in vivo testing against MCL. Clin Cancer Res; 18(22); 6227–38. 2012 AACR.

Introduction the pathogenesis of MCL (3–5). In addition, increased Mantle cell lymphoma (MCL) is an aggressive and dis- expression of MYC, Cyclin D1, and EZH2 has been associ- tinct type of B-cell malignancy that comprises up to 10% of ated with poor prognosis and correlate with poorer clinical non-Hodgkin lymphomas (1). A characteristic feature of outcome in MCL (3, 6–8). EZH2 is a core member and the MCL is the chromosomal rearrangement caused by the catalytic subunit of the polycomb protein complex, poly- translocation t(11;14)(q13;q32), which results in overex- comb repressive complex (PRC) 2, that also contains SUZ12 pression of Cyclin D1 (2, 3). Copy number variations and and EED, in which the SET domain of EZH2 mediates the associated gene expression changes involving CDKN2A histone lysine (K) methyltransferase (KMTase) activity of (p16) and CDKN2B (p15), as well as altered expression of PRC2 (9). EZH2 is a KMTase that mediates epigenetic CDK4, and p27 has also been noted and may play a role in silencing of PRC2 target genes by inducing trimethylation (3Me) of K27 on histone H3 (3MeK27H3) in the chromatin (9, 10). EZH2 overexpression in transformed cells promotes Authors' Afﬁliations: 1University of Kansas Cancer Center, Kansas City, cell proliferation and aggressiveness (11–13). In hemato- Kansas; 2H. Lee Mofﬁtt Cancer Center, Tampa, Florida; 3Novartis Institute HOX for Biomedical Research, Inc., Cambridge, Massachusetts; and 4National logic malignancies, PRC2 regulates the expression of Institutes of Health, Bethesda, Maryland genes and epigenetically represses genes including tumor Note: Supplementary data for this article are available at Clinical Cancer suppressor genes p16 (CDKN2A) and p14 (ARF; ref. 14). Research Online (http://clincancerres.aacrjournals.org/). EZH2 was shown to be preferentially overexpressed in Corresponding Author: Kapil Bhalla, University of Kansas Cancer Center, proliferating but not resting MCL cells (15). Recently, in 3901 Rainbow Blvd., 4030 Robinson, MS 1027, Kansas City, KS 66160. lymphomagenesis, EZH2-mediated gene silencing in ger- Phone: 913-945-7086; Fax: 913-588-4701; E-mail: [email protected] minal center B cells was shown to contribute to cell prolif- doi: 10.1158/1078-0432.CCR-12-0873 eration (16). Somatic gain of function mutations in tyrosine 2012 American Association for Cancer Research. 641 (Y641C) in EZH2 have also been noted to selectively

www.aacrjournals.org 6227

Fiskus et al.

mined in MCL cells. In contrast, treatment with hydroxa- Translational Relevance mate pan-histone deacetylase (HDAC) inhibitors, includ- Overexpression of EZH2 and the polycomb repressive ing vorinostat and panobinostat (PS, LBH589, Novartis complex (PRC) 2 proteins has been associated with poor Pharmaceutical Inc.), has been shown to be active against prognosis and correlates with poorer clinical outcome in MCL cells (29–31). In acute myelogenous leukemia (AML) mantle cell lymphoma (MCL). PRC2 mediates trimethy- cells, treatment with PS was shown to deplete the levels of lation of lysine 27 on histone H3 (3MeK27H3) and EZH2 and SUZ12, as well as reduce the 3MeK27H3 mark on epigenetically represses genes such as the tumor sup- the chromatin (32). PS treatment also depleted the levels of pressor p16. The effects of pharmacologic inhibition of DNMT1 and disrupted its binding to EZH2 in AML cells (33, EZH2 or EZH2 knockdown have not been tested in MCL 34). In addition, cotreatment with DZNep and PS was cells. Here, we show that treatment with the EZH2 shown to be more effective against the epigenetic targets antagonist DZNep or the pan-histone deacetylase inhib- and synergistically induced apoptosis of AML cells (34). In itor panobinostat (PS) depletes EZH2, SUZ12, and the present studies, we determined that treatment with 3MeK27H3. Cotreatment with PS-augmented DZNep DZNep depletes PRC2 complex proteins and induces mediated attenuation of PRC2, which was associated cell-cycle arrest and apoptosis of cultured and primary MCL with synergistic activity of the combination against cells. Our ﬁndings also show that the combined treatment human MCL cells in vitro and in vivo. These observations of DZNep and PS causes more depletion of PRC2 complex support the rationale to further test the in vivo anti-MCL proteins and synergistically induces apoptosis of cultured efﬁcacy of the combination of PS with an EZH2 antag- and primary MCL cells but not normal CD34 þ bone onist or combination therapies that target the deregu- marrow progenitor cells. In addition, cotreatment with lated epigenome in human MCL cells. DZNep and PS was more effective than each agent alone in inhibiting the in vivo tumor growth of JeKo-1 xenografts in NOD/SCID mice.

alter PRC2 catalytic activity and drive hypertrimethylation Materials and Methods of H3K27 in germinal center B cell lymphoma, thereby Reagents identifying EZH2 as a potentially attractive therapeutic tar- PS was kindly provided by Novartis Pharmaceuticals, Inc. get (17–20). PRC2-mediated trimethylation of H3K27 also DZNep was acquired from the National Cancer Institute recruits the multiprotein PRC1 complex, consisting among (Rockville, MD). Monoclonal BMI1, polyclonal trimethy- others of B lymphoma Mo-MLV insertion region 1 homolog lated K9 histone H3, polyclonal trimethylated K79 histone (BMI1), as well as RING1 and RING2 proteins, which H3, acetylated K16 histone H4, acetyl K56 histone H3, and mediate ubiquitylation of histone H2A on K119 associated polyclonal trimethylated K4 histone H3 antibodies were with gene repression (9, 14). BMI1 represses CDKN2A and purchased from Millipore. All other validated antibodies INK4A/ARF and cooperates with MYC in lymphomagenesis were purchased from vendors, as previously described (29, (14). In addition, the BMI1 locus has been shown to be 30, 33, 34). ampliﬁed in MCL cells (21). EZH2 has also been reported to directly control DNA methylation through its association Cell lines and cell culture with and regulation of the activity of the DNA methyltrans- MCL cell line MO2058 was obtained and maintained as ferases DNMT1, DNMT3a, and DNMT3b (22, 23). Howev- previously described (29, 30). JeKo-1 and Z-138 cells were er, genes silenced in cancer by 3MeK27H3 have been shown obtained from American Type Culture Collection (ATCC, to be independent of promoter DNA methylation indicating Manassas, VA). Both cell lines were banked after receipt, and that PRC2-mediated silencing could be an independent passaged for less than 6 months before use in these studies. mechanism for suppression of tumor suppressor genes The ATCC characterizes cell lines using short tandem repeat (24). Consistent with this, both DNA methylation and polymorphism analysis. MCL cells were maintained in transcriptional silencing of PRC2 target genes persists when culture, as previously described (29, 30). Logarithmically EZH2 expression is depleted (25, 26). growing cells were exposed to the designated concentrations 3-deazaneplanocin A (DZNep) is the cyclopentanyl ana- of DZNep and/or PS. Following these treatments, cells were log of 3-deazaadenosine that inhibits the activity of S- washed free of the drug(s) before the performance of the adenosyl-L-homocysteine (AdoHcy) hydrolase, the enzyme studies. responsible for the reversible hydrolysis of AdoHcy to adenosine and homocysteine (27). DZNep has also been Primary MCL cells shown to deplete the expression levels of EZH2 and SUZ12, Primary MCL samples and normal CD34 þ mononuclear with concomitant loss of trimethylation of K27 on histone cells were obtained with informed consent as part of a H3 and reexpression of epigenetically silenced targets such clinical protocol approved by the Institutional Review as the F-box protein FBXO32—a component of the SCF Board of University of Kansas (IRB approval # 12392). ubiquitin protein E3 ligase complex, associated with apo- Peripheral blood or bone marrow aspirate samples were ptosis of cancer cells (28). However, the effects of depletion collected and separated for mononuclear cells, as previously of EZH2 and the activity of DZNep had not been deter- described (29, 30, 33, 34).

6228 Clin Cancer Res; 18(22) November 15, 2012 Clinical Cancer Research

Combined Epigenetic Therapy against Human MCL Cells

Cell-cycle analysis in ice-cold 1X PBS, and the cell pellet was snap frozen in Following DZNep treatment, MCL cells were washed liquid nitrogen. Cell lysis, sonication, and chromatin twice with 1X PBS and fixed in 70% ethanol at 20C immunoprecipitation for EZH2 was done according to overnight. Fixed cells were washed twice with 1X PBS, the manufacturer’s protocol (Millipore). For quantitative resuspended in 1X PBS with 0.1% Triton X-100, RNAse assessment of HOXA9, RUNX3, and WNT1 promoters in A, and propidium iodide and incubated in the dark at 37C the chromatin immunoprecipitates, a SYBR Green Mas- for 15 minutes. Cell-cycle data were collected on a flow termix from Applied Biosystems was used. Relative cytometer with a 488 nmol/L laser and analyzed with enrichment in the chromatin immunoprecipitates was ModFit 3.0, as previously described (34). normalized against HOXA9, RUNX3, and WNT1 promoter DNA in the input samples, as previously described Assessment of percentage of nonviable cells (28, 34, 37). Following treatment with DZNep and/or PS, nonviable cells were determined by trypan blue dye uptake on a Assessment of apoptosis of MCL cells hemocytometer, as previously described (29, 30, 34). Alter- Untreated or drug-treated cells were stained with natively, cells were stained with propidium iodide and the annexin-V-FITC (Pharmingen) and TOPRO3 iodide, and percentages of nonviable cells were determined by flow the percentages of annexin-V-positive apoptotic cells cytometry. were determined by flow cytometry (29, 30). To analyze synergism between DZNep and PS in inducing apopto- RNA interference sis, cells were treated with DZNep (250—2,000 nmol/L) Small interfering RNAs against EZH2 were obtained and PS (5–100 nmol/L) at a constant ratio for 48 hours. from Dharmacon. JeKo-1 cells were transfected with The percentages of apoptotic cells were determined by siRNA to a final concentration of 100 nmol/L using flow cytometry, as previously described (34). The com- an Amaxa Nucleofector device. Cells were incubated for bination index (CI) for each drug combination was 48–72 hours for detection of EZH2 knockdown. Lenti- obtained by median dose effect of Chou and Talalay viral shRNAs against EZH2 were obtained from Sigma- (35) utilizing the combination index equation (assum- Aldrich. JeKo-1 cells were transduced with lentivirus ing mutual exclusivity) within the commercially avail- and incubated for 48 hours. Following this, cells were able software Calcusyn (Biosoft). CI values of less than washed with complete media and incubated an addition- 1.0 represent synergistic interaction of the 2 drugs in the al 24 hours for immunoblot analyses for 72 hours to combination. observe effects on cell proliferation, as previously described (34, 36). Cell lysis, histone isolation, and protein quantitation Untreated or drug-treated cells were centrifuged and RNA isolation and reverse transcription-PCR the cell pellets were resuspended in 200 mL of lysis buffer RNA was extracted from the cultured MCL cells using an as previously described (33, 34). Histones were extracted RNaqueous-4PCR kit (Applied Biosystems). Purified total from untreated and treated cells as previously described RNA was quantified, reverse transcribed, and quantitative (32). real-time PCR analyses for EZH2, SUZ12, EED, Cyclin D2, SMARCA2, TCF4 and EIF3A were done on the resulting SDS-PAGE and immunoblot analyses cDNA utilizing TaqMan probes from Applied Biosystems, as Seventy-five micrograms of total cell lysate was used for previously described (33, 34). SDS-PAGE. Western blot analyses of DNMT1, EZH2, SUZ12, EED, 3MeK27H3, acetyl K27H3, PARP, FBXO32, Detection and analysis of hsa-miR-101 in MCL cells Cyclin E, p16, p21, and p27 were conducted on total cell For detection of hsa-miR-101 in JeKo-1 and MO2058 lysates as previously described (29, 30, 34). Immunoblot cells, microRNAs were isolated with a kit from Applied analyses were conducted 3 times, and the same immuno- Biosystems. Enriched RNA was reverse transcribed with a blot was probed with specific antisera or monoclonal anti- stem loop primer included in the TaqMan hsa-miR-101 bodies. Representative blots were subjected to densitomet- microRNA assay following the manufacturer’s protocol ric analysis where expression levels of b-actin were used as (Applied Biosystems). Expression of hsa-miR-101 was the control. Densitometry was conducted using Image- detected by qPCR with a TaqMan probe specific to hsa- Quant 5.2 (GE Healthcare). miR-101. Relative expression of hsa-miR-101 was normalized against expression of 18S RNA. Mantle cell lymphoma xenograft JeKo-1 cells (10 million) were subcutaneously implant- Chromatin immunoprecipitation and PCR ed into the flank of NOD/SCID mice. When the average JeKo-1 cells were treated with DZNep for 16 hours. The tumor volume reached approximately 150 mm3, the fol- chromatin in the cells was cross-linked with formalde- lowing treatments were administered in cohorts of 5 mice hyde at 37C for 10 minutes then quenched with 1/20 for each treatment: vehicle alone (5% DMSO), 1 mg/kg volume of 2.5 mol/L glycine for 5 minutes at room DZNep, 10 mg/kg PS, and DZNep plus PS. DZNep temperature. The cells were washed twice for 5 minutes was administered twice per week intraperitoneally for 2

www.aacrjournals.org Clin Cancer Res; 18(22) November 15, 2012 6229

Fiskus et al.

weeks. PS was administered 3 times per week for 2 weeks. Results Tumor growth was monitored every other day by calipers Treatment with DZNep depletes PRC2 proteins EZH2 and tumor volume was calculated using the equation: and SUZ12 levels and disrupts the PRC2 complex in 1 2 [ /2 (length width )]. We selected the dose of DZNep MCL cells that had been determined to be safe in previously reported Treatment with DZNep has been previously shown to studies and combined it with a dose of PS that we had deplete EZH2 and SUZ12 levels in breast cancer, colon previously reported to be safe and biologically effective cancer and leukemia cells (28, 34, 37). Here, we deter- (29, 38). This study was conducted in duplicate with similar mined the effect of DZNep on PRC2 proteins in the results. cultured MCL JeKo-1, MO2058, and Z-138 and primary MCL cells. Exposure to DZNep for 24 hours dose depen- Statistical analysis dently reduced the protein expression of EZH2, SUZ12, Significant differences between values obtained in a and BMI1 in cultured and primary MCL cells (Fig. 1A and population of MCL cells treated with different experi- B and Supplementary Fig. S1A). Similar to our previous mental conditions were determined using the Student’s findings in AML cells, exposure to DZNep up to 1.0 mmol/ t test. For the in vivo mouse models, the final tumor L, did not significantly affect the mRNA expression of volumes at the end of treatment were used to calculate EZH2 and SUZ12, as determined by quantitative RT-PCR the difference between mice treated with DZNep or PS (Fig. 1C). As shown in Supplementary Fig. 1B, cotreat- alone versus mice treated with the combination using ment with the proteasome inhibitor bortezomib restored a 2-tailed t test. P values of less than 0.05 were assigned the DZNep-mediated depletion of EZH2 and SUZ12 significance. levels, indicating that depletion of the PRC2 proteins by

A B C JeKo-1 1.5 EZH2 MO2058 0 0.5 1.0 µmol/L, DZNep Primary MCL JeKo-1 EZH2 0 0.2 1.0 µmol/L, DZNep 1 1.0 0.80 0.56 EZH2 RQ SUZ12 0.5 1.0 0.74 0.64 SUZ12 EED EED 0 1.5 BMI1 BMI-1 SUZ12 1.0 0.73 0.50 β-Actin β-Actin 1 RQ 0.5

0 Control 0.5 1 2 D JeKo-1 E µmol/L, DZNep, 16 h 0 0.5 1.0 2.0 µmol/L, DZNep, 24 h

3MeK27 H3 JeKo-1 1.0 0.61 0.59 0.49 0 0.5 1.0 2.0 µmol/L, DZNep, 4 h IgG 3MeK4 H3 Input 1.0 1.44 1.38 1.20 DNMT1 3MeK9 H3 1.0 0.65 0.60 0.60 SUZ12 3MeK79 H3 IP: EZH2 1.0 0.91 0.82 0.58 IB: Ac K16 H4 EED 1.0 0.87 043 0.43 Ac K27 H3 EZH2 Ac K56 H3

Histone H3

Figure 1. Treatment with DZNep depletes expression of PRC2 proteins EZH2, SUZ12, and associated 3MeK27H3 levels in cultured human MCL cells. A, immunoblot analyses of EZH2, SUZ12, EED, BMI1, and b-actin in JeKo-1 cells following 24 hours treatment with DZNep. B, immunoblot analyses of EZH2, SUZ12, EED, BMI1, and b-actin in primary MCL cells after treatment with DZNep for 24 hours. C, quantitative RT-PCR for EZH2 and SUZ12 in MO2058 and JeKo1 treated with DZNep for 16 hours. The relative quantity (RQ) of each mRNA was normalized against glyceraldehyde-3-phosphate dehydrogenase expression. D, immunoblot analyses of 3Me K27H3, 3MeK4H3, 3MeK79H3, 3MeK9H3, AcK16H4, AcK27H3, AcK56H3, and histone H3 on acid-extracted histones from JeKo-1 cells treated with DZNep for 24 hours. E, JeKo-1 cells were treated with DZNep for 4 hours. Following this, EZH2 was immunoprecipitated and immunoblot analyses were conducted for DNMT1, SUZ12, EED, and EZH2 on the immunoprecipitates. The numbers beneath the blots represent densitometric evaluation of band intensity of the proteins bound to EZH2 in the immunoprecipitates.

6230 Clin Cancer Res; 18(22) November 15, 2012 Clinical Cancer Research

Combined Epigenetic Therapy against Human MCL Cells

A B JeKo-1 0 0.5 1.0 µmol/L, DZNep JeKo-1 FBXO32 1.0 2.47 3.07 1.4 1.2 1.2 HOXA9 WNT1 RUNX3 p27 1.2 1 1 1.0 1.10 1.30 1 0.8 0.8 p21 0.8 1.0 1.63 2.01 0.6 0.6 0.6 p16 0.4 0.4 1.0 1.25 1.25 0.4 Fold enrichment

(relative to (relative control) 0.2 0.2 Cyclin D1 0.2 1.0 0.67 0.54 0 0 0 Cyclin E 0 1.0 2.0 0 1.0 2.0 0 1.0 2.0 µmol/L, DZNep, 16 h µmol/L, DZNep, 16 h µmol/L, DZNep, 16 h β-Actin

Figure 2. Treatment with DZNep attenuates EZH2 binding at target promoters, induces the cell-cycle inhibitors p16, p21, and p27 and depletes Cyclin D1 and Cyclin E expression in MCL cells. A, quantitative PCR of the HOXA9, WNT1, and RUNX3 promoters in EZH2 chromatin immunoprecipitates from JeKo-1 cells following treatment with DZNep. Fold enrichment data are relative to the control and are expressed as a ratio of the CT of the chIP DNA versus the CT for the input samples. B, immunoblot analyses of FBXO32, p27, p21, p16, Cyclin D1, Cyclin E, and b-actin in JeKo-1 cells after treatment with DZNep for 24 hours.

DZNep was mediated by proteasomal degradation. In caused significant promoter hypomethylation (Supple- MCL cells, depletion of the PRC2 complex proteins by mentary Fig. S3C). DZNep was associated with attenuation of the repressive histone mark 3MeK27H3 and upregulation of 3MeK4H3 DZNep treatment induces cell-cycle arrest and (Fig. 1D and data not shown). DZNep treatment did not apoptosis of MCL cells alter the expression levels of 3MeK9H3 and 3MeK79H3, We next determined the effects of DZNep treatment on but induced acetylation of H3K27, H3K56, and H4K16 the cell cycle of MO2058 and JeKo-1 cells. As shown in Table (Fig. 1D). Treatment with DZNep for as short as 4 hours 1, treatment of MO2058 and JeKo-1 cells resulted in a dose- markedly depleted the binding of EZH2 with other PRC2 dependent accumulation of cells in the G0–G1 phase, with a proteins, SUZ12 and EED, as well as diminished binding concomitant decline in the number of cells in S-phase of the of EZH2 to DNMT1 in MCL cells (Fig. 1E). DZNep- cell cycle (P < 0.05). Next, we determined the effects of mediated alterations in the levels of PRC2 proteins and DZNep treatment on induction of apoptosis in MO2058, in the chromatin marks led to decline in the binding of JeKo-1, and Z-138 cells. Treatment with DZNep dose depen- EZH2 to its known target promoters, for example, dently induced apoptosis in all 3 MCL cell lines, although Z- HOXA9, WNT1, and RUNX3, as discerned by chromatin 138 cells were more sensitive to the lower concentrations of immunoprecipitation analysisofthepromotersofthese DZNep compared with MO2058 and JeKo-1 cells (Fig. 3A genes (Fig. 2A; refs. 28, 34, 37). Both in cultured and and Supplementary Fig. S4A). The apoptotic effects in primary MCL cells, treatment with DZNep also reduced MO2058 and JeKo-1 cells were further enhanced following the levels of the PRC1 protein BMI1 (Fig. 1A and B). The a 72-hour exposure to DZNep (data not shown). DZNep- effects of DZNep on PRC2 and PRC1 proteins and chro- induced caspase 3 activity was also shown by induction of matin marks were associated with decline in the levels of PARP cleavage, a hallmark of apoptosis (Fig. 3B and Sup- Cyclin D1 and Cyclin E, but increase in the levels of p27, plementary Fig. S4B). p21, p16, and the prodeath F-Box protein FBXO32, an E3 ubiquitin ligase (28, 34), in the cultured MCL cells (Fig. DZNep-mediated induction of p16, p21, p27, and 2B and Supplementary Fig. S1A). DZNep treatment sig- FBXO32 is mechanistically linked to depletion of EZH2 nificantly increased the mRNA levels of the PRC2 target in MCL cells genes SMARCA2, EIF3A,andTCF4 (Supplemental Fig. To determine whether the DZNep-mediated decline in S2A-C). Recently, the expression of EZH2 was shown to EZH2 was mechanistically linked to increase in the levels of be inhibited by microRNA-101 in cancer cells (39, 40). p21, p27, and FBXO32, we also determined the effect of Although MO2058 and JeKo-1 cells possess the pre- siRNA to EZH2 in JeKo-1 cells. Figure 4A shows that, as microRNA for hsa-miR-101 (Supplementary Fig. S3A), compared with the control siRNA, treatment with siRNA to treatment with DZNep did not significantly alter the EZH2 for 48 hours depleted EZH2 mRNA, but modestly expression of the mature hsa-miR-101 in MO2058 and increased the levels of EED mRNA. SUZ12 mRNA was JeKo-1 cells (Supplementary Fig. S3B). Treatment with unaffected (data not shown). Treatment with siRNA to DZNep (1.0–2.0 mmol/L) also had no effect on the EZH2 was associated with decline in protein levels of EZH2 methylation status of the p16 promoter in MO2058 cells, and SUZ12 but not of EED (Fig. 4B). In addition, there whereas treatment with the DNMT1 inhibitor decitabine was no effect of EZH2 siRNA on the protein levels of

www.aacrjournals.org Clin Cancer Res; 18(22) November 15, 2012 6231

Fiskus et al.

Table 1. Treatment with DZNep induces cell-cycle arrest of MCL cells in a dose-dependent manner

% of cells

Cells and treatment G0–G1 SG2/M MO2058 Untreated 35.51 1.20 60.98 0.47 3.51 0.77 0.5 mmol/L DZNep 36.75 0.56 58.26 0.52 4.99 0.26 1.0 mmol/L DZNep 46.63 1.87 46.67 1.90 6.70 0.11 2.0 mmol/L DZNep 52.89 1.23 39.42 0.86 7.69 0.47 JeKo-1 Untreated 49.08 1.12 48.70 1.02 2.22 0.12 0.5 mmol/L DZNep 60.22 0.10 36.46 0.23 3.32 0.32 1.0 mmol/L DZNep 67.77 0.98 28.76 1.21 3.47 0.28 2.0 mmol/L DZNep 72.06 0.54 24.18 0.70 3.76 0.20

NOTE: Cell-cycle status of MO2058 and JeKo-1 cells treated with the indicated concentrations of DZNep for 24 hours. Values represent the mean of 3 independent experiments SEM.

DNMT1 and HDAC2, whereas HDAC1 levels were slightly Cotreatment with DZNep and PS shows superior anti- increased (Fig. 4B). Attenuation of PRC2 proteins by EZH2 activity, induces more p21 and p27, and exerts EZH2 siRNA and the impairment of the PRC2 complex synergistic apoptotic effects against MCL cells activity were associated with decline in the levels of We next determined the effects of combined treatment 3MeK27H3 but induction of acetylated K27H3 and 3Me with DZNep and PS in MCL cells. Figure 6A and B show that, H3K4 levels in MCL cells (Fig. 4C). This was accompanied as compared with treatment with either agent alone, cotreat- by increase in the levels of p16, p21, and p27, with modest ment with DZNep (250–2,000 nmol/L) and PS synergisti- decline in Cyclin D1 levels (Fig. 4C). EZH2 siRNA treat- cally induced apoptosis of MO2058 and JeKo-1 cells, as ment also induced the F-Box protein FBXO32 levels in highlighted by the combination indices of less than 1.0 MCL cells (Fig. 4C). We also determined the effects of transduction of shRNA to EZH2 on the proliferation and apoptosis of MCL cells. Depletion of EZH2 and SUZ12 by EZH2 shRNA signiﬁcantly inhibited cell proliferation A 50 MO2058 without inducing apoptosis of MCL cells (Fig. 4D and 45 JeKo-1 data not shown). 40 35 Treatment with the pan-HDAC inhibitor panobinostat 30 depletes PRC2 complex proteins with concomitant 25 decline in 3MeK27H3 in MCL cells 20

We next determined whether treatment with the pan- % Apoptosis 15 HDAC inhibitor PS would deplete the levels of EZH2 and 10 SUZ12 with concomitant decline in the levels of 5 3MeK27H3 in MCL cells, as was observed in cultured 0 and primary AML cells (32). Figure 5A shows that PS 0 0.5 2.0 treatment depletes EZH2 and SUZ12 levels, as well as µmol/L, DZNep, 48 h attenuates the levels of BMI1 and DNMT1 in MCL cells B (Fig. 5A and Supplementary Fig. S4C). This was accom- MO2058 JeKo-1 panied by depletion of 3MeK27H3 and marked upregula- 0 0.5 1.0 2.0 0 0.5 1.0 2.0 µmol/L, DZNep, 24 h tion of acetylated-K27H3 and acetylated-K16H4 in the PARP MCL cells (Fig. 5B). In addition, PS treatment also inhib- cleaved PARP ited K119H2A ubiquitylation, possibly because of decline β-Actin in PRC1 activity in MCL cells (9, 14). Treatment with PS also signiﬁcantly depleted the levels of Cyclin D1 in all 3 cell lines (Fig. 5A and SupplementaryFig.4C).Consistent Figure 3. Treatment with DZNep induces dose-dependent apoptosis of with previous reports, here also we observed that, in MCL cells. A, MO2058 and JeKo-1 cells were treated with DZNep as – conjunction with its effects on epigenetic mechanisms, indicated for 48 hours. The percentages of annexin-V positive apoptotic cells were determined by ﬂow cytometry. Columns, mean of 3 PS also induced apoptosis of the cultured MCL cells experiments; bars, SEM. B, immunoblot analyses of PARP and b-actin in (Fig. 5C and D; refs. 29–31). MO2058 and JeKo-1 cells after treatment with DZNep for 24 hours.

6232 Clin Cancer Res; 18(22) November 15, 2012 Clinical Cancer Research

Combined Epigenetic Therapy against Human MCL Cells

2 A EZH2 JeKo-1 B JeKo-1 1.8 EED 1.6 + - 100 nmol/L control siRNA Figure 4. RNA interference– 1.4 -+100 nmol/L EZH2 siRNA mediated depletion of EZH2 induces 1.2 EZH2 p16, p27, p21, and FBXO32 1 1.0 0.48 expression and inhibits cell 0.8 SUZ12 1.0 0.49 proliferation of MCL cells. A, 0.6 quantitative RT-PCR for EZH2 and to control) RQ (relative 0.4 EED1 EED following 48-hour transfection 0.2 with scrambled or EZH2 siRNA. The 0 DNMT1 relative expression was normalized Control siRNA EZH2 siRNA HDAC1 against GAPDH. B and C, 100 nmol/L siRNA, 48 h immunoblot analyses of EZH2, HDAC2 SUZ12, EED, DNMT1, HDAC1, C JeKo-1 β-Actin HDAC2, 3MeK27H3, AcK27H3, FBXO32, p16, p27, p21, CyclinD1, + - 100 nmol/L control siRNA and b-actin in JeKo-1 cells following -+100 nmol/L EZH2 siRNA D 72-hour transfection with scrambled 3MeK27 H3 P = 0.0003 or EZH2 siRNA. D, JeKo-1 cells were 1.0 0.60 1.2 transduced with nontargeting (NT) AcK27 H3 JeKo-1 shRNA or EZH2 shRNA for 48 hours. 1.0 1.45 1 5 Then, 1 10 cells were plated in 3MeK4 H3 0.8 triplicate and cell growth was 1.0 1.39 sh NT sh NT sh EZH2 measured for an additional 72 hours. EZH2 FBXO32 0.6 Graph shows cell growth of 2 1.0 1.35 SUZ12 independent experiments β-Actin p16 0.4 conducted in triplicate. The inset p27 shows the level of EZH2 and SUZ12 (in Cells/mL millions) 0.2 NT shRNA knockdown in the cells. p21 EZH2 shRNA Cyclin D1 0 1.0 0.77 48 72 96 120 β-Actin Hours posttransduction

determined through median dose effect isobologram mor effects against the JeKo-1 xenografts (P < 0.05). Neither analyses. Superior activity of the combination was also each agent alone nor the combination of DZNep and PS noted in Z-138 cells (Supplementary Fig. 5C). This activ- induced weight loss or other physical signs of toxicity in the ity of the combination was also associated with greater xenograft bearing NOD/SCID mice. attenuation of EZH2, SUZ12 and the 3MeK27H3 chromatin mark, with a concomitant, greater induction of Cotreatment with DZNep and PS shows greater 3MeK4H3 mark in the cultured MCL cells (Fig. 6C and cytotoxicity against primary MCL cells than either Supplemental Fig. 5A and B). Cotreatment with DZNep agent alone and PS also caused greater decline in DNMT1 levels. It is Finally, we determined the effects of treatment with noteworthy that, as compared with each agent alone, DZNep and/or PS on the cell viability of patient-derived combined treatment with DZNep and PS induced greater primary MCL cells versus normal CD34þ bone marrow increase in the levels of FBXO32, p21 and p27 in the progenitor cells. Figure 7A shows that exposure to either cultured MO2058 MCL Cells (Fig.6C).Similareffects DZNep or PS for 48 hours induced more loss of viability in 6 were also observed in JeKo-1 and Z-138 cells (Supple- primary MCL cell samples, as compared with normal mentary Fig. 5A). CD34þ progenitor cells. In addition, compared with treatment with each agent alone, cotreatment with DZNep (0.5 Treatment with DZNep and/or PS inhibits tumor or 2.0 mmol/L) and PS (50 nmol/L) induced significantly growth of JeKo-1 xenografts greater lethality of MCL versus normal progenitor cells (Fig. We next determined the in vivo anti-MCL activity of 7A). Similar to the effects observed in the cultured MCL cell treatment with DZNep and/or PS in JeKo-1 cell xenografts lines, combined treatment with DZNep and PS was syner- in NOD/SCID mice. In cohorts of mice, JeKo-1 cells were gistically lethal against primary MCL cells (Fig. 7B). On 1 implanted in the flanks, and the treatment with each agent primary sample, where adequate numbers of MCL cells were or vehicle alone was begun after the flank tumors achieved a available, immunoblot analyses showed that combined size of approximately 150 mm3. As shown in Fig. 6D, treatment with PS and DZNep, as compared with treatment although treatment with DZNep or PS alone also caused with DZNep alone, caused greater decline in the levels of inhibition of tumor growth, combined treatment with EZH2, SUZ12, BMI1, and Cyclin D1 expression levels in DZNep and PS exerted superior growth inhibitory, antitu- primary MCL cells (Fig. 7C).

www.aacrjournals.org Clin Cancer Res; 18(22) November 15, 2012 6233

Fiskus et al.

ACJeKo-1 MO2058 0 10 20 50 0 20 50 100 nmol/L, PS, 24 h 100 DNMT-1 90 JeKo-1 cells Figure 5. Treatment with PS 80 depletes DNMT1 and the PRC2 EZH2 70 complex proteins resulting in loss 60 SUZ12 50 of trimethylation of K27 on histone EED 40 H3. A, immunoblot analyses of 30 DNMT1, EZH2, SUZ12, EED, BMI1, BMI1 Apoptosis % 20 Cyclin D1, and b-actin in MO2058 10 Cyclin D1 and JeKo-1 cells after treatment β 0 -Actin 01020 50 with PS for 24 hours. B, nmol/L, PS, 48 h immunoblot analyses of JeKo-1 MO2058 3MeK27H3, AcK27H3, AcK16 H4, B D Acetyl H3, Acetyl H4, and Histone 0 10 20 0 20 50 100 nmol/L, PS, 24 h 80 H3 in MO2058 and JeKo-1 cells 3MeK27 H3 MO2058 cells 70 after treatment with PS for 24 Ubiq Histone H2A 60 hours. C and D, JeKo-1 and 50 MO2058 cells were treated with PS Ac. K27H3 for 48 hours. The percentages of 40 apoptotic cells were determined by Ac. K16 H4 30

% Apoptosis ﬂow cytometry. Columns, mean of 20 Ac. H3 3 experiments; bars, SEM. 10 Ac. H4 0 02050 100 Histone H3 nmol/L, PS, 48 h

Discussion exposure to DZNep induced the expression of FBXO32 Deregulated epigenome plays a pathogenic role and (atrogin) and depleted Cyclin E (28, 32, 37), as well as up contributes to the aggressive biology in human MCL regulated the levels of p16, p21, and p27 in MCL cells. As (3, 7, 8, 14). In this, the specific role of overexpression of SUZ12 is aberrantly overexpressed and promotes survival the PRC2 and PRC1 proteins EZH2, SUZ12, and BMI1 has of MCL cells (42), DZNep-mediated depletion of SUZ12 also been elucidated (3, 5, 7, 14). Our present studies show may also impair survival of MCL cells. Collectively, these that a combination therapy targeting the deregulated epi- effects of DZNep treatment explain the growth inhibitory genome by cotreatment with DZNep and PS exerts superior and apoptotic effects of DZNep in MCL cells. in vitro and in vivo activity against MCL cells, as compared Notably, the PRC1 protein BMI1 is commonly ampli- with treatment with each agent alone. The KMTase activity fied and overexpressed in MCL cells (9, 14). Our findings of EZH2 and PRC2 is dependent on S-adenosylmethionine show that DZNep treatment also depleted BMI1, thereby (SAM) for maintaining histone methylation and gene affecting PRC1 activity in MCL cells. Although this was silencing (10). By inhibiting the SAM-dependent KMTase not studied here, DZNep may be exerting this effect on activity of EZH2 and PRC2, treatment with DZNep attenu- BMI1 levels by perturbing the recently described, EZH2- ates its catalytic KMTase activity within the PRC2 in MCL regulated miRs that regulate the expression of BMI1 (43). cells (10, 28). We show here, that exposure to DZNep for By inhibiting both PRC2 and PRC1 activity, DZNep as little as 4 hours results in dissociation of EZH2 from treatment reduced the levels of 3MeK27H3 and ubiqui- other PRC2 complex proteins and to DNMT1, before the tylated-K119H2A chromatin marks, but increased the depletion of EZH2 and SUZ12 in MCL cells. DZNep acetylation of K27H3, K56H3 and K16H4. While the treatment did not significantly affect the mRNA level of precise underlying mechanism for this is unknown, expo- EZH2, instead promoted the proteasomal degradation of sure to DZNep also increased the levels of 3MeK4H3 in EZH2. Parenthetically, a recent study showed that induc- MCL cells. Although its impact here is unknown, this is tion of PRAJA1, a RING finger-containing ubiquitin knowntobeapermissivechromatinmarkmediatedby ligase, by DZNep treatment caused rapid ubiquitination trithorax (TRX) family of proteins facilitating gene expres- of EZH2, SUZ12 and EED in cells, which was reversed by sion involved in stem cell-lineage differentiation (44). treatment with a proteasome inhibitor (41). Treatment Collectively, DZNep-mediated inhibition of PRC2 and with DZNep also did not affect the expression levels of PRC1, and the notable alterations in the chromatin hsa-miR-101, which is known to down regulate the levels marks: decreased 3MeK27H3 and increased 3MeK4H3, of EZH2 (39, 40). Consistent with DZNep-mediated are also likely to affect the bivalent chromatin mark, attenuation of EZH2 levels, in MCL cells, we observed which is known to promote lineage differentiation in decreased binding of EZH2 to the promoters of the transformed stem cells (45). While PRC2 is known to known EZH2 target genes, including HOXA9, WNT1, and recruit DNMTs to the promoters of PRC2 target genes and RUNX3. In addition, and consistent with prior reports, PRC2 inhibition should affect DNA methylation of the

6234 Clin Cancer Res; 18(22) November 15, 2012 Clinical Cancer Research

Combined Epigenetic Therapy against Human MCL Cells

A 2.0 MO2058 1.5 C

CI 1.0 MO2058

0.5 - ++- 1.0 µmol/L, DZNep -- ++50 nmol/L PS 0 0 0.2 0.4 0.6 0.8 1.0 EZH2 Fractional effect 1.0 0.64 0.83 0.40 Figure 6. Cotreatment with DZNep DZNep PS Fa CI SUZ12 and PS exerts synergistic anti-MCL (nmol/L) (nmol/L) 1.0 0.70 0.86 0.32 activity and enhances DZNep- 600 30 0.4242 0.457 750 37.5 0.4216 0.584 EED mediated depletion of EZH2, SUZ12, 800 40 0.4223 0.619 and 3MeK27H3 and induction of 900 45 0.4534 0.536 3MeK27 H3 FBXO32, p21, and p27 in MCL cells. 1,000 50 0.5744 0.223 1.0 0.44 0.49 0.35 1,500 75 0.5355 0.459 3MeK4 H3 A and B, MO2058 and JeKo-1 cells 2,000 100 0.70 0.155 were treated with DZNep (dose range 1.0 1.86 1.62 2.47 0.25–2.0 mmol/L) and PS (dose range B 2.0 DNMT1 JeKo-1 5–100 nmol/L) at a fixed ratio for 48 hours. The percentages of apoptotic 1.5 FBXO32 cells were determined by flow 1.0 1.57 1.20 1.78 cytometry. Median dose effect and CI 1.0 Cyclin D1 isobologram analyses were 1.0 0.47 0.59 0.40 p21 conducted using the commercially 0.5 available software Calcusyn. 1.0 1.59 1.21 1.70 Combination indices less than 1.0 0 p27 0 0.2 0.4 0.6 0.8 1.0 1.0 1.47 1.95 2.75 indicate the synergistic interaction of Fractional effect these 2 agents. C, immunoblot β-actin DZNep PS Fa CI analyses of EZH2, SUZ12, EED, (nmol/L) (nmol/L) 3MeK27H3, 3MeK4H3, DNMT1, 250 5 0.2804 0.775 FBXO32, Cyclin D1, p21 p27, and 375 7.5 0.3992 0.685 b 500 10 0.43 0.809 -actin in MO2058 cells treated with 750 15 0.6087 0.626 DZNep and/or PS for 24 hours. D, 800 16 0.7421 0.390 tumor growth of JeKo-1 cells 1,000 20 0.7189 0.540 1,500 30 0.7909 0.580 implanted in the flank of NOD/SCID mice and treated as indicated for 2 weeks. , tumor volumes significantly D 3,500 less (P ¼ 0.02) in the combination Vehicle

than treatment with DZNep alone at )

3 3,000 10 mg/kg PS the end of treatment. , tumor ﬁ P ¼ volumes signi cantly less ( 0.008) 2,500 1 mg/kg DZNep in combination than treatment with DZ + PS PS alone at the end of treatment. 2,000

1,500 1 tumor volume (mm * 1,000

3 * JeKo-

500 150 mm

0 13581115

Days of treatment

target gene promoters, this was not the case for p16 gene tion of p16 expression in MCL cells. BMI1 has also been promoter where CpG methylation was found to be unaf- shown to be required to reinforce bivalent domains at key fected by DZNep treatment. However, both EZH2 and gene loci for maintaining lineage speciﬁcation poised for BMI1-mediated chromatin effects have been implicated activation in hematopoietic stem cells (48). in the repression of Ink4a/Arf locus encoding p16 and DZNep treatment in MCL cells increased the expression ARF (p19; refs. 46, 47). Therefore, notwithstanding the of the cell cycle inhibitory proteins p16, p21 and p27 in absence of its effects on DNA methylation, it is not MCL cells and caused induction of the E3 ubiquitin ligase surprising that treatment with DZNep led to up regula- FBXO32 (28, 32, 34, 37). DZNep induced FBXO32 levels

www.aacrjournals.org Clin Cancer Res; 18(22) November 15, 2012 6235

Fiskus et al.

A 100 B 1.5 90 Primary MCL cells (n = 6) Primary MCL cells Figure 7. Cotreatment with DZNep and PS exerts synergistic cytotoxic 80 Normal CD34+ cells (n = 3) * 70 1.0 effects against primary MCL cells. 60 CI A, primary MCL (n ¼ 6) and normal 50 0.5 CD34þ cells (n ¼ 3) were treated 40 † with DZNep and/or PS for 48 hours. 30 20 0 , loss of viability values % Nonviable cells 0 0.2 0.4 0.6 0.8 1.0 ﬁ P < 10 Fractional effect signi cantly greater ( 0.05) than 0 those resulting from treatment with DZNep PS Fa CI † trol (nmol/L) (nmol/L) either agent alone. , values Con DZNep DZNep ep + PS ep + PS ﬁ þ N N 1,000 20 0.641 0.111 signi cantly less in normal CD34 Z Z 1,500 30 0.645 0.161 50 nmol/L PS progenitor cells than those 2,000 40 0.625 0.258 0.5 µmol/L 2.0 µmol/L observed in primary MCL cells. B, µmol/L D µmol/L D 2,500 50 0.725 0.128 0.5 2.0 median dose effect and Primary MCL C 1.5 isobologram analyses of primary Primary MCL cells MCL cells treated with DZNep and 1.0 PS for 48 hours. CI values less than

CI 1.0 indicate the synergistic

0.5 interaction of these 2 agents. C, immunoblot analyses of EZH2, 1.0 µmol/L DZNep 20 nmol/L PS 20 nmol/L Control DZNep + PS DZNep + SUZ12, BMI1, Cyclin D1, and 0 EZH2 0 0.2 0.4 0.6 0.8 1.0 b-actin in primary MCL cells Fractional effect treated with DZNep and/or PS for SUZ12 DZNep PS Fa CI 24 hours. D, schematic model of BMI1 (nmol/L) (nmol/L) the mechanism of action of DZNep 1,000 20 0.393 0.570 and PS in MCL cells. Treatment 1,500 30 0.402 0.813 Cyclin D1 2,000 40 0.551 0.557 with DZNep results in the inhibition β-Actin 2,500 50 0.573 0.637 of EZH2 expression and depletion of lysine 27 trimethylation on Histone H3 (H3K27Me3). This Anti-MCL activity of knocking down EZH2 and histone deacetylases results in induction of p16, p21, and D p27. PS inhibits the activity of Class Panobinostat I (HDAC1, HDAC2, and HDAC3) DZNep and Class II HDACs, particularly HDAC6, the deacetylase for hsp90. HDAC6 Inhibition of HDAC6 activity results in the hyperacetylation of hsp90, thereby reducing chaperone hsp90 association of hsp90 with EZH2. This leads to depletion of EZH2 HDAC expression and subsequent loss of Ac 1, 2, 3 H3K27Me3. Inhibition of hsp90 EZH2 hsp90 function also results in depletion of CDK4 and cyclin D1 expression levels. PS treatment also induces acetylation of Histone H3 and H4 CDK4 Cyclin D1 and upregulates p16, p21, and p27 in MCL cells. Combined treatment p16 with DZNep and PS causes greater p21 depletion of H3K27Me3 and H3/H4 Acetylation H3K27Me3 p27 greater induction of p16, p21, and p27 in MCL cells.

were associated with decline in the levels of Cyclin E, which inhibition in MCL cells. This may be because DZNep is known to be targeted by FBXO32 (28). Induction of treatment induced significantly higher levels of FBXO32 in FBXO32 has also been shown to promote apoptosis MCL cells than treatment with EZH2 siRNA. Other off-target (28, 33). Similar to DZNep treatment, depletion of EZH2 effects of DZNep may also be responsible for this discrep- by siRNA also induced the levels of p16, p21, p27, and ancy. Other studies have also shown that DZNep is not a FBXO32, which indicates that DZNep-mediated depletion specific inhibitor of EZH2, and can affect the methylation of EZH2 is responsible for modulating these protein levels status of chromatin marks other than H3K27 at high con- and growth inhibition in MCL cells. It is notable that while centrations (49). Searching for more specific and less toxic treatment with DZNep induced growth inhibition and EZH2 antagonists is a clear priority but prototype, direct apoptosis, EZH2 shRNA treatment only induced growth EZH2 antagonists are not yet available for testing.

6236 Clin Cancer Res; 18(22) November 15, 2012 Clinical Cancer Research

Combined Epigenetic Therapy against Human MCL Cells

As was reported for other cancer cell-types, treatment mice implanted with human AML cell xenografts without with PS also depleted EZH2, SUZ12, and BMI1 in MCL resulting in toxicity to the mice (34). It is noteworthy that, cells. Concomitantly, this was associated with decline in so far, neither DZNep nor any of its active analogues has the 3MeK27H3 but increase in 3MeK4H3, acetylated-K27 been administered to humans. On the other hand, Phase H3, and acetylated-K16 H4 chromatin marks. PS treat- I/II clinical studies have highlighted the activity of PS in ment also inhibited H2A ubiquitylation, possibly because patients with hematologic malignancies, especially those of decline in PRC1 activity in MCL cells (9, 14). Impor- with lymphoma (50). Findings presented here support tantly, treatment with PS also significantly depleted the the rationale to further test the in vivo anti-MCL efficacy of levelsofCyclinD1,aswellasinducedapoptosisof the combination of PS with an EZH2 antagonist. In cultured and primary MCL cells. It should be noted that addition, our findings also create the rationale to develop we have previously documented multiple nonepigenetic andtestcombinationtherapiesthattargetthederegulated mechanisms that may also contribute to the anti-MCL epigenome in human MCL cells. activity of PS (29, 30). Therefore, the overall anti-MCL activity of PS is likely to be because of multiple mechan- Disclosure of Potential Conflicts of Interest isms. Our findings also show that, as compared with K.N. Bhalla: commercial research grant and honoraria of speakers’ bureau from Novartis Pharmaceuticals. No potential conflicts of interest were treatment with each agent alone, cotreatment with disclosed by the other authors. DZNep and PS resulted in greater depletion of EZH2, SUZ12, DNMT1, Cyclin D1, and 3MeK27H3, with con- Authors' Contributions comitant induction of 3MeK4H3, FBXO32, p21, and p27. Conception and design: P.W. Atadja, K.N. Bhalla These molecular perturbations resulting from cotreat- Development of methodology: K.N. Bhalla ment with DZNep and PS are likely to have contributed Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): W. Fiskus, R. Rao, R. Balusu, E.M. Sotomayor, S.L. to the superior anti-MCL activity of the combination (Fig. Hembruff, K.N. Bhalla 7D). Indeed, cotreatment with DZNep and PS synergis- Analysis and interpretation of data (e.g., statistical analysis, biosta- tically induced apoptosis of cultured and primary MCL tistics, computational analysis): W. Fiskus, U. Mudunuru, K.N. Bhalla Writing, review, and/or revision of the manuscript: S. Ganguly, E.M. cells. Notably, combined treatment with DZNep and PS Sotomayor, J.E. Smith, P.W. Atadja, K.N. Bhalla was only slightly more toxic than each agent alone against Administrative, technical, or material support (i.e., reporting or orga- þ nizing data, constructing databases): R. Rao, R. Balusu, J. Tao, J.E. Smith, normal CD34 progenitor cells. The combination also S.L. Hembruff, K.N. Bhalla induced significantly more apoptosis in primary MCL Study supervision: K.N. Bhalla versus normal CD34þ progenitor cells. These results Development and provision of critical reagents: P.W. Atadja Other: Provided a key reagent (drug) and contributed with mechanis- indicate that cotreatment with DZNep and PS is selec- tic insights on the mode of action of the drug related to the study, V.E. tively more toxic against MCL cells. This is further con- Marquez firmed by our findings that the combination exerts greater The costs of publication of this article were defrayed in part by the in vivo antitumor effects against the JeKo-1 xenograft payment of page charges. This article must therefore be hereby marked model, without inducing weight loss or other physical advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate side effects during treatment of the NOD/SCID mice. In this fact. the current study, we used previously determined doses of Received March 14, 2012; revised July 13, 2012; accepted August 21, 2012; DZNep and PS that improved survival of NOD/SCID published OnlineFirst August 29, 2012.

References 1. Andersen NS, Jensen MK, de Nully Brown P, Geisler CH. A Danish 7. Visser HP, Gunster MJ, Kluin-Nelemans HC, Manders EM, Raaphorst population-based analysis of 105 mantle cell lymphoma patients: FM, Meijer CJ, et al. The polycomb group protein EZH2 is upregulated incidences, clinical features, response, survival and prognostic fac- in proliferating, cultured human mantle cell lymphoma. Br J Haematol tors. Eur J Cancer 2002;38:401–8. 2001;112:950–8. 2. Sander B. Mantle cell lymphoma: recent insights into pathogenesis, 8. Kienle D, Katzenberger T, Ott G, Saupe D, Benner A, Kohlhammer H, clinical variability, and new diagnostic markers. Semin Diagn Pathol et al. Quantitative gene expression deregulation in mantle-cell lym- 2011;28:245–55. phoma: correlation with clinical and biologic factors. J Clin Oncol 3. Jares P, Colomer D, Campo E. Genetic and molecular pathogenesis of 2007;25:2770–7. mantle cell lymphoma: perspectives for new targeted therapeutics. Nat 9. Bracken AP, Helin K. Polycomb group proteins: navigators of lineage Rev Cancer 2007;7:750–62. pathways led astray in cancer. Nat Rev Cancer 2009;9:773–84. 4. Hartmann EM, Campo E, Wright G, Lenz G, Salaverria I, Jares P, et al. 10. Simon JA, Lange CA. Roles of the EZH2 histone methyltransferase in Pathway discovery in mantle cell lymphoma by integrated analysis of cancer epigenetics. Mutat Res 2008;647:21–9. high-resolution gene expression and copy number proﬁling. Blood 11. Kleer CG, Cao Q, Varambally S, Shen R, Ota I, Tomlins SA, et al. EZH2 2010;116:953–61. is a marker of aggressive breast cancer and promotes neoplastic 5. Royo C, Salaverria I, Hartmann EM, Rosenwald A, Campo E, BeaS. transformation of breast epithelial cells. Proc Natl Acad Sci U S A The complex landscape of genetic alterations in mantle cell lymphoma. 2003;100:11606–11. Semin Cancer Biol 2011;21:322–34. 12. Varambally S, Dhanasekaran SM, Zhou M, Barrette TR, Kumar-Sinha 6. Nagy B, Lundan T, Larramendy ML, Aalto Y, Zhu Y, Niini T, et al. C, Sanda MG, et al. The polycomb group protein EZH2 is involved in Abnormal expression of apoptosis-related genes in haematological progression of prostate cancer. Nature 2002;419:624–9. malignancies: overexpression of MYC is poor prognostic sign in 13. Bachmann IM, Halvorsen OJ, Collett K, Stefansson IM, Straume O, mantle cell lymphoma. Br J Haematol 2003;120:434–41. Haukaas SA, et al. EZH2 expression is associated with high

www.aacrjournals.org Clin Cancer Res; 18(22) November 15, 2012 6237

Fiskus et al.

proliferation rate and aggressive tumor subgroups in cutaneous mel- inhibitors in mantle cell lymphoma cells in vitro and in vivo. Mol Cancer anoma and cancers of the endometrium, prostate, and breast. J Clin Ther 2011;10:1686–97. Oncol 2006;24:268–73. 32. Fiskus W, Pranpat M, Balasis M, Herger B, Rao R, Chinnaiyan A, et al. 14. Martin-Perez D, Piris MA, Sanchez-Beato M. Polycomb proteins in Histone deacetylase inhibitors deplete enhancer of zeste 2 and asso- hematologic malignancies. Blood 2010;116:5465–75. ciated polycomb repressive complex 2 proteins in human acute 15. van Kemenade FJ, Raaphorst FM, Blokzijl T, Fieret E, Hamer KM, Satijn leukemia cells. Mol Cancer Ther 2006;5:3096–104. DP, et al. Coexpression of BMI-1 and EZH2 polycomb-group proteins 33. Fiskus W, Buckley K, Rao R, Mandawat A, Yang Y, Joshi R, et al. Pano- is associated with cycling cells and degree of malignancy in B-cell non- binostat treatment depletes EZH2 and DNMT1 levels and enhances Hodgkin lymphoma. Blood 2001;97:3896–901. decitabine mediated de-repression of JunB and loss of survival of 16. Velichutina I, Shaknovich R, Geng H, Johnson NA, Gascoyne RD, human acute leukemia cells. Cancer Biol Ther 2009;8:939–50. Melnick AM, et al. EZH2-mediated epigenetic silencing in germinal 34. Fiskus W, Wang Y, Sreekumar A, Buckley KM, Shi H, Jillella A, et al. center B cells contributes to proliferation and lymphomagenesis. Combined epigenetic therapy with the histone methyltransferase Blood 2010;116:5247–55. EZH2 inhibitor 3-deazaneplanocin A and the histone deacetylase 17. Morin RD, Johnson NA, Severson TM, Mungall AJ, An J, Goya R, et al. inhibitor panobinostat against human AML cells. Blood 2009;114: Somatic mutations altering EZH2 (Tyr641) in follicular and diffuse large 2733–43. B-cell lymphomas of germinal-center origin. Nat Genet 2010;42: 35. Chou TC, Talalay P. Quantitative analysis of dose-effect relationships: 181–5. the combined effects of multiple drugs or enzyme inhibitors. Adv 18. Sneeringer CJ, Scott MP, Kuntz KW, Knutson SK, Pollock RM, Richon Enzyme Regul 1984;22:27–55. VM, et al. Coordinated activities of wild-type plus mutant EZH2 drive 36. Balusu R, Fiskus W, Rao R, Chong DG, Nalluri S, Mudunuru U, et al. tumor-associated hypertrimethylation of lysine 27 on histone H3 Targeting levels or oligomerization of nucleophosmin 1 induces dif- (H3K27) in human B-cell lymphomas. Proc Natl Acad Sci U S A ferentiation and loss of survival of human AML cells with mutant NPM1. 2010;107:20980–5. Blood 2011;118:3096–106. 19. Yap DB, Chu J, Berg T, Schapira M, Cheng SW, Moradian A, et al. 37. Jiang X, Tan J, Li J, Kivimae€ S, Yang X, Zhuang L, et al. DACT3 is an Somatic mutations at EZH2 Y641 act dominantly through a mecha- epigenetic regulator of Wnt/beta-catenin signaling in colorectal cancer nism of selectively altered PRC2 catalytic activity, to increase H3K27 and is a therapeutic target of histone modifications. Cancer Cell trimethylation. Blood 2011;117:2451–9. 2008;13:529–41. 20. Ryan RJ, Nitta M, Borger D, Zukerberg LR, Ferry JA, Harris NL, et al. 38. Chen J, Fiskus W, Eaton K, Fernandez P, Wang Y, Rao R, et al. EZH2 Codon 641 Mutations are Common in BCL2-Rearranged Ger- Cotreatment with BCL-2 antagonist sensitizes cutaneous T-cell lym- minal Center B Cell Lymphomas. PLoS One 2011;6:e28585. phoma to lethal action of HDAC7-Nur77-based mechanism. Blood 21. Bea S, Tort F, Pinyol M, Puig X, Hernandez L, Hernandez S, et al. BMI-1 2009;113:4038–48. gene amplification and overexpression in hematological malignancies 39. Varambally S, Cao Q, Mani RS, Shankar S, Wang X, Ateeq B, et al. occur mainly in mantle cell lymphomas. Cancer Res 2001;61:2409–12. Genomic loss of microRNA-101 leads to overexpression of histone 22. Vire E, Brenner C, Deplus R, Blanchon L, Fraga M, Didelot C, et al. The methyltransferase EZH2 in cancer. Science 2008;322:1695–9. Polycomb group protein EZH2 directly controls DNA methylation. 40. Friedman JM, Liang G, Liu CC, Wolff EM, Tsai YC, Ye W, et al. The Nature 2006;439:871–4. putative tumor suppressor microRNA-101 modulates the cancer epi- 23. Schlesinger Y, Straussman R, Keshet I, Farkash S, Hecht M, Zimmer- genome by repressing the polycomb group protein EZH2. Cancer Res man J, et al. Polycomb-mediated methylation on Lys27 of histone H3 2009;69:2623–9. pre-marks genes for de novo methylation in cancer. Nat Genet 41. Zoabi M, Sadeh R, di Bie P, Ciechanover A. PRAJA1 is a ubiquitin ligase 2007;39:232–6. for the polycomb repressive complex 2 proteins. Biochem Biophys 24. Kondo Y, Shen L, Cheng AS, Ahmed S, Boumber Y, Charo C, et al. Res Commun 2011;408:393–8. Gene silencing in cancer by histone H3 lysine 27 trimethylation inde- 42. Martín-Perez D, Sanchez E, Maestre L, Suela J, Vargiu P, Di Lisio L, pendent of promoter DNA methylation. Nat Genet 2008;40:741–50. et al., Deregulated expression of the polycomb group protein SUZ12 25. McGarvey KM, Greene E, Fahrner JA, Jenuwein T, Baylin SB. DNA targets genes characterizes mantle cell lymphoma. Am J Pathol methylation and complete transcriptional silencing of cancer genes 2010;177:930–42. persist after depletion of EZH2. Cancer Res 2007;67:5097–102. 43. Cao Q, Mani RS, Ateeq B, Dhanasekaran SM, Asangani IA, Prensner 26. Hernandez-Mu noz~ I, Taghavi P, Kuijl C, Neefjes J, van Lohuizen M. JR, et al. Coordinated regulation of polycomb group complexes Association of BMI1 with polycomb bodies is dynamic and requires through microRNAs in cancer. Cancer Cell 2011;20:187–99. PRC2/EZH2 and the maintenance DNA methyltransferase DNMT1. 44. Schuettengruber B, Martinez AM, Iovino N, Cavalli G. Trithorax group Mol Cell Biol 2005;25:11047–58. proteins: switching genes on and keeping them active. Nat Rev Mol 27. Glazer RI, Hartman KD, Knode MC, Richard MM, Chiang PK, Tseng Cell Biol 2011;12:799–814 CK, et al. 3-Deazaneplanocin: a new and potent inhibitor of S-adeno- 45. Taberlay PC, Kelly TK, Liu CC, You JS, De Carvalho DD, Miranda TB, sylhomocysteine hydrolase and its effects on human promyelocytic et al. Polycomb-repressed genes have permissive enhancers that leukemia cell line HL-60. Biochem Biophys Res Commun 1986;135: initiate reprogramming. Cell 2011;147:1283–94. 688–94. 46. Kotake Y, Cao R, Viatour P, Sage J, Zhang Y, Xiong Y. pRB family 28. Tan J, Yang X, Zhuang L, Jiang X, Chen W, Lee PL, et al. Pharmacologic proteins are required for H3K27 trimethylation and polycomb repres- disruption of polycomb-repressive complex 2-mediated gene repression complexes binding to and silencing p16INK4A tumor suppressor sion selectively induces apoptosis in cancer cells. Genes Dev gene. Genes Dev 2007;21:49–54. 2007;21:1050–63. 47. Dhawan S, Tschen SI, Bhushan A. Bmi-1 regulates the Ink4a/Arf locus 29. Rao R, Nalluri S, Fiskus W, Savoie A, Buckley KM, Ha K, et al. Role of to control pancreatic beta-cell proliferation. Genes Dev 2009;23: CAAT/enhancer binding protein homologous protein in panobinostat- 906–11. mediated potentiation of bortezomib-induced lethal endoplasmic 48. Oguro H, Yuan J, Ichikawa H, Ikawa T, Yamazaki S, Kawamoto H, et al. reticulum stress in mantle cell lymphoma cells. Clin Cancer Res Poised lineage specification in multipotential hematopoietic stem and 2010;16:4742–54. progenitor cells by the polycomb protein Bmi1. Cell Stem Cell. 2010; 30. Rao R, Lee P, Fiskus W, Yang Y, Joshi R, Wang Y, et al. Co-treatment 6:279–86. with heat shock protein 90 inhibitor 17-dimethylaminoethylamino-17- 49. Miranda TB, Cortez CC, Yoo CB, Liang G, Abe M, Kelly TK , et al. demethoxygeldanamycin (DMAG) and vorinostat: a highly active com- DZNep is a global histone methylation inhibitor that reactivates devel- bination against human mantle cell lymphoma (MCL) cells. Cancer Biol opmental genes not silenced by DNA methylation. Mol Cancer Ther Ther 2009;8:1273–80. 2009;8:1579–88. 31. Dasmahapatra G, Lembersky D, Son MP, Attkisson E, Dent P, Fisher 50. Copeland A, Buglio D, Younes A. Histone deacetylase inhibitors in RI, et al. Carfilzomib interacts synergistically with histone deacetylase lymphoma. Curr Opin Oncol 2010;22:431–6.

6238 Clin Cancer Res; 18(22) November 15, 2012 Clinical Cancer Research

Superior Efficacy of a Combined Epigenetic Therapy against Human Mantle Cell Lymphoma Cells

Warren Fiskus, Rekha Rao, Ramesh Balusu, et al.

Clin Cancer Res 2012;18:6227-6238. Published OnlineFirst August 29, 2012.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-12-0873

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2012/08/29/1078-0432.CCR-12-0873.DC1

Cited articles This article cites 50 articles, 24 of which you can access for free at: http://clincancerres.aacrjournals.org/content/18/22/6227.full#ref-list-1

Citing articles This article has been cited by 4 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/18/22/6227.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/18/22/6227. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.