promoter methylation analysis reveals that kcna5 ion channel silencing supports ewing ... ·...

Chromatin, Epigenetics, and RNA Regulation

Promoter Methylation Analysis Reveals ThatKCNA5 Ion Channel Silencing Supports EwingSarcoma Cell ProliferationKatherine E. Ryland1,2,3, Allegra G. Hawkins2,3, Daniel J.Weisenberger4,5, Vasu Punj5,Scott C. Borinstein6, Peter W. Laird7, Jeffrey R. Martens8, and Elizabeth R. Lawlor2,3,9

Abstract

Polycomb proteins are essential regulators of gene expressionin stem cells and development. They function to reversiblyrepress gene transcription via posttranslational modification ofhistones and chromatin compaction. In many human cancers,genes that are repressed by polycomb in stem cells are subject tomore stable silencing via DNA methylation of promoter CpGislands. Ewing sarcoma is an aggressive bone and soft-tissuetumor that is characterized by overexpression of polycombproteins. This study investigates the DNA methylation statusof polycomb target gene promoters in Ewing sarcoma tumorsand cell lines and observes that the promoters of differentiationgenes are frequent targets of CpG-island DNA methylation. Inaddition, the promoters of ion channel genes are highly dif-ferentially methylated in Ewing sarcoma compared with non-malignant adult tissues. Ion channels regulate a variety ofbiologic processes, including proliferation, and dysfunction of

these channels contributes to tumor pathogenesis. In particular,reduced expression of the voltage-gated Kv1.5 channel has beenimplicated in tumor progression. These data show that DNAmethylation of the KCNA5 promoter contributes to stableepigenetic silencing of the Kv1.5 channel. This epigeneticrepression is reversed by exposure to the DNA methylationinhibitor decitabine, which inhibits Ewing sarcoma cell prolif-eration through mechanisms that include restoration of theKv1.5 channel function.

Implications: This study demonstrates that promoters of ionchannels are aberrantly methylated in Ewing sarcoma andthat epigenetic silencing of KCNA5 contributes to tumor cellproliferation, thus providing further evidence of the importanceof ion channel dysregulation to tumorigenesis.Mol Cancer Res; 14(1);26–34. �2015 AACR.

IntroductionA fundamental trait of cancer cells is their ability to sustain

proliferation (1). Cancer cells accomplish this by hijacking phys-iologic pathways and silencing or mutating tumor-suppressorgenes that control cell-cycle progression (1). Moreover, the pro-liferative phenotype of cancer cells is known to contribute totumor relapse (2–4), a major impediment to cancer cures. There-fore, continued elucidation of the mechanisms that promote

cancer cell proliferation is key for developing novel anticancertherapies.

Disruptions to gene expression in cancer can be achieved byboth genetic and epigenetic mechanisms that lead to eitheraberrant induction or repression of target transcripts (reviewedin refs. 5, 6). Although genetic loss of function is usually achievedby gene deletion or inactivating mutations, epigenetic loss offunction ismost often associated with aberrant repression of genetranscription mediated by posttranslational histone modifica-tions and DNA methylation. In normal stem cells, polycombgroup proteins reversibly repress target genes and, in so doing,prevent differentiation, maintain stemness, and control develop-ment until appropriate developmental cues are received (7, 8).Importantly, the promoters of these stem cell targets of polycombcomplexes are frequently targeted for DNA hypermethylation incancer (9–11). In this way, more stable, and potentially irrevers-ible, gene silencing is effected, and cancer cells become locked in astem cell–like state.

Ewing sarcoma is an aggressive bone and soft-tissue tumor thatpresents most often in adolescents and young adults. It is a tumorof presumed stem cell origin and is characterized by an undiffer-entiated cellular phenotype and the presence of a tumor-initiatingoncogenic fusion gene, most commonly EWS–FLI1 (12). Ewingsarcoma is genetically quiet, with a low rate of genetic mutation(13–15), and tumor pathogenesis is largely mediated by EWS–FLI1 and its impact on the epigenome (16, 17). Overexpression ofpolycomb proteins, in particular EZH2 and BMI-1, is evident in

1Department of Pharmacology, University of Michigan, Ann Arbor,Michigan. 2Translational Oncology Program, University of Michigan,Ann Arbor, Michigan. 3Department of Pediatrics, University of Michi-gan, Ann Arbor, Michigan. 4Department of Biochemistry and Molec-ular Biology, University of SouthernCalifornia, LosAngeles,California.5Keck School of Medicine, University of Southern California, LosAngeles, California. 6Department of Pediatrics, Vanderbilt University,Nashville, Tennessee. 7Van Andel Research Institute, Grand Rapids,Michigan. 8DepartmentofPharmacologyandTherapeutics,Universityof Florida, Gainesville, Florida. 9Department of Pathology, Universityof Michigan, Ann Arbor, Michigan.

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

Corresponding Author: Elizabeth R. Lawlor, University of Michigan, NCRCBuilding 520, Rm 1352, 1600 Huron Parkway, Ann Arbor, MI 48109-2800. Phone:734-615-4814; Fax: 734-764-9017; E-mail: [email protected]

doi: 10.1158/1541-7786.MCR-15-0343

�2015 American Association for Cancer Research.

MolecularCancerResearch

Mol Cancer Res; 14(1) January 201626

on February 13, 2021. © 2016 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst November 16, 2015; DOI: 10.1158/1541-7786.MCR-15-0343

http://mcr.aacrjournals.org/

nearly all Ewing sarcoma tumor cells, and both proteins contrib-ute to the tumor phenotype (18–21). In addition, altered DNAmethylation of cancer-associated gene promoters has beendescribed, and preclinical studies demonstrate that exposure ofEwing sarcoma cells to hypomethylating agents inhibits tumorgrowth (22, 23).

In the current study, we used a custom-designed DNA meth-ylation array to investigate whether the promoters of polycombtarget genes are targeted for abnormal DNA methylation inEwing sarcoma. Our results reveal that, like other cancer, CpGislands in the promoters of differentiation-associated genes aretargets of DNA methylation. In addition, they uncover theunanticipated finding that promoters of ion channel genes arefrequent targets of aberrant DNA methylation. In particular,these studies demonstrate that DNA hypermethylation contri-butes to epigenetic repression of the KCNA5 locus and that theresulting suppression of the Kv1.5 ion channel supports cancercell proliferation.

Materials and MethodsCell lines and tumor tissues

Ewing sarcoma cell lines were provided by Dr. Timothy TricheChildren's Hospital Los Angeles (CHLA, Los Angeles, CA) and theChildren's Oncology Group cell bank (www.cogcell.org). Humanvascular endothelial cells (HuVEC) were obtained from Lonza(191027). MRC-5 cells were obtained from the ATCC (CCL-171).Human bone marrow–derived mesenchymal stem cell (MSC)lines andhuman embryonic stem-cell (hESC)-derived neural creststem cells (NCSC) were obtained as previously described (24).Identities were confirmed by short tandem repeat profiling. Allcell lines were cultured in standard cell culture media, supple-mentedwith 10% FBS (Atlas Biologicals, Inc.; F-500-A) at 37�C in5% CO2. Ewing sarcoma tumor specimens were acquired fromthe Vanderbilt University pathology archives. Approval from theVanderbilt University Institutional Review Board was grantedprior to tissue acquisition.

Proliferation analysisCells were plated at a density of 200,000 cells per 6-well plate

and left for 24 hours prior to a 72-hour drug treatment where thecells were treated every 24 hours. Brightfield images of the cellswere captured on the Olympus CKX41 microscope on the 10�objective by the Lumenera Infinity 3-1 1.4 Megapixel camera.Proliferation was determined by trypan-blue exclusion and EdUincorporation. 5-ethynyl-2'-deoxyuridine (EdU) was added tofresh medium of the cells after the 72-hour incubation for 2hours before harvest at a concentration of 10 mmol/L. EdUincorporation was determined with the Click-iT Plus EdU AlexaFluor 488 flow cytometry assay Kit (Life Technologies; C10632)following the manufacturer's instructions, performed on theAccuri C6 flow cytometer and analyzed using FlowJo V10.

Pharmacologic studiesDiphenyl phosphine oxide-1 (DPO-1; 310 nmol/L; Tocris

Bioscience; 2533) was prepared in ethanol, 40Aminopyridine(40AP; 50 mmol/L; Sigma-Aldrich; 275875) was prepared in anaqueous solution, and 5-aza-20-deoxycytidine (50AZA-CdR/deci-tabine; 100 nmol/L; Sigma-Aldrich; A3656) was diluted indimethyl sulfoxide (Fisher; D128-500). Cells were pretreated atthese concentrations for 72 or 96 hours.

Quantitative real-time PCRTotal RNA extraction was performed using the RNeasy Plus

Mini Kit (Qiagen; 74136) or Quick-RNA Miniprep (Zymo;R1055), and cDNA was generated using iScript (Bio-Rad;BIO1708891). qRT-PCR was performed using validated SYBRprimers (KCNA5, GAPDH, HPRT; IDT). Analysis was performedin triplicate using the Lightcycler 480 System (Roche AppliedScience). Data were analyzed by normalizing average Ct values ofthe gene of interest (KCNA5) to the geometric mean of referencegenes (HPRT and GAPDH) within each sample using the DDCtmethod. Primers for KCNA5 were as follows: Forward: 50-GTAACG TCA AGG CCA AGA GC-30; Reverse: 50-TCC CAT TCC CTACTC CAC TG-30.

Statistical analysisStatistical analyses were performed using GraphPad Prism 6. P

values of less than 0.05 across at least 3 independent experimentswere considered significant.

DNA methylationDNA methylation was interrogated on a previously custom-

designed Illumina Golden-Gate bead array. This array consists of1,536 probes designed to detect DNAmethylation at CpG islandswithin the promoters of >1,400 genes that are established targetsof polycomb regulation in human embryonic stem cells (8). Twoindependent batches of samples were analyzed using this array,the first included 4 Ewing sarcoma tumors and 5 Ewing sarcomacell lines and the second comprised 52 samples from 11 differentnontransformed adult tissues that were obtained at rapid autop-sy (25). Analysis of DNA methylation (as determined by betavalues) was restricted to 1,297 probes that were not targeted topromoters of X- or Y-linked genes and were successfully andreproducibly detected in over 75% of the samples. Probes thatcontained repetitive sequence of more than 10 base pairs or thatcontained SNPswithin CpG islands were also excluded, resultingin evaluable data for 1,204 loci. Differential methylation wasdefined as a median difference in beta value of at least 0.2 and aBonferroni multiple test-corrected P value of less than 0.05between Ewing sarcoma and nonmalignant adult tissues. Geneontology was determined using DAVID and enriched categoriesdetermined relative to all genes, as well as relative to the 1,204evaluable loci (26).

MethyLight studiesMethyLight analyses were carried out as previously described

(27). Briefly, genomic DNA was isolated using the GenomicDNA Clean & Concentrator Kit (Zymo Research; D4011). Sodi-um bisulfite conversion of genomic DNA was performed usingthe EZ DNA Methylation Kit (Zymo Research; D5002). Aftersodium bisulfite conversion, the genomic DNA was amplifiedusing MethyLight, a fluorescence-based real-time quantitativePCR, as previously described (27, 28), with the use of the EpiTectMethyLight PCR Kit (Qiagen; 59496) and EpiTect PCR controlDNA set (Qiagen; 59695). The assay for KCNA5 consistedof Forward primer: 50-ATCGTAATCGGTTTAGTTTCGACG-30;Reverse primer: 50-AATAAAAACGTATCTCGTCCGCG-30; Probe:6FAM-CGTTAACCGACCTCCGCAAACGACC-BHQ1 (IDT; AppliedBiosystems; and Biosearch Technologies). Control probes andconditions have been published previously (27). Analysis wasperformed in triplicate using the Lightcycler 480 System (RocheApplied Science), and the average percentage of methylated

Methylation of KCNA5 Promotes Ewing Sarcoma Proliferation

www.aacrjournals.org Mol Cancer Res; 14(1) January 2016 27




reference (PMR) for each sample was determined as detailedpreviously (27).

ResultsDNAmethylation of polycomb target gene promoters in Ewingsarcoma differs from normal adult tissues

Ewing sarcoma is characterized by an undifferentiated histol-ogy, and transcriptional profiling and tumor biology studiessuggest that this primitive state is a consequence of both cellularorigin and a block in cellular differentiation (18, 20, 21, 24).DNAhypermethylation at the promoters of developmentally criticalpolycomb target genes has been observed in many adult tumorsand is associatedwith transcriptional silencing of genes that directcell differentiation (refs. 10, 11 and reviewed in ref. 6). Given thehigh level of polycombprotein expression that is evident in Ewingsarcoma, and the critical roles of these proteins in tumor path-ogenesis, we hypothesized that abnormal DNA methylation ofpolycomb target genes might be a feature of Ewing sarcoma. Tobegin to address this, we investigated the methylation status ofgene promoters using a custom-designed Illumina GoldenGatebead array (see Materials and Methods). Five Ewing sarcoma celllines and four primary tumors were analyzed, and their DNAmethylation data were compared with a panel of 52 normal adulttissues. As shown, despite their diverse origins, promoter DNAmethylation was remarkably similar in all adult tissues, but wasdistinct from Ewing sarcoma (Fig. 1A). Notably, Ewing sarcomatumors and cell lines clustered together and showed a similaroverall pattern of DNA methylation, demonstrating that cellculture was not the primary driver of the abnormal DNA meth-ylation profiles. Comparison of Ewing sarcoma with normaltissues identified 547 probes with differences in median betavalue of �0.2, and 498 of these probes reached statistical signif-icance, as described in Materials and Methods (SupplementaryTable S1). A total of 265 loci showed increased methylation inEwing sarcoma, and 222 loci had lower levels ofmethylation thanin adult tissues. As expected, an increase in DNAmethylation wassignificantly more likely at CpG islands than nonislands. Specif-ically, 378 differentially methylated probes were located withinCpG islands, and nearly two thirds of them (N ¼ 243, 64%)showed increased methylation in Ewing sarcoma. Conversely,120 differentially methylated probes were not located withinannotated CpG islands, and the vast majority (N ¼ 98, 82%)showed reduced levels ofmethylation in Ewing sarcoma (Fig. 1B).

Promoters of ion channel encoding genes are differentiallymethylated in Ewing sarcoma

In order to better characterize the biology of the genes withdifferentiallymethylated promoters, weperformed gene ontologyanalyses to identify enriched biologic processes and molecularfunctions. To begin, we assessed enrichment of the gene ontologycategories relative to the whole genome. As expected, given thatthe array was developed to directly interrogate polycomb targets,genes with differentially methylated promoters were highlyenriched for embryonic transcription factors involved in mor-phogenesis, differentiation, and development (SupplementaryTable S2). In particular, numerous HOX genes were differentiallymethylated in Ewing sarcoma (Supplementary Table S1). This isconsistent with our recent studies, which showed that expressionof HOX genes is widely abnormal in Ewing sarcoma comparedwith adult tissues (29). In addition to the expected enrichment of

developmental transcription factors, this analysis also identifiedan unanticipated enrichment of genes with molecular functionsrelated to potassium ion binding, transport, and channel activity(Supplementary Table S2B). To determine if enrichment of thesefunctional categories was merely a reflection of the array designrather than a true enrichment, gene ontology analysis was repeat-ed and enrichment determined relative to the set of 1,204 evalu-able probes rather than the whole genome. This analysis con-firmed that biologic and molecular processes involved in ionhomeostasis were the most significantly enriched categoriesamong differentially methylated genes (Fig. 1C and D). In par-ticular, genes involved in calcium homeostasis were most prom-inently associatedwith increased promotermethylation (Fig. 1C),whereas potassium-associated geneswere overrepresented amongloci with reduced levels of DNA methylation (Fig. 1D). Thus, inaddition to the expected pattern of altered DNA methylation atdevelopmental transcription factors, this analysis also identifiedion regulatory genes as prime targets of differentialmethylation inEwing sarcoma.

The promoter of the voltage-gated potassium ion channel geneKCNA5 is DNA methylated in cancer

The flux of potassium, sodium, chloride, and calcium ionsacross the cytoplasmic membrane is regulated by a complex arrayof voltage-gated channels, and control of intracellular ion levelsby these channels is essential for cell proliferation and survival. Inparticular, voltage-gated ion channels play critical roles in themaintenance of cellular homeostasis (30, 31), and specific dereg-ulation of potassium ion channels has recently been broadlyimplicated in cancer pathogenesis (32). Twenty-two potassiumion channel genes have been reported to be differentiallyexpressed in a wide variety of human cancers, and 20 of thesewere found to be overexpressed in the context of malignantdisease (reviewed in ref. 32). In contrast, expression of KCNQ1and of Kv1.5 has been reported to be inversely correlated withclinical aggression in gastrointestinal tumors (33) and in lym-phoma and gliomas (34–36), respectively. Thus, deregulatedexpression of potassium ion channels is common in cancer, butthe mechanisms underlying their deregulation remain largely ill-defined. Significantly, we recently reported that expression of theKCNA5 gene, which encodes Kv1.5, is reduced in Ewing sarcomaand that the locus is reversibly epigenetically repressed by thepolycomb proteins BMI-1 and EZH2 (37). Therefore, we hypoth-esized thatmore permanent silencing of theKCNA5 locus byDNAmethylation may be an additional mechanism of more stablechannel suppression in Ewing sarcoma. To address this, we firstidentified potassium ion channel genes among the list of 498differentially methylated loci in Ewing sarcoma. Twenty potassi-um ion channel gene promoters were included in the list, andfourteen showed a relative reduction in DNA methylation intumors compared with nonmalignant adult tissues. Conversely,six showed an increase in DNA methylation (Fig. 2A). Amongrelatively hypomethylated loci, increased expression of channelsencoded by KCNA1 (Kv1.1), KCNC4 (Kv3.4), and KCNK2(K2P2.1) has been reported in cancer (32). Whether a relativereduction in DNA methylation at these gene promoters contri-butes to channel overexpression remains to be elucidated. Withrespect to the six loci with increased methylation in Ewingsarcoma, only downregulation of KCNA5 (Kv1.5) has beenimplicated in cancer pathogenesis (34–37), thus leading us tocontinue to focus our studies on this channel. Significantly,

Ryland et al.

Mol Cancer Res; 14(1) January 2016 Molecular Cancer Research28




KCNA5 promoter methylation was detected in both Ewing sar-coma tumors and cell lines, confirming that, in at least someEwing sarcoma cells, polycomb-dependent repressionof the locusis complemented by CpG-island DNA methylation (Fig. 2B).

To validate the data from the smaller array-based study, weanalyzed the KCNA5 locus in a larger and completely indepen-dent set of primary tumors and cells using MethyLight analysis.A methylation-specific probe was designed that allowed quan-tification of DNA methylation at the KCNA5 locus, at a site thatwas several hundred bases downstream of the GoldenGatearray probe site, but still within the promoter-associated CpG

island (Fig. 2C). This analysis confirmed that the KCNA5 locusis methylated in Ewing sarcoma cell lines and tumors comparedwith both nonmalignant fibroblasts and endothelial cells aswell as bone marrow–derived MSC, a putative cell of Ewingsarcoma origin (Fig. 2D). In contrast, the KCNA5 locus washighly methylated in NCSCs, another putative cell of tumororigin (Fig. 2D). Thus, these data suggest that differences inDNA methylation exist between stem cells of different devel-opmental origins. It is noteworthy that, compared with theGoldenGate array probe, the MethyLight probe detected moreabundant DNAmethylation in the cell lines than in the primary

Figure 1.The abnormal DNA methylation profile of polycomb target genes in Ewing sarcoma. A, analysis of 52 normal adult tissues, 5 Ewing sarcoma cell lines, and4 primary Ewing sarcoma tumors using a custom GoldenGate bead array demonstrates differences in the methylation profile of polycomb target gene promotersin Ewing sarcoma samples compared with normal adult tissue. B, in Ewing sarcoma, increased methylation is more prevalent in promoters with CpG islandsthan in promoters that do not contain CpG islands. ��, P < 0.0001 by the Fisher exact test. C and D, gene ontology analysis of differentially methylatedpromoters identifies calcium homeostasis genes as selective targets of increased promoter methylation (C) and potassium ion channel genes as beingenriched among genes with decreased promoter methylation (D). Enrichment was determined relative to the set of 1,204 probes on the array and isexpressed as the negative log of the enrichment score (P value).






tumors. This suggests that broader methylation of locus mightbe an adaptive response of cells to prolonged in vitro culture, orthat Ewing sarcoma cells with more extensive DNAmethylationof the KCNA5 gene are more amenable to generating a cell line(Fig. 2D). However, given that MSCs, fibroblasts, and endo-thelial cell lines showed no evidence of significant DNA meth-ylation, it is clear that cell culture alone cannot explainincreased methylation of the KCNA5 locus in Ewing sarcomacells. In support of this, data from the MethHC database (38)shows significant tumor-specific KCNA5 promoter DNA meth-ylation in all of the 18 different tumor types analyzed (Sup-plementary Fig. S1). Thus, DNA methylation at the KCNA5promoter is a common attribute of multiple tumor types in vivo,thereby providing a potential mechanism to explain priorobservations that Kv1.5 is downregulated in cancer.

Decitabine induces loss of DNA methylation at the KCNA5locus and an increase in gene expression

Having established that KCNA5 is methylated in Ewing sarco-ma cells, we next sought to determine if exposing cells to thehypomethylating agent, decitabine, would reverse DNA methyl-ation and lead to increased gene expression. At high doses,decitabine is cytotoxic, but at low doses, it prevents DNA meth-ylation without inducing cell death, and the efficacy of decitabineas an epigenetic therapy is best achieved with subcytotoxic doses,which in vitro usually range from20 to 300 nmol/L (39). Exposureof Ewing sarcoma cells to 100 nmol/L decitabine did not inducecell death and led to reduced methylation of KCNA5 in all threeEwing sarcoma cell lines (Fig. 3A). Consistent with DNA meth-ylation being a mechanism of gene repression, we also observed

Figure 2.DNA methylation of KCNA5 in Ewing sarcoma. A, DNA methylation state(expressed as beta values) of differentiallymethylated potassium ion channelgenes, as determined by a custom Illumina GoldenGate DNA methylationarray. All twenty genes were significantly differentially methylated in Ewingsarcoma compared with normal adult tissues (median difference in betavalues � 0.2 and corrected P value < 0.05). B, KCNA5 promoter DNAmethylation (�� , P < 0.001). C, site of MethyLight primers and probe forevaluation the KCNA5 locus. D, MethyLight analysis of 6 Ewing sarcomacell lines, 21 Ewing sarcoma primary tumors, 2 nontransformed cell lines(HuVEC, lung fibroblasts: MRC5), MSC, and NCSCs. Data points arepresented as mean values of replicate experiments. MethyLight data arepresented as PMR. �� , P < 0.005 and �� , P < 0.001 by the Student t test(mean � SEM).

Figure 3.Demethylation of the KCNA5 locus is associated with increased transcriptexpression. A, MethyLight analysis of the KCNA5 locus in Ewing sarcoma cellsafter treatment with 100 nmol/L decitabine. Expression is presented as PMRrelative to untreated cells. B, site of PCR primers for evaluation of KCNA5transcript expression. C, qRT-PCR of KCNA5 transcript expression in Ewingsarcoma cell lines after a 72- or 96-hour treatment with 100 nmol/Ldecitabine. Expression is normalized to the geometric mean of HPRT andGAPDH within each sample and fold change in decitabine relative to thevehicle-treated cells. � , P < 0.05 by the Student t test (mean � SEM, n ¼ 3).

Ryland et al.





that expression of the KCNA5 transcript increased concomitantlywith loss of DNAmethylation (Fig. 3C). However, the correlationbetween loss of methylation and derepression of gene expressionwas not linear over time, indicating that other mechanisms ofgene regulation also contribute to modulation of KCNA5 expres-sion in these cells.

Kv1.5 channel function inhibits Ewing sarcoma cellproliferation

Wenext evaluated whether DNAmethylation of KCNA5 affectsthe proliferative phenotype of Ewing sarcoma cells. Exposure ofEwing sarcoma cells to low-dose decitabine for 72 hours had aprofound impact on cell expansion (Fig. 4A). Trypan-blue exclu-sion displaying either total live cell count (Fig. 4B) or total deadcell count (Fig. 4C) and EdU incorporation (Fig. 4D) wereperformed to determine whether this effect was due to an increasein cell death or a decrease in proliferation. As shown, no signif-icant loss of viability was observed (Fig. 4C). In contrast, total livecell count was significantly reduced (Fig. 4B), and incorporationof EdU was reduced (loss of 2nd peak in histogram; Fig. 4D),revealing that decitabine inhibits cell-cycle progression and pro-liferation of Ewing sarcoma cells. To determine if reduced pro-liferation in decitabine-treated cells was a result of loss of KCNA5repression, we used the pharmacologic compounds 40AP andDPO-1 (Fig. 5A). These two compounds are highly specificinhibitors of the Kv1.5 channel when used at pharmacologicallyvalidated doses (40–43). Notably, they function to block theKv1.5 channel function by inhibiting the flux of potassium ionsacross the channel pore and altering transmembrane current,without altering the levels of channel expression. Significantly,

blocking the Kv1.5 channel with either agent partially restored theproliferative phenotype of decitabine-treated Ewing sarcoma cells(Fig. 5B and C). Together, these data reveal that epigeneticsilencing of the Kv1.5 channel, by DNA methylation of KCNA5,contributes to Ewing sarcoma cell proliferation and that this effectis mediated, at least in part, by the function of Kv1.5 as a negativeregulator of cell-cycle progression (Fig. 6).

DiscussionCancer cells hijack and dysregulate epigenetic mechanisms to

dynamically alter gene expression and drive tumor pathogenesis.In the current studies, we assessed theDNAmethylation profile ofpolycomb target gene promoters in the context of Ewing sarcoma.From these studies, we discovered that, like other human cancers,CpG islands in the promoters of polycomb target genes areselectively methylated in Ewing sarcoma. However, we also dis-covered the unanticipated finding that genes involved in calciumand potassium homeostasis are also differentially methylatedrelative to nonmalignant adult tissues. More specifically, ourstudies identified potassium ion channels as frequent targets ofaberrant DNA methylation in Ewing sarcoma and revealed thatthe promoter of KCNA5 is relatively hypermethylated in tumorcells. In addition, interrogation of public databases showed thatthe KCNA5 locus is reproducibly hypermethylated, relative toadjacent normal tissues, in a wide variety of human cancers (38).We recently showed that polycomb-dependent repression ofKCNA5 contributes to Ewing sarcoma cell survival under condi-tions of hypoxia (37). Thus, the current study validates our priorobservation that the Kv1.5 channel suppression contributes to the

Figure 4.Decitabine inhibits Ewing sarcoma cellproliferation. A, Brightfield images ofthree Ewing sarcoma cell lines withvehicle or 100 nmol/L decitabine. B,trypan-blue–based proliferationassays determined TC-71, A4573, andA673 cells have a significant reductionin cell expansion after a 72-hourtreatmentwith 100nmol/L decitabine.� , P < 0.05 by the Student t test (mean� SEM, n ¼ 3–5). C, decitabinetreatment has no effect on cell deathas compared with vehicle control, asdetermined by direct counting oftrypan-blue–stained cells. D,fluorescence signal from three Ewingsarcoma cells exposed to 10 mmol/LEdU for 2 hours following 72-hourtreatment with vehicle or 100 nmol/Ldecitabine. The graphs show a clearseparation of nonproliferating andproliferating cells in vehicle-treatedcells, whereas the decitabine-treatedcells lack a clear separation due to areduction in or complete loss ofproliferating cells.






cancer phenotype and provides new evidence that cancer cellproliferation in ambient conditions is promoted, in part, by stableDNA methylation–dependent repression of the KCNA5 locus.

Potassium channels are transmembrane proteins that controlcellular ion concentrations and homeostasis. Given that cellproliferation and survival are intimately linked to intracellularpotassium ion levels (44–46), dysregulation of potassium chan-nels can affect the cancer phenotype. Consistent with this, aber-rant expression of ion channels is prevalent in cancer (47, 48).Potassium-conducting channels also regulate the biophysicalproperties of a cell, controlling intracellular potassium concen-trations andmaintaining resting membrane potential (31). Thus,hijacking of potassium ion channel function by cancer cells hasthe potential to affect cell proliferation, cell survival, migration,and differentiation (32). Interestingly, however, although mostpotassium channels have been found to be overexpressed incancer, Kv1.5 is one of only two potassium channels that isdownregulated (32).

Knowledge regarding the physiologic functions of Kv1.5 innormal cell biology provides insights into why this channel inparticular would be selectively silenced in cancer. In myoblasts,high Kv1.5 expression leads to accumulation of cyclin-dependentkinase inhibitors and impairs cell-cycle progression at the G1 to Stransition (49), suggesting the channel is a cell-cycle checkpointregulator and could function as a tumor-suppressor gene. Ourstudies begin to illuminate the potential importance of thisfunction in the context of Ewing sarcoma. Specifically, we havefound thatDNAmethylation and repressionofKCNA5 contributeto cell-cycle progression and that reversion of promoter methyl-ation and KCNA5 de-repression is associated with growth inhi-bition and cell-cycle arrest. Moreover, specific pharmacologicinhibition of the Kv1.5 channel function partially restores pro-liferation in Ewing sarcoma cells that have been exposed todecitabine. Thus, these studies together provide evidence that

reactivation of silenced Kv1.5 channels can inhibit cancer prolif-eration, likely by permitting efflux of Kþ ions and inhibiting G1 toS transition. Alternatively, emerging data suggest that there are

Figure 6.Overview of the epigenetic repression of KCNA5 in Ewing sarcoma cancercells. Survival of Ewing sarcoma cells in hypoxia and maintenance of aproliferative state are supported by epigenetic suppression of the KCNA5locus. The PcG proteins, BMI-1 and EZH2, and their respective histonemodifications (H2AK119Ub1 and H3K27me3), along with DNA methylation,serve to reversibly and more stably silence the locus, respectively. Stablesilencing of the KCNA5 locus by promoter DNA methylation contributes toEwing sarcoma cell proliferation. The Kv1.5 channel has been shown tofunction as a G1–S cell-cycle checkpoint regulator. Therefore, silencing of thislocus contributes to the ability of Ewing sarcoma cells to bypass cell-cyclecheckpoints and maintain a proliferative phenotype.

Figure 5.Blocking the Kv1.5 channel functionpartially restores proliferativephenotype to decitabine-treatedEwing sarcoma cells. A, diagramdemonstrating the open channelblockers, DPO-1 and 40AP. Bothcompounds prevent Kþ

flux out of theKv1.5 channel. Blocking the Kv1.5channel function with 50 mmol/L 40AP(B) or 310 nmol/L DPO-1 (C) indecitabine-treated cells partiallyrestores Ewing sarcoma cellproliferation. � , P < 0.05 by theStudent t test (mean� SEM, n¼ 3–5).

Ryland et al.





other functions for potassium channels that do not depend ontheir roles as ion conductors (32, 50). The known effect of both4'AP and DPO-1 is to block Kþ efflux and alter transmembranecurrent. Whether noncanonical roles exist for Kv1.5 outside ofregulating Kþ

flux andmembrane depolarization exist andwheth-er they contribute to its function as a tumor suppressor remain tobe elucidated.

The contribution of ion channel deregulation to tumor path-ogenesis remains a relatively unexplored area of investigation incancer biology. We have now shown that the Kv1.5 channel isreproducibly targeted for epigenetic repression, by both poly-comb proteins (37) and by DNA methylation (this study), andthat repression contributes to cancer cell survival and prolifera-tion. These findings provide compelling evidence to support thedesignation of KCNA5 as a tumor-suppressor gene in humancancers, and suggest that further studies into its mechanism ofaction as a cell-cycle regulator will reveal how loss of the Kv1.5function supports the proliferative cancer phenotype.

Disclosure of Potential Conflicts of InterestD.J. Weisenberger is consultant at Zymo Research Corporation; has owner-

ship interest (including patents) in USC patent royalty payment; and is aconsultant/advisory boardmember for Zymo Research Corporation. P.W. Lairdhas received honoraria from the Speakers Bureau of Merck SD. No potentialconflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: K.E. Ryland, S.C. Borinstein, J.R. Martens, E.R. LawlorDevelopment of methodology: K.E. Ryland, A.G. Hawkins, D.J. Weisenberger,S.C. Borinstein

Acquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): D.J. Weisenberger, S.C. Borinstein, P.W. Laird,E.R. LawlorAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): K.E. Ryland, V. Punj, S.C. Borinstein, P.W. Laird,E.R. LawlorWriting, review, and/or revisionof themanuscript:K.E. Ryland, A.G.Hawkins,D.J. Weisenberger, S.C. Borinstein, P.W. Laird, J.R. Martens, E.R. LawlorAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): K.E. Ryland, E.R. LawlorStudy supervision: J.R. Martens, E.R. Lawlor

AcknowledgmentsThe authors thank members of the Lawlor and Martens laboratories for

helpful discussion and technical support.

Grant SupportK.E. Ryland received support from Pharmacological Sciences Training Pro-

gram T32 GM007767. A.G. Hawkins received support from the UM CancerBiology Training Grant 5T32 CA009676-20. J.R. Martens was supported by anNIH grant number, R01HL070973. E.R. Lawlor was supported by an NIH grantnumber, 1R01 CA134604, a Stand Up to Cancer Innovative Research Grant(grant number SU2C-AACR-IRG-1309), and CureSearch Children's OncologyGroup and Nick Currey Fund. Stand Up To Cancer is a program of theEntertainment Industry Foundation administered by the American Associationfor Cancer Research. S.C. Borinstein received support from the Rally Foundationfor Cancer Research, the St. Baldrick's Foundation, and Hyundai Hope onWheels.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received August 11, 2015; revised October 28, 2015; accepted November 4,2015; published OnlineFirst November 16, 2015.

References1. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell

2011;144:646–74.2. Liu Y, Vlatkovic L, Saeter T, Servoll E, Waaler G, Nesland JM, et al. Is the

clinical malignant phenotype of prostate cancer a result of a highlyproliferative immune-evasive B7-H3-expressing cell population? Int J Urol2012;19:749–56.

3. Bianchini G, Pusztai L, Karn T, Iwamoto T, Rody A, Kelly C, et al. Prolif-eration and estrogen signaling can distinguish patients at risk for earlyversus late relapse among estrogen receptor positive breast cancers. BreastCancer Res 2013;15:R86.

4. Barboza N, Minakhina S, Medina DJ, Balsara B, Greenwood S, Huzzy L,et al. PDCD2 functions in cancer cell proliferation and predicts relapsedleukemia. Cancer Biol Ther 2013;14:546–55.

5. Jones PA, Baylin SB. The epigenomics of cancer. Cell 2007;128:683–92.6. Baylin SB, Jones PA. A decade of exploring the cancer epigenome -

biological and translational implications. Nat Rev Cancer 2011;11:726–34.

7. Laugesen A, Helin K. Chromatin repressive complexes in stem cells,development, and cancer. Cell Stem Cell 2014;14:735–51.

8. Lee TI, Jenner RG, Boyer LA, Guenther MG, Levine SS, Kumar RM, et al.Control of developmental regulators by Polycomb in human embryonicstem cells. Cell 2006;125:301–13.

9. Ohm JE, McGarvey KM, Yu X, Cheng L, Schuebel KE, Cope L, et al. A stemcell-like chromatin pattern may predispose tumor suppressor genes toDNA hypermethylation and heritable silencing. Nat Genet 2007;39:237–42.

10. Schlesinger Y, Straussman R, Keshet I, Farkash S, Hecht M, Zimmerman J,et al. Polycomb-mediated methylation on Lys27 of histone H3 pre-marksgenes for de novo methylation in cancer. Nat Genet 2007;39:232–6.

11. Widschwendter M, Fiegl H, Egle D, Mueller-Holzner E, Spizzo G, Marth C,et al. Epigenetic stem cell signature in cancer. Nat Genet 2007;39:157–8.

12. BalamuthNJ,Womer RB. Ewing's sarcoma. LancetOncol 2010;11:184–92.

13. Tirode F, Surdez D, Ma X, Parker M, Le Deley MC, Bahrami A, et al.Genomic landscape of Ewing sarcoma defines an aggressive subtypewith co-association of STAG2 and TP53 mutations. Cancer Discov 2014;4:1342–53.

14. Crompton BD, Stewart C, Taylor-Weiner A, Alexe G, Kurek KC, CalicchioML, et al. The genomic landscape of pediatric Ewing sarcoma. CancerDiscov 2014;4:1326–41.

15. Brohl AS, Solomon DA, Chang W, Wang J, Song Y, Sindiri S, et al. Thegenomic landscape of the Ewing Sarcoma family of tumors reveals recur-rent STAG2 mutation. PLoS Genet 2014;10:e1004475.

16. Tomazou EM, Sheffield NC, Schmidl C, Schuster M, Schonegger A, Datlin-ger P, et al. Epigenome mapping reveals distinct modes of gene regulationand widespread enhancer reprogramming by the oncogenic fusion proteinEWS-FLI1. Cell Rep 2015;10:1082–95.

17. Riggi N, Knoechel B, Gillespie SM, Rheinbay E, Boulay G, Suva ML, et al.EWS-FLI1 utilizes divergent chromatin remodelingmechanisms to directlyactivate or repress enhancer elements in Ewing sarcoma. Cancer Cell2014;26:668–81.

18. Douglas D, Hsu JH, Hung L, Cooper A, Abdueva D, van Doorninck J, et al.BMI-1 promotes Ewing sarcoma tumorigenicity independent of CDKN2Arepression. Cancer Res 2008;68:6507–15.

19. Hsu JH, Lawlor ER. BMI-1 suppresses contact inhibition and stabilizes YAPin Ewing sarcoma. Oncogene 2011;30:2077–85.

20. Richter GH, Plehm S, Fasan A, Rossler S, Unland R, Bennani-Baiti IM, et al.EZH2 is a mediator of EWS/FLI1 driven tumor growth and metastasisblocking endothelial and neuro-ectodermal differentiation. ProcNatl AcadSci U S A 2009;106:5324–9.

21. Riggi N, Suva M-L, Suva D, Cironi L, Provero P, Tercier S, et al. EWS-FLI-1expression triggers a Ewing's sarcoma initiation program in primaryhuman mesenchymal stem cells. Cancer Res 2008;68:2176–85.

22. Hurtubise A, Bernstein ML, Momparler RL. Preclinical evaluation of theantineoplastic action of 5-aza-20-deoxycytidine and different histone






deacetylase inhibitors on human Ewing's sarcoma cells. Cancer Cell Int2008;8:16.

23. Park HR, Jung WW, Kim HS, Park YK. Microarray-based DNAmethylationstudy of Ewing's sarcoma of the bone. Oncol Lett 2014;8:1613–7.

24. von LevetzowC, Jiang X, Gwye Y, von LevetzowG, Hung L, Cooper A, et al.Modeling initiation of Ewing sarcoma in human neural crest cells. PLoSOne 2011;6:e19305.

25. Byun HM, Siegmund KD, Pan F,Weisenberger DJ, Kanel G, Laird PW, et al.Epigenetic profiling of somatic tissues from human autopsy specimensidentifies tissue- and individual-specific DNA methylation patterns. HumMol Genet 2009;18:4808–17.

26. Huang da W, Sherman BT, Zheng X, Yang J, Imamichi T, Stephens R, et al.Extracting biological meaning from large gene lists with DAVID. CurrProtoc Bioinformatics 2009;Chapter 13:Unit 13 1.

27. CampanM,WeisenbergerDJ, Trinh B, Laird PW.MethyLight.MethodsMolBiol 2009;507:325–37.

28. Lang F, RitterM,GamperN,Huber S, Fillon S, Tanneur V, et al. Cell volumein the regulation of cell proliferation and apoptotic cell death. Cell PhysiolBiochem 2000;10:417–28.

29. Svoboda LK, Harris A, Bailey NJ, Schwentner R, Tomazou E, von LevetzowC, et al. Overexpression of HOX genes is prevalent in Ewing sarcoma and isassociated with altered epigenetic regulation of developmental transcrip-tion programs. Epigenetics 2014;9:1613–25.

30. Rao VR, Perez-Neut M, Kaja S, Gentile S. Voltage-gated ion channels incancer cell proliferation. Cancers 2015;7:849–75.

31. Yang M, Brackenbury WJ. Membrane potential and cancer progression.Front Physiol 2013;4:185.

32. Huang X, Jan LY. Targeting potassium channels in cancer. J Cell Biol2014;206:151–62.

33. Than BL, Goos JA, Sarver AL, O'SullivanMG, Rod A, Starr TK, et al. The roleof KCNQ1 in mouse and human gastrointestinal cancers. Oncogene2014;33:3861–8.

34. Arvind S, Arivazhagan A, Santosh V, Chandramouli BA. Differentialexpression of a novel voltage gated potassium channel–Kv 1.5 in astrocy-tomas and its impact on prognosis in glioblastoma. Br J Neurosurg 2012;26:16–20.

35. Bielanska J, Hernandez-Losa J, Perez-Verdaguer M, Moline T, Somoza R,Ramon y Cajal S, et al. Voltage-dependent potassium channels Kv1.3 andKv1.5 in human cancer. Curr Cancer Drug Targets 2008;9:904–14.

36. Preussat K, Beetz C, Schrey M, Kraft R, Wolfl S, Kalff R, et al. Expression ofvoltage-gated potassium channels Kv1.3 and Kv1.5 in human gliomas.Neurosci Lett 2003;346:33–6.

37. Ryland KE, Svoboda LK, Vesely ED, McIntyre JC, Zhang L, Martens JR, et al.Polycomb-dependent repression of the potassium channel-encoding geneKCNA5promotes cancer cell survival under conditions of stress. Oncogene2015;34:4591–600.

38. Huang WY, Hsu SD, Huang HY, Sun YM, Chou CH, Weng SL, et al.MethHC: a database of DNA methylation and gene expression in humancancer. Nucleic Acids Res 2015;43(Database issue):D856–61.

39. Tsai HC, Li H, Van Neste L, Cai Y, Robert C, Rassool FV, et al. Transientlow doses of DNA demethylating agents exert durable anti-tumor effectson hematological and epithelial tumor cells. Cancer Cell 2012;21:430–46.

40. Stump G, Wallace A, Regan C, Lynch J. In vivo antiarrhythmic and cardiacelectrophysiologic effects of a novel diphenylphosphine oxide IKur blocker(2-Isopropyl-5-methylcyclohexyl) diphenylphosphine oxide. J PharmacolExp Ther 2005;315:1362–7.

41. Fedida D, Wible B, Wang Z, Fermini B, Faust F, Nattel S, et al. Identity of anovel delayed rectifier current fromhumanheart with a cloned Kþ channelcurrent. Circ Res 1993;73:210–6.

42. Lagrutta A, Wang J, Fermini B, Salata JJ. Novel, potent inhibitors of humanKv1.5 Kþ channels and ultrarapidly activating delayed rectifier potassiumcurrent. J Pharmacol Exp Ther 2006;317:1054–63.

43. Yamane T, Furukawa T, Hiraoka M. 4-Aminopyridine block of the non-inactivating cloned Kþ channel Kv1.5 expressed in Xenopus oocytes. Am JPhysiol 1995;269(2 Pt 2):H556–64.

44. Burg ED, Remillard CV, Yuan JXJ. Kþ channels in apoptosis. J MembraneBiol 2006;209:3–20.

45. Fiske JL, Fomin VP, BrownML, Duncan RL, Sikes RA. Voltage-sensitive ionchannels and cancer. Cancer Metastasis Rev 2006;25:493–500.

46. Urrego D, Tomczak AP, Zahed F, Stuhmer W, Pardo LA. Potassiumchannels in cell cycle and cell proliferation. Philos Trans R Soc Lond BBiol Sci 2014;369:20130094.

47. Huber SM. Oncochannels. Cell Calcium 2013;53:241–55.48. Pedersen SF, Stock C. Ion channels and transporters in cancer: patho-

physiology, regulation, and clinical potential. Cancer Res 2013;73:1658–61.

49. Villalonga N, Martinez-Marmol R, Roura-Ferrer M, David M, Valen-zuela C, Soler C, et al. Cell cycle-dependent expression of Kv1.5 isinvolved in myoblast proliferation. Biochim Biophys Acta 2008;1783:728–36.

50. Sun XX, Bostrom SL, Griffith LC. Alternative splicing of the eag potassiumchannel gene in Drosophila generates a novel signal transduction scaf-folding protein. Mol Cell Neurosci 2009;40:338–43.


Ryland et al.




2016;14:26-34. Published OnlineFirst November 16, 2015.Mol Cancer Res Katherine E. Ryland, Allegra G. Hawkins, Daniel J. Weisenberger, et al. Silencing Supports Ewing Sarcoma Cell Proliferation

Ion ChannelKCNA5Promoter Methylation Analysis Reveals That

Updated version

10.1158/1541-7786.MCR-15-0343doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2015/11/14/1541-7786.MCR-15-0343.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/14/1/26.full#ref-list-1

This article cites 49 articles, 10 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/14/1/26.full#related-urls

This article has been cited by 1 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://mcr.aacrjournals.org/content/14/1/26To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-15-0343

http://mcr.aacrjournals.org/content/suppl/2015/11/14/1541-7786.MCR-15-0343.DC1

http://mcr.aacrjournals.org/content/14/1/26.full#ref-list-1

http://mcr.aacrjournals.org/content/14/1/26.full#related-urls

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/14/1/26


promoter methylation analysis reveals that kcna5 ion channel silencing supports ewing ... ·...

Documents