WHSC1 Promotes Oncogenesis through Regulation of NIMA ... · Regulation of NIMA-Related Kinase-7 in Squamous Cell Carcinoma of the Head and Neck Vassiliki Saloura1, Hyun-Soo Cho1,2,

Chromatin, Gene, and RNA Regulation

WHSC1 Promotes Oncogenesis throughRegulation of NIMA-Related Kinase-7 inSquamous Cell Carcinoma of the Head and NeckVassiliki Saloura1, Hyun-Soo Cho1,2, Kazuma Kiyotani1, Houda Alachkar1, Zhixiang Zuo1,Makoto Nakakido1, Tatsuhiko Tsunoda3, Tanguy Seiwert1, Mark Lingen1, Jonathan Licht4,Yusuke Nakamura1, and Ryuji Hamamoto1

Abstract

Squamous cell carcinoma of the head and neck (SCCHN) is arelatively common malignancy with suboptimal long-term prog-nosis, thus new treatment strategies are urgently needed. Over thelast decade, histone methyltransferases (HMT) have been recog-nized as promising targets for cancer therapy, but their mecha-nism of action in most solid tumors, including SCCHN, remainsto be elucidated. This study investigated the role of Wolf–Hirsch-horn syndrome candidate 1 (WHSC1), an NSD family HMT, inSCCHN. Immunohistochemical analysis of locoregionallyadvanced SCCHN, dysplastic, and normal epithelial tissue speci-mens revealed that WHSC1 expression and dimethylation ofhistone H3 lysine 36 (H3K36me2) were significantly higher inSCCHN tissues than in normal epithelium. Both WHSC1 expres-sion and H3K36me2 levels were significantly correlated withhistologic grade. WHSC1 knockdown in multiple SCCHN cell

lines resulted in significant growth suppression, induction ofapoptosis, and delay of the cell-cycle progression. Immunoblotand immunocytochemical analyses in SCCHN cells demonstrat-ed that WHSC1 induced H3K36me2 and H3K36me3. Microarrayexpression profile analysis revealed NIMA-related kinase-7(NEK7) to be a downstream target gene of WHSC1, and chro-matin immunoprecipitation (ChIP) assays showed that NEK7was directly regulated by WHSC1 through H3K36me2. Further-more, similar toWHSC1, NEK7 knockdown significantly reducedcell-cycle progression, indicating that NEK7 is a key player in themolecular pathway regulated by WHSC1.

Implications: WHSC1 possesses oncogenic functions in SCCHNand represents a potential molecular target for the treatment ofSCCHN. Mol Cancer Res; 13(2); 293–304. �2014 AACR.

IntroductionOver the last decade, epigenetic regulators have been implicated

as key factors in many pathways relevant to cancer developmentand progression, such as cell-cycle regulation (1–3), invasiveness(4), signaling pathways (5), chemoresistance (6) and immuneevasion (7). The three basic systems of epigenetic regulation areDNA methylation of gene regulatory regions, histone proteinmodifications, including acetylation,methylation, ubiquitination,phosphorylation and sumoylation, andnoncodingRNAs.Histonemethylation is dynamically regulated by two different types ofenzymes called histone methyltransferases and histone demethy-

lases. At present, approximately 50different histone lysinemethyl-transferases (HKMT; ref. 8), 10histone argininemethyltransferases(HRMT; ref. 8), and 30 histone demethylases (HDMT; ref. 9) havebeen identified, but their biologic functions are not fully charac-terized. However, because of their frequent overexpression and/orsomatic mutations in a variety of cancer types, extensive efforts forthe development of drugs targeting these enzymes have beeninitiated over the past several years (10, 11). In this regard, animportant group of HKMTs are the nuclear receptor Suppressor ofvariegation 3–9 Su(var)3–9, Enhancer of zeste and Trithorax(SET)-domain-containing (NSD) family members of HKMTs(NSD-HKMTs) NSD1, NSD2/WHSC1/MMSET, and NSD3/WHSC1L1, which modulate the expression of genes throughmethylation of lysine 36 on histone H3 (12). These HKMTs share70% to 75%homology in their amino acid sequences and contain4 basic domains which are also conserved in other development-associated proteins: a Pro-Trp-Trp-Pro (PWWP) motif, which is aDNA methyl-lysine and methyl-arginine histone binder, a planthomeodomain PHD-type zinc finger (Cys3-His-Cys4) with meth-yl-lysine binding affinity, a high-mobility-group (HMG) boxwhichhasDNA-binding capacity, anda SETdomain that possessesthe methyltransferase activity (13).

We previously reported thatWHSC1 is overexpressed in varioustypes of human cancer compared with nonneoplastic tissues. Wealso showed that knockdown of WHSC1 in bladder and lungcancer cell lines significantly suppressed the growth of cancercells through cell-cycle arrest and that WHSC1 regulated the Wnt

1SectionofHematologyandOncology,UniversityofChicago,Chicago,Illinois. 2Genomics Structure Research Center, Korea Research Insti-tute of Bioscience and Biotechnology, Daejeon, 305-333, Republic ofKorea. 3Laboratory for Medical Science Mathematics, RIKEN Centerfor Integrative Medical Sciences, Tsurumi-ku, Yokohama, Kanagawa,Japan. 4Section of Hematology and Oncology, Northwestern Univer-sity, Chicago, Illinois.

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

Corresponding Author: Ryuji Hamamoto, University of Chicago, 5841 S. Mary-land Avenue, MC2115, Chicago, IL 60637. Phone: 773702-0933; Fax: 773-702-9385; E-mail: [email protected]

doi: 10.1158/1541-7786.MCR-14-0292-T

�2014 American Association for Cancer Research.

MolecularCancerResearch

www.aacrjournals.org 293

on August 30, 2020. © 2015 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst October 3, 2014; DOI: 10.1158/1541-7786.MCR-14-0292-T

http://mcr.aacrjournals.org/

pathway through interactionwithb-catenin (14).Hudlebusch andcolleagues also examined theexpressionofWHSC1 in3,774cancertissues and 904 corresponding normal tissues and found signifi-cantly higher expression in a variety of cancers (15). Stec andcolleaguesdescribed thehaploinsufficiencyofWHSC1asa causeofWolf–Hirschhorn syndrome, a growth developmental disorder,and reported the chromosomal translocation t(4;14)(p16.3;q32.3) of theWHSC1 gene and the immunoglobulin heavy-chain(IgH) promoter in multiple myeloma that led to significantoverexpression of WHSC1 (16). Moreover, it has been shownthat knockdown of WHSC1 in multiple myeloma cell linesremarkably suppresses growth and regulates apoptosis, cell cycle,invasion, and DNA repair pathways (17). Martinez-Garcia andcolleagues showed thatWHSC1-mediatedoncogenesis inmultiplemyeloma is related to activation of gene expression of clusters ofgenes through increased dimethylation of lysine 36 on histoneH3(H3K36me2; ref. 18). Kuo and colleagues further reported that themain chromatin-modifying end product ofWHSC1 is H3K36me2and showed that this was sufficient for activation of oncogenicprogramming favoring myelomagenesis (19). Furthermore,WHSC1 has been reported to facilitate p53-binding protein 1recruitment in the process of DNA damage repair, suggesting thatoverexpression ofWHSC1maypotentially contribute to resistanceof cancer cells to DNA-damaging chemotherapy agents (20).

Despite various advances in treatment, long-term survival forpatients with squamous cell carcinoma of the head and neck(SCCHN) remains suboptimal, hence new therapeutic optionsare urgently required. Epigenetic dysregulation has mainly beenstudied as a potential mechanism of progression from dysplasiato SCCHN through aberrant DNA promoter methylation ofgenes involved in carcinogenesis (21, 22). Few studies thoughhave explored the role of histone modifications in the patho-genesis of SCCHN. The Cancer Genome Atlas (TCGA) projectrecently reported that the NSD family of HKMTs is altered in29% of patients with SCCHN in a mutually exclusive pattern,with 9% of patients having recurrent amplifications inWHSC1L1 (8p11.23) and 10% having recurrent mutations inNSD1, implying that alterations in these genes may function asdriver events in the oncogenesis of SCCHN (23). Frequent andmutually exclusive genetic alterations in the NSD-HKMTswere also found in lung squamous cell carcinoma (28%) andbreast invasive carcinoma (18%; ref. 23). On the basis of thesedata, we decided to further investigate WHSC1 as a potentialnovel therapeutic target for SCCHN. In the present study, wedemonstrate that WHSC1 is significantly overexpressed inSCCHN, its knockdown causes cell death through apoptosisand that it promotes cell-cycle progression through activationof NEK7, indicating its potential role as an oncogenic force inSCCHN.

Materials and MethodsImmunohistochemistry and head and neck cancer tissuemicroarrays

The expression pattern of WHSC1 in 149 SCCHN, 19 dysplas-tic, and 18 normal epithelial tissue sections were examined byimmunohistochemistry (IHC). SCCHN sections were derivedfrom biopsies of patients with local or locoregionally advanceddisease previous to treatment with either surgery with or withoutadjuvant chemoradiation or definitive chemoradiation. Slides ofparaffin-embedded squamous cell carcinoma tumor specimens,

dysplastic, and normal epithelial tissues were deparaffinized,rehydrated, and sections were treated with antigen retrievalbuffer (pH 6; DAKO, S2367) in a steamer for 20 minutes at96�C. Anti-WHSC1 (Abcam, ab75359, dilution: 1:400) andanti-H3K36me2 (Cell Signaling Technology, #2901, dilution:1:100) antibodies were applied on tissue sections for 1-hourincubation at room temperature (RT), followed by detectionusing the Bond Refine system (Leica). Following TBS wash, theantigen–antibody binding was detected with the Bond Refinepolymer detection system (DS9800, Leica Biosystems) and theDABþ chromogen (DAKO, K3468). Tissue sections were brieflyimmersed in hematoxylin for counterstaining and were coveredwith cover glass slides. An expert head and neck cancer pathol-ogist performed semiquantitative analysis of WHSC1 andH3K36me2 staining using a 4-grade scale defined as follows:negative, grade 0; mild, grade þ1; moderate, grade þ2; andstrong staining intensity, grade þ3. Use of tissues was approvedby the University of Chicago Institutional Review Board (IRB12–2125 and IRB 12–2117).

Cell cultureSquamous cell carcinoma cell lines UD-SCC2, SCC-23, SCC-

25, SCC-35, UT-SCC-40, HN-SCC-135, HN-SCC-151, HN-SCC-166, PE/CA-PJ15, OECM-1, BICR31, 93VU147T, FaDu, JSQ-3,HN-5, and HN-6 were derived from patients with locoregionallyadvanced SCCHN and were kindly provided by Dr. TanguySeiwert (The University of Chicago, Chicago, IL). Detailed char-acteristics of each cell line are shown in Supplementary Table S1.UD-SCC-2, SCC-23, SCC-25, SCC-35, HN-SCC-135, HN-SCC-151, HN-SCC-166, and JSQ-3 were maintained in DMEM/F12medium, 10% FBS, 1% penicillin/streptomycin, and 2 nmol/LL-glutamine. UT-SCC-40 cells were maintained in Eagle's medi-um, 10% FBS, 1% penicillin/streptomycin, and 2 nmol/LL-glutamine. PE/CA-PJ15 was maintained in IDMEM, 10% FBS,1% penicillin/streptomycin, and 2 nmol/L L-glutamine. HN-5,HN-6, BICR31, and 93VU147T cells were maintained in DMEMwith 10% FBS, 1% penicillin/streptomycin, and 2 nmol/LL-glutamine. OECM-1 and FaDu cells were maintained in RPMImedium, 10% FBS, 1% penicillin/streptomycin, and 2 nmol/LL-glutamine. KGM cells (normal human keratinocytes, Lonza,00192627) were maintained in KGM-Gold keratinocyte growthmedium supplemented by BPE, transferrin, insulin, hEGF, hydro-cortisone, epinephrine, transferrin, and gentamicin/amphotericinB (KGM-Gold Bullet kit 00192060).

Expression vector constructionAn entire coding sequence of WHSC1 (GenBank Accession:

BC166668) was amplified from human testis cDNAs using KOD-Plus-Neo (TOYOBO, KOD-401) DNA polymerase and clonedinto pCAGGSn3FC vector between NotI and XhoI restrictionenzyme sites (pCAGGS-WT-WHSC1-FLAG). To prepare anenzyme-inactive WHSC1, the coding sequence of SET domainwas deleted from the entire coding sequence of WHSC1(pCAGGS-WHSC1-DSET-FLAG).

Quantitative real-time PCRSpecific primers for humanGAPDH (housekeeping gene), SDH

(housekeeping gene), WHSC1, NEK7, MAPK8, and HIPK3 weredesigned (primer sequences in Supplementary Table S2). PCRreactions were performed using ViiA 7 real-time PCR system (LifeTechnologies) following the manufacturer's protocol.

Saloura et al.

Mol Cancer Res; 13(2) February 2015 Molecular Cancer Research294




Western blottingNuclear extracts were prepared using the Nuclear Extract Kit

(Active Motif) to examine protein levels of WHSC1, and whole-cell extracts were used to examine protein levels of cytoplasmicNEK7 and ACTB. Primary antibodies used were anti-WHSC1(Abcam, ab75359, dilution: 1:1,000), anti-NEK7 (Thermo Scien-tific, H.691.4, dilution: 1:1,000) and anti-ACTB (Sigma-Aldrich,A5441, dilution: 1:4,000). For detection of histonemarks, nuclearextracts were prepared using the Nuclear Extract kit (Active Motif)and 1 mg of each extract was loaded for each experiment. Anti-bodies used were anti-H3K36me2 (Millipore, 07–369, dilution:1:4,000), anti-H3K36me3 (Abcam, ab9050, dilution: 1:1,000),anti-phospho H3 serine 10 (Millipore, 06–570, dilution:1:1,000), and anti-H3 (Abcam, ab1791, dilution: 1:20,000). Ananti-FLAG antibody (Sigma-Aldrich, F7425, dilution: 1:4,000)was used to assess efficacy of transfection of cell lines with FLAG-WHSC1-WT and FLAG -WHSC1-SET-deleted (DSET).

siRNA transfection and cell growth assaysMISSION_ siRNA oligonucleotide duplexes were purchased

from Sigma-Aldrich for targeting the human WHSC1 transcripts(SASI Hs02_00309678 and SASI Hs02_00309679). siNegativecontrol (siNC), which consists of 3 different oligonucleotideduplexes, were used as control siRNAs (Cosmo Bio). The siRNAsequences are described in Supplementary Table S3 and siAS(control) was obtained from QIAGEN (AllStars Negative ControlsiRNA, SI03650318). SCCHN cells were plated overnight in 24-well plates (2� 104 to 4� 104 cells per well) andwere transfectedwith siRNA duplexes (50 nmol/L final concentration) usingLipofectamine RNAimax (Life Technologies) for 144 hours (6days) to 192 hours (8 days) with retransfection performed at day5. The number of viable cells was measured using the CellCounting Kit-8 (Dojindo) at days 6 to 8 (24, 25).

Chromatin immunoprecipitation assaysChromatin immunoprecipitation (ChIP) assays were per-

formed using ChIP Assay kit (Millipore, 17–295) according tothe manufacturer's protocol (26–29). Briefly, WHSC1 and frag-mented chromatin complexes were immunoprecipitated withanti-WHSC1 (Abcam, ab75359, dilution: 1: 100) and anti-H3K36me2 (Millipore, 07–369, dilution: 1:100) antibodies 48hours after transfection of UD-SCC-2 cells with siRNAs forWHSC1. DNA fragments were quantified by real-time PCR usingNEK7-ChIP forward primer (50-GGATGTTTACACTCCTGACA-GC-30) and NEK7-ChIP reverse primer (50-GCGTCCGGAGT-GCGAGCCTAGC-30). We also conducted ChIP assays using pri-mers for GAPDH (Takara Bio, #5311) and b-globin (Takara Bio,#5316)whose expressions were not changed afterWHSC1 knock-down and confirmed that H3K36 dimethylation levels of thesegenes were not changed (Supplementary Fig. S1).

Cell-cycle analysis and apoptosis assaysThe bromodeoxyuridine (BrdUrd) flow kit (BD Pharmingen)

was used to determine the cell-cycle kinetics. The assay was per-formed according to the manufacturer's instructions (7, 30, 31).Briefly, cells were seeded overnight in 10-cm tissue culture dishesand treated with siRNAs (50 nmol/L) described as above usingmedium with 10% FBS without antibiotics for 72 hours, followedby addition of 10 mmol/L BrdUrd. Cells were harvested at 72 hoursand fixed in a solution containing paraformaldehyde and saponin.Then samples were incubated with DNase for 1 hour at 37�C and

fluorescein isothiocyanate (FITC)-conjugated anti-BrdUrd anti-body (dilution: 1:50) was added for 20 minutes at RT. Total DNAwas stained with 7-amino-actinomycin D (7-AAD), followed byflow cytometric analysis. Apoptosis was measured using theAnnexin V Apoptosis Detection Kit (Biovision) according to themanufacturer's protocol.

Microarray hybridization and statistical analysisUD-SCC2 cells were plated in 6-well dishes and treated with 2

different WHSC1-specific siRNAs (50 nmol/L) and 2 siRNAnegative controls (siNegative control, siAS) in triplicates for 48and 72 hours. Purified total RNA isolated from these samples waslabeled and hybridized onto Affymetrix GeneChip U133 Plus 2.0oligonucleotide arrays (Affymetrix) according to the manufac-turer's instructions (32–34). Probe signal intensities were nor-malized by RMA and Quantile normalization methods (using Rand Bioconductor). Next, signal intensity fluctuation due tointerexperimental variation was estimated. Each experiment wasreplicated (n¼ 1 and 2), and the SD of log2(intensity2/intensity1)was calculated for each set of intensity ranges with the midpointsbeing at log2[(intensity1þ intensity2)/2]¼ 5, 7, 9, 11, 13, and 15.We modeled intensity variation using the formula SD[log2(in-tensity2/intensity1)] ¼ a{log2[(intensity1 þ intensity2)/2]}) þ band estimated parameters a and b using the method of leastsquares. Using these values, the SD of intensity fluctuation wascalculated. The signal intensities of each probe were then com-pared between siWHSC1 (EXP) and controls (siNC or siAS)(CONT) and tested for up/downregulation by calculating the z-score: log2(intensityEXP/intensityCONT)/{a[log2[(intensityEXP þintensityCONT)/2]] þ b}. Resultant P values for the replicationsets were multiplied to calculate the final P value of each probe.These procedures were applied to each comparison: siNC versussiWHSC1, siAS versus siWHSC1, and siNC versus siAS, respec-tively. We determined up/downregulated gene sets as those thatsimultaneously satisfied the following criteria: (i) The Benjamini–Hochberg false discovery rate (FDR) � 0.05 for siNC versussiWHSC1, (ii) FDR � 0.05 for siAS versus siWHSC1 and theregulation direction is the same as (i) and (iii) siNC versus siAShas the direction opposite to (i) and (ii) orP>0.05 for siNCversussiAS. Finally, we performed a pathway analysis using the hyper-geometric distribution test, which calculates the probability ofoverlap between the up/downregulated gene set and each GeneOntology (GO) category compared against another gene list thatis randomly sampled. We applied the test to the identified up/downregulated genes to test whether or not they are significantlyenriched (FDR� 0.05) in each category of "Biological processes"(857 categories) as defined by the GO database.

ImmunocytochemistryFaDu cells (SCCHN cell line with low expression of WHSC1)

were plated in two 10-cm dishes, and next day, they weretransfected with FLAG-WHSC1-WT using Fugene HD (RocheApplied Science, 8 mg plasmid, 24 mL Fugene HD, 450 mLOPTIMEM) in 10 mL of cell culture medium with 10% FBS.After 2 days of transfection, cells were trypsinized and reseededin 2-chamber glass slides at a density of 2 � 104 cells per well(one well per experimental condition). After 24 hours ofincubation, medium was removed and cells were washed 2times with 1 mL of PBS. Following suctioning of PBS, 1 mL of4% paraformaldehyde was added to each well for 30 minutes at4�C to fix the cells. Subsequently, cells were washed with PBS

WHSC1 Upregulates NEK7 in Head and Neck Cancer

www.aacrjournals.org Mol Cancer Res; 13(2) February 2015 295




three times for 5 minutes each time at RT. 0.1% Triton X-100was added for 3 minutes at RT to permeabilize the cells andsamples were washed with PBS 3 times for 5 minutes each time.Then cells were blocked with 3% BSA for 1 hour at RT andincubated with primary antibodies: mouse anti-WHSC1(Abcam, ab75359, dilution: 1:500) and rabbit anti-H3K36me2(Cell Signaling Technology, #2901, dilution: 1:500) in a 1-mLsolution of 3% BSA at 4�C overnight. Next day, cells werewashed 4 times with 1 mL of PBS and secondary antibodieswere added (anti-mouse Alexa 594 -red- for FLAG, anti-rabbitAlexa 488 -green- for H3K36me2, dilution: 1:500) for 1 hour atRT with gentle shaking. Following this, cells were washed 4times with PBS and mounting medium with DAPI (VECTA-

SHIELD, Vector Laboratories) was added on each well. Thewells were finally covered with a glass slide. Confocal micros-copy (Leica 2D-Photon microscope) was used for the observa-tion of stained cells. ImageJ software was used to analyze theimages and quantify integrated intensities of the FLAG-WHSC1and H3K36me2.

Statistical analysisSpearman rank correlation coefficientswere calculated to assess

associations between WHSC1 and H3K36me2 expression levels(using the 4-point IHC scale). Expression levels were dichoto-mized as 0 to 1 versus 2 to 3. IHC score, stage, T-stage, N-stage,grade, and smoking status were treated as an ordinal variable, age

Normal squamousepithelium

SCCHN+1 +2 +3

A B C

SCCHN+1 +2 +3

Normal squamousepithelium

D E

Dysplasia (19)Normal(18) SCCHN(149)0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

WH

SC

1 ex

pres

sion

P < 0.0001**

Per

cent

age

of S

CC

HN

sec

tions

20

40

60

80

100

120

0Well Moderate Poor

High WHSC1Low WHSC1

High H3K36me2Low H3K36me2

Per

cent

age

of S

CC

HN

sec

tions

0

20

40

60

80

100

120

Well Moderate Poor

P = 0.0032 P = 0.003

H3K

36m

e2 le

vels

Dysplasia (13)

Normal(16)

SCCHN(96)

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

P < 0.0001*P < 0.0001*

P = 0.0007**

P = 0.0003*

P = 0.0485*

F

Lymph node Lymph node

Figure 1.WHSC1 is strongly overexpressed in locoregionally advanced SCCHN. A, IHC staining for WHSC1 (Abcam, ab75359) in TMAs (tissue microarrays) of patients withlocoregionally advanced SCCHN (n ¼ 149). Normal squamous epithelium sections (n ¼ 18) were used as baseline reference. Dysplastic squamous epithelialsections were also stained (n¼ 19). Slides were counterstained with hematoxylin and eosin. WHSC1 is mildly expressed in the basal and parabasal layers of normalsquamous epithelium with a nuclear localization. This expression increases significantly in SCCHN with a mixed cytoplasmic and nuclear expression: 27%, 40%, and33% of SCCHN sections were þ1, þ2, and þ3, respectively. Lymph node sections were used as a positive control for WHSC1 staining. B, WHSC1 protein expressionlevels are significantly higher in SCCHN than in normal epithelium and dysplastic squamous epithelium (� , P < 0.0001; Mann–Whitney U test). WHSC1 proteinexpression levels also increase significantly among normal, dysplastic squamous epithelium, and SCCHN (�� , P < 0.0001; Kruskal–Wallis test). C, IHC staining forH3K36me2 (Cell Signaling Technology, #2901) in TMAs of patients with locoregionally advanced SCCHN (n ¼ 96). Normal squamous epithelium sections (n ¼ 16)were used as baseline reference. Dysplastic squamous epithelial sections were also stained (n ¼ 13). Slides were counterstained with hematoxylin and eosin.H3K36me2 is mildly expressed in the basal and parabasal layers of normal squamous epithelium with nuclear localization. Lymph node sections were used as apositive control for H3K36me2 staining. D, H3K36me2 protein expression levels are significantly higher in SCCHN and dysplastic squamous epithelium than innormal squamous epithelium (� , P ¼ 0.0003 and P ¼ 0.0485; Mann–Whitney U test). H3K36me2 expression also increases significantly among normal, dysplasticsquamous epithelium, and SCCHN (�� , P ¼ 0.0007; Kruskal–Wallis test). E, high WHSC1 protein levels (þ2, þ3) correlate with poor grade in locoregionallyadvanced SCCHN. About 87% of the poorly differentiated SCCHN samples had high WHSC1 expression levels (P ¼ 0.0032; Cochrane–Armitage test). F, highH3K36me2 levels (þ2, þ3) correlate with poor grade in locoregionally advanced SCCHN. Almost 100% of the poorly differentiated SCCHN samples had highH3K36me2 expression levels (P ¼ 0.003; Cochrane–Armitage test).

Saloura et al.





as a continuous variable, and gender was treated as a categoricalvariable. Association between IHC scores and each factor wasevaluated using the Student t test, Cochran–Armitage trend test,and Fisher exact test. Statistical analyses were carried out using theR statistic program (http://www.r-project.org/).

ResultsCorrelation of WHSC1 expression and dimethylated H3K36 inadvanced SCCHN

We first examined the expression levels of WHSC1 anddimethylation status of lysine 36 on histone H3 (H3K36me2)in SCCHN by IHC analysis. We used tissue microarrays derivedfrom 149 patients with locoregionally advanced SCCHN forWHSC1 and of 96 patients for H3K36me2. We also obtainedclinical information for 92 of 149 cases examined forWHSC1 and54 of 96 cases for H3K36me2. Nineteen dysplastic epitheliumand 18 normal squamous epithelium samples were also exam-ined for WHSC1 and H3K36me2. Figure 1A shows represen-tative results of IHC applying a 4-grade scoring system (IHCscore 0, þ1, þ2, þ3). In the cases with strong staining (þ3),WHSC1 protein was observed in both the cytoplasm and thenucleus, whereas in cases with mild staining (þ1) WHSC1 wasobserved mainly in the nucleus. IHC analysis revealed moder-ate or strong (þ2, þ3) staining for WHSC1 in 73% of SCCHNsamples and absent or mild (0, þ1) staining in the remaining27% of cases, whereas 95% of the normal epithelium samplesshowed mild (þ1) staining. WHSC1 expression was signifi-cantly higher in SCCHN than in dysplastic and normal epithe-lium tissues (P < 0.0001; Mann–Whitney U test). Significant

differences in WHSC1 expression among normal, dysplastic,and SCCHN cases were also confirmed by the Kruskal–Wallistest (P < 0.0001; Fig. 1B).

H3K36me2 staining was observed in all of 16 normal, 13dysplastic, and 96 SCCHN samples (Fig. 1C), but their inten-sities were significantly different. H3K36me2 levels were strongin approximately 40% of the SCCHN cases (40 of 96) andmoderate in 38% of the cases (37 of 96). Approximately 80%of 13 dysplastic tissues stained moderately (10 of 13), andone case was scored as þ3. In the normal controls, nearly30% was moderately or strongly stained (5 of 16) and theremaining 70% was mildly stained (11 of 16). H3K36me2levels were significantly higher in SCCHN and dysplastic tissuesthan in normal controls (normal-dysplastic epithelium: P ¼0.0485, normal-SCCHN: P ¼ 0.0003, Mann–Whitney U test).The Kruskal–Wallis test confirmed significant differences inH3K36me2 levels among normal epithelium, dysplastic epi-thelium, and SCCHN (P ¼ 0.0007; Fig. 1D). In addition, weexamined the correlation between WHSC1 expression levelsand dimethylated H3K36me2 in 123 samples and found astatistically significant positive correlation (Spearman rankcorrelation coefficient, r ¼ 0.545, P < 0.0001).

We then investigated the correlation of the WHSC1 expressionand dimethylated H3K36me2 with various clinical parameters,including grade, tumor size (T), nodal stage (N), TNM stage,smoking status, HPV status, age, and gender, in 92 (for WHSC1)and 53 (for H3K36me2) patients for whom clinical informationwas available. Results are summarized in Table 1. Although mostof the clinicopathologic parameters were not associated withWHSC1 and dimethylated H3K36me2, we found significant

Table 1. Clinicopathologic correlations of dichotomized WHSC1 and H3K36me2 expression by IHC in locoregionally advanced SCCHN

WHSC1 H3K36me2Clinicopathologic parameters Lowð0;þ1Þ

ðn¼17ÞHighðþ2;þ3Þ

n¼75P Lowð0;þ1Þ

ðn¼9ÞHighðþ2;þ3Þ

n¼44P

Gender 1.00 0.086Female 5 21 0 14Male 12 54 7 25

Age 51 � 11.6 57 � 9.2 0.01 59 � 5.7 55 � 10.7 0.34Smoking status 0.064 0.36�0 person-years 4 12 1 91–39 person-years 9 25 3 15�40 person-years 2 29 0 13

Grade 0.0032 0.003Well differentiated 8 5 6 6Moderately 2 25 1 13differentiatedPoorly differentiated 2 13 0 9

Stage 0.061 0.068I 4 5 3 6II 0 1 0 1III 2 6 2 5IV 12 60 2 25

T 0.98 0.071T1 5 14 4 11T2 2 15 2 8T3 3 14 1 7T4 8 27 0 10

N 0.042 0.43N0 9 19 3 17N1 2 8 2 2N2 6 41 2 17N3 1 7 0 4

HPV status 0.19 0.166p16þ 10 36 7 28p16� 2 24 0 12






correlation of both WHSC1 and H3K36me2 with poorly differ-entiated histologic grade (WHSC1: P ¼ 0.0032, H3K36me2: P ¼0.003, Cochrane–Armitage test; Fig. 1E and F). In this cohort ofpatients, overall and recurrence-free survival analyses were notfeasible due to lack of power.

Overexpression of WHSC1 in SCCHN cell lines and its criticalrole in cell proliferation and survival

To clarify the physiologic function of WHSC1 in SCCHN, weexamined the expression profile and the effect of WHSC1knockdown on the proliferation of SCCHN cells. Quantitative

real-time PCR demonstrated that all SCCHN cell lines examinedshowed higher levels of WHSC1 than in a control humannormal keratinocyte cell line (KGM; Fig. 2A). We next trans-fected UD-SCC-2 cells, an HPV-positive SCCHN cell line highlyexpressing WHSC1, with control siRNA (siNC) and 2-indepen-dent specific siRNAs targeting WHSC1 (siWHSC1 #1 and #2). Asshown in Fig. 2B, WHSC1 expression was knocked down at bothmRNA and protein levels. Subsequently, we transfected thesesiRNAs into 1 HPV-positive (UD-SCC-2) and 3 HPV-negativecell lines (UM-SCC-35, HN-SCC-151, and PE/CA-PJ15) endog-enously overexpressing WHSC1 and examined their effect on

C

A

0

1

2

3

4

5

6

7

8

9

KG

MH

N-5

HN

-6JS

Q-3

UD

-SC

C-2

UM

-SC

C-2

3U

T-S

CC

-40

HN

-SC

C-1

35H

N-S

CC

-151

HN

-SC

C-1

66S

CC

-25

SC

C-3

5P

E/C

A-P

J15

OE

CM

-1B

ICR

3193

VU

147T

Rel

ativ

e W

HS

C1

expr

essi

onB Western

Rel

ativ

e W

HS

C1

expr

essi

on

0.2

0.4

0.6

0.8

1.0

1.2

siNC #1 #2siWHSC1

Real-time PCR

UD-SCC-2

0

0.2

0.4

0.6

0.8

1.0

1.2*

*

Rel

ativ

e ce

ll nu

mbe

r

0

0.2

0.4

0.6

0.8

1.0

1.2

Rel

ativ

e ce

ll nu

mbe

r

UM-SCC-35

**

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 30

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Day 3

PE/CA-PJ15

Rel

ativ

e ce

ll nu

mbe

r

Day 6

siNC

siWHSC1

HN-SCC-151

Day 3 Day 6

Rel

ativ

e ce

ll nu

mbe

r0

0.20.40.60.81.01.21.41.61.82.0

**

**

FaDu

Rel

ativ

e ce

ll nu

mbe

r

1: siNC2: siWHSC1#13: siWHSC1#2

DsiNC siWHSC1

0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

siNC #1 #2

siWHSC1

IB:WHSC1

IB:ACTB

Day 6Day 8Day 8

*

Rel

ativ

e nu

mbe

r of

ap

opto

tic c

ells

0

0.5

1.0

1.5

2.0

2.5

Figure 2.WHSC1 is overexpressed and its knockdown causes significant growth suppression and apoptosis in SCCHN cell lines. A, quantitative real-time PCR of WHSC1 in 15SCCHN cell lines compared with a normal control keratinocyte cell line (KGM). mRNA levels were normalized by GAPDH. B, knockdown of WHSC1 mRNA andprotein levels using 2 different WHSC1-specific siRNAs (siWHSC1#1 and siWHSC1#2). UD-SCC-2 cells were treated with control siRNA (siNC), siWHSC1#1, andsiWHSC1#2 for 3 days.mRNA and regular protein extractionwere performed.WHSC1 (150 kD)was blottedwith anti-WHSC1 (HPA015801, Sigma-Aldrich). ACTBwasused as a loading control. C, cell cytotoxicity assays by Cell Counting Kit-8 (Dojindo) in 4 SCCHN cell lines with endogenous overexpression of WHSC1. siRNA-mediated WHSC1 knockdown caused significant growth suppression in one HPV-positive (UD-SCC-2) and 3 HPV-negative (SCC-35, HN-SCC-151, CE/PA-PJ15)SCCHN cell lines after 6 to 8 days of treatment (� , P <0.05). siWHSC1 treatment of a cell linewith low expression ofWHSC1 (FaDu) did not cause growth suppression.Each condition was plated in quadruples, and statistical comparisons performed by the Student t test. D, assessment of apoptosis in UD-SCC-2 cells usingtheAnnexinV assay.UD-SCC-2 cellswereplated in 10-cmdishes and treatedwith siNCand siWHSC1 for 5 days. Cellswere collected andprocessedusing an annexinVassay (Biodivision). Annexin V–positive cells increased by an average of approximately 2 times in the siWHSC1-treated cells compared with siNC-treated cells.Three independent experiments were performed and the average of ratios of relative increase in apoptosis [r ¼ (percentage of early þ late apoptotic SCC2cells treated with siWHSC1)/(percentage of earlyþ late apoptotic SCC2 cells treated with siNC)] was calculated and statistical comparisons were performed by theStudent t test (� , P < 0.05).

Saloura et al.





cell growth using the Cell Counting Kit-8 (Dojindo). WHSC1knockdown resulted in significant growth suppression of all 4cell lines (Fig. 2C), suggesting that WHSC1 possesses a signif-icant role in the proliferation of SCCHN cells. Treatment ofFaDu cells with low expression of WHSC1 with WHSC1-specificsiRNAs did not result in growth suppression (Fig. 2C andSupplementary Fig. S2), indicating that the observed effect inSCCHN cells was unlikely to be due to an off-target effect. Toexamine whether WHSC1 knockdown induced death throughapoptosis, we performed Annexin V assays in UD-SCC-2 cells.WHSC1 knockdown resulted in doubling of the Annexin V–positive cells from 25.4% in the siRNA control–treated (siNC)group to 51.3% in the siWHSC1-treated group (Fig. 2D). Thesedata indicate that knockdown of WHSC1 induces growth sup-pression and apoptosis of SCCHN cells.

Induction of global histone methylation changes of H3K36 byWHSC1 in SCCHN cells

To evaluate whether overexpression of WHSC1 induces globalhistone methylation changes in SCCHN cells, we used a knock-down system of PE/CA-PJ15 cells transfected with a WHSC1-

specific siRNA (siWHSC1#1). Following 5 days of culture, nuclearextracts were prepared and immunoblotted with anti-dimethy-lated H3K36 and anti-trimethylated H3K36 antibodies. A signif-icant decrease in di- and trimethylatedH3K36was observed in thecells treated with siWHSC1 compared with those treated withcontrol siRNA (Fig. 3A). To evaluate the effect of WHSC1 in theaforementioned histone marks in a gain-of-function system,FaDu cells with very low expression of WHSC1 (SupplementaryFig. S2) were transfected with FLAG-WHSC1-WT and FLAG-WHSC1-DSET (enzyme-inactive type of WHSC1) for 2 days(14), andWestern blot analyses were performed to assess changesin di- and trimethylated H3K36. Both di- and trimethylatedH3K36 levels in FaDu cells transfected with FLAG-WHSC1-WTwere higher than those in FLAG-WHSC1-DSET–transfected FaDucells (Fig. 3B).

To further examine the relationship between WHSC1 over-expression and H3K36me2 levels in SCCHN cells, we performedimmunocytochemistry (ICC) using anti-FLAG and anti-H3K36me2 antibodies in FaDu cells transfected with theFLAG-WHSC1-WT (Fig. 3C). We quantified the fluorescenceintensity by Fiji ImageJ software and found statistically higher

C D

0

0.02

0.04

0.06

0.08

0.10

0.12

0.14

FLAG-WHSC1Negative

H3K

36m

e2 le

vels

per

are

a

FLAG-WHSC1Positive

U = 0Z = 0.23646P = 0.0082

A

IB: H3K36me3

IB: H3K36me2

IB: WHSC1

siNC siWHSC1

IB: Histone H3

siNC siWHSC1

H3K36me2

Rel

ativ

e H

3K36

me2

leve

ls

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Rel

ativ

e H

3K36

me3

leve

ls

siNC siWHSC1

B

WHSC1-WTWHSC1-

ΔSET

IB: H3K36me3

IB: H3K36me2

IB: WHSC1

IB: Histone H3

H3K36me3

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Rel

ativ

e H

3K36

me2

leve

ls

H3K36me2

WHSC

1-WT

WHSC1-

ΔSET

0

0.5

1.0

1.5

2.0

2.5

3.0

Rel

ativ

e H

3K36

me3

leve

ls

WHSC

1-WT

WHSC1-

ΔSET

H3K36me3

FLAG (WHSC1)

H3K36me2

DAPI

Merge DIC

10 µm

Figure 3.WHSC1 induces global histone di- and trimethylation of H3K36 in SCCHN cells. A,WHSC1 knockdown causes global decrease in H3K36me2 and H3K36me3 in PE/CA-PJ15 cells. After treatment of UD-SCC-2 cells with siWHSC1 for 5 days, Western blotting was performed for H3K36me2 (Millipore, 07–369), H3K36me3 (Abcam,ab9050), and WHSC1 (Abcam, ab75359) in nuclear extracts of UD-SCC-2 cells. Histone H3 (Abcam, ab1791) was used as a loading control. B, FaDu cells withendogenously low levels of WHSC1 were transfected with pCAGGS-WT-WHSC1-FLAG (wild-type 150-kD WHSC1) and pCAGGS-WHSC1-DSET-FLAG (150-kDWHSC1 with deleted SET domain) plasmids and Western blotting was performed for H3K36me2, H3K36me3, and WHSC1. H3K36me2 and H3K36me3 levels werehigher in theWT-WHSC1-FLAG than in theWHSC1-DSET-FLAG–transfected FaDu cells. Histone H3was used as a loading control. C, FaDu cells were transfectedwithFLAG-WHSC1-WT and ICC was performed with anti-FLAG (Sigma-Aldrich, M2, dilution: 1:500, red) and anti-H3K36me2 (Cell Signaling Technology, #2901,dilution: 1:500, green) antibodies. Nuclei were stained with 40 ,6-diamidino-2-phenylindole (DAPI). Transfected FaDu cells exhibited higher WHSC1 and H3K36me2integrated densities. D, FaDu cells transfected with FLAG-WHSC1-WT cells showed significantly higher levels of H3K36me2 expression than untransfectedcells (P ¼ 0.0082; Mann–Whitney U test).






staining of H3K36me2 in FLAG-WHSC1-WT cells than in theuntransfected cells (P < 0.0001; Mann–Whitney U test; Fig. 3D).Interestingly, we also observed that the mean H3K36me2 fluo-rescence levels of FLAG-WHSC1-DSET–transfected cells were sig-nificantly lower than that of untransfected cells (P ¼ 0.0367;Mann–Whitney U test), indicating a possible dominant-negativeeffect of FLAG-WHSC1-deltaSET in H3K36me2. These resultsindicate that overexpression of WHSC1 enhances methylationof histone H3 lysine 36 in SCCHN cells.

Direct regulation NEK7 transcription by WHSC1Next, we attempted to identify the genes directly regulated by

WHSC1 to further clarify the biologic function of WHSC1 andelevated H3K36 methylation in SCCHN cells. We transfectedUD-SCC2 cells with 2 control siRNAs and 2 independent

WHSC1 siRNAs and total RNA was extracted to conduct micro-array expression profile analysis (Affymetrix platform) at 48and 72 hours after transfection. We used these time points toavoid confounding of our analysis by death-associated path-ways, given that cell death with WHSC1 knockdown was notedonly after at least 4 days of siRNA treatment. Hypergeometricdistribution analysis revealed significant downregulation of 26genes by more than 50% reduction (Supplementary Fig. S3 andSupplementary Table S4). Among these genes, NIMA-relatedkinase-7 (NEK7), homeodomain-interacting protein kinase-3(HIPK3), and mitogen-activated protein kinase 8 (MAPK8)were previously reported to be involved in cell-cycle regulationand mitosis, apoptosis, and chemoradioresistance. NEK7 is aserine/threonine kinase which is required for cell-cycle progres-sion through mitotic spindle formation and cytokinesis

A

D

B C

IB: WHSC1

IB: NEK7

IB: ACTB

siNC #1 #2siWHSC1

0

0.2

0.4

0.6

0.8

1.0

1.2

siNCsiWHSC1

#1 #2

Rel

ativ

e pr

otei

n le

vels

WHSC1NEK7

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Mock WHSC1-WT

Rel

ativ

e N

EK

7ex

pres

sion

leve

ls

*

E

TSS

NEK7 transcriptional regulation region

ChIP-f/r

9IP: anti-WHSC1

siWHSC1

87

6543

2

10

Per

cent

age

of in

put [

×10

2]

siNC0

1

2

3

4

5

6

7

8

siWHSC1 siNCP

erce

ntag

e of

inpu

t [×

102]

IP: anti-H3K36me2

* *

WHSC1NEK7

Rel

ativ

e m

RN

A e

xpre

ssio

n le

vels

0

0.2

0.4

0.6

0.8

1.0

1.2

siNCsiWHSC148 h

#1 #2 siNCsiWHSC1#1 #2

72 h

1.4

*

*

**

*

*

**

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Rel

ativ

e m

RN

A e

xpre

ssio

n le

vels

siNCsiWHSC1

#1 #2

*

*

**

WHSC1NEK7

100 bp1st exon 1st intron

Figure 4.NEK7 is a direct downstream target and is transcriptionally regulated byWHSC1 throughH3K36 dimethylation in HPV-positive andHPV-negative SCCHN cell lines. A,quantitative real-timePCRofNEK7 after siRNA-mediated knockdownofWHSC1 inHPV-positiveUD-SCC-2 cellswith 2 different siRNAs at 48 and72 hours. NEK7wassignificantly downregulated at 48 and 72 hours in the siWHSC1-treated cells compared with controls. Conditions were plated in triplicates. Statisticalcomparisons between siWHSC1 and siNC groups were performed using the Student t test (� , P < 0.05). B, quantitative real-time PCR of NEK7 after siRNA-mediatedknockdown of WHSC1 in HPV-negative PE/CA-PJ15 cells with 2 different siRNAs at 72 hours. NEK7 is significantly downregulated at 72 hours in the siWHSC1-treated cells compared with controls. Conditions were plated in triplicates. Statistical comparisons between siWHSC1 and siNC groups were performed using theStudent t test (� , P < 0.05). C, confirmation of NEK7 decrease at the protein level after WHSC1 knockdown in PE/CA-PJ15 cells. Regular and nuclear proteinextracts were blotted for NEK7 (Thermo-Scientific, H.691.4) and WHSC1 (Abcam, ab75359), respectively. ACTB was used as a loading control. D, quantitative real-time PCR for NEK7 after transfection of FaDu cells with pFLAG-WHSC1-WT. NEK7 mRNA transcript levels increased by 35% compared with control FaDu cellstransfected with mock-plasmid (FLAG-Mock). Conditions were performed in triplicates and statistical comparisons between the mock- and WT-WHSC1-FLAG-transfected cells were performed using the Student t test (� , P < 0.05). E, ChIP assay was performed in UD-SCC-2 cells treated with siWHSC1 for 3 days. Top,schematic diagram of the NEK7 promoter region. PCR amplified fragments are positioned by nucleotide number relatives to TSS (arrows). Bottom, real-timePCR analysis using primer pairs as described under Materials andMethods. Immunoprecipitationwas performedwith ChIP grade anti-WHSC1 (Abcam, ab75359) andanti-H3K36me2 (Millipore, 07–369) antibodies. Amplicons were significantly decreased in both anti-WHSC1 IP group and anti-H3K36me2 IP group. The y-axisshows a percentage of the input chromatin. Results are the mean � SD of 3 independent experiments, and P values were calculated using the Student t test(� , P < 0.05).

Saloura et al.





(35, 36). HIPK3 is a serine/threonine kinase which was shownto negatively regulate apoptosis by phosphorylating FADD andinhibiting the interaction between FADD and caspase-8, thusinducing resistance to Fas-R–mediated apoptosis (37). MAPK8is a serine/threonine kinase involved in epithelial transforma-tion, migration, differentiation, and transcriptional regulationand has also been implicated in induction of radioresistance inSCCHN cells (38). These potential downstream genes werevalidated with quantitative real-time PCR in UD-SCC-2 cellsafter knockdown of WHSC1 (Fig. 4A and Supplementary Fig.S4). Among these genes, NEK7 was previously reported toregulate proliferation of cancer cells (36, 39), which was aphenotype observed with WHSC1 knockdown. In addition,NEK7 was confirmed to be significantly decreased after knock-down of WHSC1 at the protein level. Therefore, we focused onNEK7 for further analysis.

Decrease of NEK7 expression after WHSC1 knockdown wasalso observed in the HPV-negative PE/CA-PJ15 cell at the tran-scriptional and protein levels (Fig. 4B and C). Concordantly,transfection of FaDu cells with the FLAG-WHSC1-WT plasmidsignificantly enhanced the expression levels of NEK7 (Fig. 4D).

To assesswhetherWHSC1directly regulates the transcription ofNEK7, we conducted ChIP assays in UD-SCC-2 cells transfectedwith either siNC or siWHSC1 using ChIP-grade antibodies forWHSC1 and H3K36me2 (Fig. 4E). Quantitative PCR for NEK7showed a 54% reduction in the gene levels between the siNC-treated and siWHSC1-treated UD-SCC-2 cells, with a concordant47%decrease in the levels ofH3K36me2. This result supports that

NEK7 is transcriptionally regulated by WHSC1 through dimethy-lation of histone H3 lysine 36.

To examine whether WHSC1-mediated transcriptional regu-lation of NEK7 resulted in a phenotypic effect on cell-cycleprogression, flow cytometry for cell-cycle analysis was per-formed in UD-SCC-2 cells after knockdown of WHSC1 orNEK7. UD-SCC-2 cells were transfected with siNC, siWHSC1,or siNEK7, and we synchronized the cell cycle at the G0–G1

phase with aphidicolin treatment. After release of the cell cyclein growth arrested cells, cell-cycle analysis using flow cytometryshowed that the percentage of cells in the G2–M phase wasdecreased from 44.1% in the siNC-treated cells to 19.4% and30.5% in the siWHSC1- and siNEK7-treated cells respectively.At the same time, the percentage of cells in the G0–G1 phase wasincreased from 26.3% in the siNC-treated cells to 49.6% and38.6% in the siWHSC1- and siNEK7-treated cells, respectively(Fig. 5A and B). Previous reports showed that NEK kinases,including NEK7, play important roles in the cell-cycle check-point regulation at G1–S, intra-S, and G2–M phases in additionto their established functions during mitosis (39, 40), and ourresults appear to be consistent with these findings. In addition,we prepared cell lysates 24 hours after the release of cell cycleand conducted Western blotting for phosphorylated H3 Serine10, a maker of mitosis. Consistently, phosphorylated H3S10was decreased in cells treated with either siWHSC1 or siNEK7compared with controls (Fig. 5C), indicating that mitoticcells were diminished after treatment with either siWHSC1or siNEK7. Because WHSC1 knockdown produced a similar

A

siNC

siWHSC1

siNEK7

(A)

DNA content

Cel

l nu

mbe

r

DNA content

Cel

l nu

mbe

r

DNA content

Cel

l nu

mbe

r

DNA content

Cel

l nu

mbe

r

DNA content

Cel

l nu

mbe

r

DNA content

Cel

l nu

mbe

r

siNCsiWHSC1siNEK7

Per

cent

age

of c

ells

in e

ach

phas

e

0

10

20

30

40

50

60

Sub-G1 G0–G1 S G2–M

IB: H3S10ph

IB: ACTB

siNC siWHSC1 siNEK7

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

siNC siWHSC1 siNEK7

H3S

10ph

leve

ls n

orm

aliz

ed b

y A

CT

B

(B) B C

Figure 5.WHSC1 delays cell-cycle progression through transcriptional regulation of NEK7. A, histograms of cell-cycle analysis of UD-SCC-2 cells synchronized withaphidicholin (5 mg/mL). UD-SCC-2 cells were plated in 10-cm dishes and treated with siNC, siWHSC1, and siNEK7 for 3 days. On the second day of transfection, cellswere exposed to aphidicholin for 48 hours. Following that, cells were released and cell-cycle analysiswas performed at 0 hour (A) and 36 hours (B). G2–Mphase cellsweredecreased from44.1% in the siNCgroup to 19.4% in the siWHSC1 group and30.5% in the siNEK7group,whereasG0–G1 phase cells increased from26.3% to49.6%in the siWHSC1 group and 36.6% in the siNEK7 group. Three independent experiments were performed and demonstrated delay in cell-cycle progression. B,numerical analysis of the flow cytometry result (A), classifying cells by cell-cycle status. C, Western blotting for phospho-H3-Ser10 (Millipore, 06–570) in regularextracts from UD-SCC-2 cells treated with siNC, siWHSC1, and siNEK7 at 24 hours after release from aphidicholin synchronization. Phospho-H3-Ser10 decreased inboth siWHSC1 and siNEK7 groups compared with controls.






phenotypic effect with NEK7 knockdown, that is, a delay in thecell-cycle progression, NEK7 is likely to be one of the key factorsin the molecular pathway regulated by WHSC1 in SCCHN cells.Taken together, these results suggest that WHSC1 plays a criticalrole in the cell-cycle progression of SCCHN cells through directactivation of NEK7 expression.

DiscussionHMTs are a group of histone modifiers that are emerging as

attractive candidates for drug development (41–44). Aninhibitor of EZH2, an HMT, has already been introduced inphase I trials with the goal to target patients with refractorydiffuse large B-cell lymphoma (DLBCL) and follicular lym-phoma with the Y641 and A677 EZH2 activating mutations(45, 46). In this study, we showed that the HMT WHSC1 issignificantly overexpressed in 73% of locoregionally advancedSCCHN tissues and provide evidence that supports the sig-nificant pathophysiologic role of WHSC1 in SCCHN. WHSC1expression significantly increased with transition from normalto dysplastic epithelium and squamous cell carcinoma, indi-cating a vital role of WHSC1 in the initial stages of head andneck carcinogenesis. High WHSC1 expression and H3K36me2levels were associated with poor differentiation, which sug-gests that WHSC1 may drive a dedifferentiation reprogram-ming of epithelial cells. This is in accordance with the obser-vation that WHSC1 is highly expressed during embryonicdevelopment (16, 47), although its physiologic function inthis setting has not been elucidated yet. It is possible thatWHSC1 may allow for the maintenance of stemness andcellular plasticity which is normally required during embry-onic development before the initiation of differentiationtoward a specific cell fate.

Our results show thatWHSC1 is important for cell proliferationand its knockdown induces delay in the cell-cycle progression inSCCHN cell lines. We identified that this effect is mediated byNEK7and thatNEK7 is a direct downstream target geneofWHSC1in SCCHN cells. NEK7 belongs to the NEK family of proteinkinases which have a prominent role in cell-cycle control andmitotic spindle formation (48–50). A number of studies on theNEK family of kinases have shown their potential role in onco-genesis (36, 51–55), whereas a recent study reported that highlevels of NEK7were associated with poor survival in patients withgallbladder carcinoma (51). Furthermore, the TCGA databaserevealed that amplification of NEK7 frequently occurred in var-ious types of cancer (Supplementary Fig. S5), suggesting thatactivation of NEK7 may be important in oncogenesis. Takentogether, the WHSC1–NEK7 pathway is likely to play a criticalrole in the oncogenesis of SCCHN. Further functional analysis iswarranted to explore the importance of this pathway as a target of

SCCHN therapy, as well as other aspects of the WHSC1-depen-dent network, such as its potential effect on apoptosis andchemoradioresistance pathways through regulation HIPK3 andMAPK8.

In conclusion, this study underlines the possible oncogenicactivity of WHSC1 in SCCHN. As the methylation networks ofWHSC1 in cancer are still largely unknown, research in this fieldwill advance our knowledge and potentially accelerate thedevelopment of truly novel therapeutics for SCCHN. Develop-ment of specific inhibitors targeting WHSC1 may be a prom-ising approach to improve the treatment outcomes for patientswith SCCHN.

Disclosure of Potential Conflicts of InterestJ. Licht reports receiving a commercial research grant from Epizyme and is a

consultant/advisory boardmember for Abbvie, GlaxoSmithKline, and Celgene.Y. Nakamura reports receiving a commercial research grant from and is aconsultant/advisory board member for Oncotherapy Science. No potentialconflicts of interest were disclosed by the other authors.

Authors' ContributionsConception and design: V. Saloura, H.-S. Cho, T. Seiwert, J. Licht, Y. Nakamura,R. HamamotoDevelopment of methodology: V. Saloura, H.-S. Cho, T. Tsunoda,R. HamamotoAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): V. Saloura, H. Alachkar, M. Nakakido, M. Lingen,R. HamamotoAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): V. Saloura, K. Kiyotani, H. Alachkar, Z. Zuo,M. Nakakido, T. Tsunoda, T. Seiwert, M. Lingen, Y. Nakamura, R. HamamotoWriting, review, and/or revision of the manuscript: V. Saloura, T. Tsunoda,T. Seiwert, M. Lingen, Y. Nakamura, R. HamamotoAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): V. Saloura, R. HamamotoStudy supervision: V. Saloura, Y. Nakamura, R. Hamamoto

AcknowledgmentsThis work was funded in part by a Conquer Cancer Foundation of ASCO

Young Investigator Award. Anyopinions,findings, and conclusions expressed inthismaterial are those of the author(s) and do not necessarily reflect those of theAmerican Society of Clinical Oncology� or the Conquer Cancer Foundation.The authors thank Drs. Ezra Cohen, Kenbun Sone, Lianhua Piao, Deng Zhen-zong, and Gottfried Von Keudell for technical support and helpful discussion.They also thank the members of Nakamura laboratory in the University ofChicago for kind support.

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

Received May 21, 2014; revised September 8, 2014; accepted September 23,2014; published OnlineFirst October 3, 2014.

References1. ChoHS,Hayami S, ToyokawaG,MaejimaK, Yamane Y, Suzuki T, et al. RB1

methylation by SMYD2enhances cell cycle progression through an increaseof RB1 phosphorylation. Neoplasia 2012;14:476–86.

2. Cho HS, Suzuki T, Dohmae N, Hayami S, Unoki M, Yoshimatsu M, et al.Demethylation of RB regulator MYPT1 by histone demethylase LSD1promotes cell cycle progression in cancer cells. Cancer Res 2011;71:1–6.

3. Hayami S, YoshimatsuM, VeerakumarasivamA, UnokiM, Iwai Y, TsunodaT, et al. Overexpression of the JmjC histone demethylase KDM5B in human

carcinogenesis: involvement in the proliferation of cancer cells through theE2F/RB pathway. Mol Cancer 2010;9:59.

4. Ezponda T, Popovic R, Shah MY, Martinez-Garcia E, Zheng Y, Min DJ,et al. The histone methyltransferase MMSET/WHSC1 activates TWIST1 topromote an epithelial-mesenchymal transition and invasive properties ofprostate cancer. Oncogene 2013;32:2882–90.

5. Mohammad HP, Baylin SB. Linking cell signaling and the epigeneticmachinery. Nat Biotech 2010;28:1033–8.

Saloura et al.





6. Baylin SB. Resistance, epigenetics and the cancer ecosystem. Nat Med2011;17:288–9.

7. Tomasi TB,Magner WJ, Khan AN. Epigenetic regulationof immune escapegenes in cancer. Cancer Immunol Immunoth 2006;55:1159–84.

8. Richon VM, Johnston D, Sneeringer CJ, Jin L, Majer CR, Elliston K, et al.Chemogenetic analysis of human protein methyltransferases. Chem BiolDrug Des 2011;78:199–210.

9. Cloos PA, Christensen J, Agger K, Helin K. Erasing themethylmark: histonedemethylases at the center of cellular differentiation and disease. GenesDev 2008;22:1115–40.

10. Copeland RA, Solomon ME, Richon VM. Protein methyltransferasesas a target class for drug discovery. Nat Rev Drug Discov 2009;8:724–32.

11. Kooistra SM, Helin K. Molecular mechanisms and potential functions ofhistone demethylases. Nat Rev Mol Cell Biol 2012;13:297–311.

12. Volkel P, Angrand PO. The control of histone lysine methylation inepigenetic regulation. Biochimie 2007;89:1–20.

13. Douglas J, Coleman K, Tatton-Brown K, Hughes HE, Temple IK, Cole TR,et al. Evaluation of NSD2 and NSD3 in overgrowth syndromes. Eur J HumGenet 2005;13:150–3.

14. Toyokawa G, Cho HS, Masuda K, Yamane Y, Yoshimatsu M, Hayami S,et al. Histone lysine methyltransferase wolf-hirschhorn syndrome candi-date 1 is involved in human carcinogenesis through regulation of the Wntpathway. Neoplasia 2011;13:887–98.

15. Hudlebusch HR, Santoni-Rugiu E, Simon R, Ralfkiaer E, Rossing HH,Johansen JV, et al. The histonemethyltransferase and putative oncoproteinMMSET is overexpressed in a large variety of human tumors. Clin CancerRes 2011;17:2919–33.

16. Stec I, Wright TJ, van Ommen GJ, de Boer PA, van Haeringen A,Moorman AF, et al. WHSC1, a 90 kb SET domain-containing gene,expressed in early development and homologous to a Drosophiladysmorphy gene maps in the Wolf-Hirschhorn syndrome critical regionand is fused to IgH in t(4;14) multiple myeloma. Hum Mol Genet1998;7:1071–82.

17. Lauring J, Abukhdeir AM, Konishi H, Garay JP, Gustin JP, Wang Q, et al.The multiple myeloma associated MMSET gene contributes to cellularadhesion, clonogenic growth, and tumorigenicity. Blood 2008;111:856–64.

18. Martinez-Garcia E, Popovic R,MinDJ, Sweet SM, Thomas PM,Zamdborg L,et al. The MMSET histone methyl transferase switches global histonemethylation and alters gene expression in t(4;14) multiple myeloma cells.Blood 2011;117:211–20.

19. Kuo AJ, Cheung P, Chen K, Zee BM, Kioi M, Lauring J, et al. NSD2 linksdimethylation of histone H3 at lysine 36 to oncogenic programming. MolCell 2011;44:609–20.

20. Pei H, Zhang L, Luo K, Qin Y, Chesi M, Fei F, et al. MMSET regulateshistone H4K20 methylation and 53BP1 accumulation at DNA damagesites. Nature 2011;470:124–8.

21. Worsham MJ, Stephen JK, Chen KM, Havard S, Shah V, Gardner G, et al.Delineating an epigenetic continuum in head and neck cancer. Cancer Lett2014;342:178–84.

22. Lingen MW, Pinto A, Mendes RA, Franchini R, Czerninski R, TilakaratneWM, et al. Genetics/epigenetics of oral premalignancy: current status andfuture research. Oral Dis 2011;17(Suppl 1):7–22.

23. Weinstein JN,Collisson EA,Mills GB, Shaw KR,Ozenberger BA, Ellrott K,et al. The cancer genome atlas pan-cancer analysis project. Nat Genet2013;45:1113–20.

24. Hayami S, Kelly JD, Cho HS, Yoshimatsu M, Unoki M, Tsunoda T, et al.Overexpression of LSD1 contributes to human carcinogenesis throughchromatin regulation in various cancers. Int J Cancer 2011;128:574–86.

25. YoshimatsuM, ToyokawaG,Hayami S,UnokiM, TsunodaT, FieldHI, et al.Dysregulation of PRMT1 and PRMT6, Type I arginine methyltransferases,is involved in various types of human cancers. Int J Cancer 2011;128:562–73.

26. HamamotoR, Furukawa Y,MoritaM, Iimura Y, Silva FP, LiM, et al. SMYD3encodes a histonemethyltransferase involved in the proliferation of cancercells. Nat Cell Biol 2004;6:731–40.

27. Hamamoto R, Silva FP, TsugeM,Nishidate T, Katagiri T, Nakamura Y, et al.Enhanced SMYD3 expression is essential for the growth of breast cancercells. Cancer Sci 2006;97:113–8.

28. Cho HS, Kelly JD, Hayami S, Toyokawa G, Takawa M, Yoshimatsu M,et al. Enhanced expression of EHMT2 is involved in the proliferation ofcancer cells through negative regulation of SIAH1. Neoplasia 2011;13:676–84.

29. Toyokawa G, Cho HS, Iwai Y, Yoshimatsu M, Takawa M, Hayami S, et al.The histone demethylase JMJD2B plays an essential role in human carci-nogenesis through positive regulation of cyclin-dependent kinase 6. Can-cer Prev Res 2011;4:2051–61.

30. Takawa M, Cho HS, Hayami S, Toyokawa G, Kogure M, Yamane Y, et al.Histone lysine methyltransferase SETD8 promotes carcinogenesis byderegulating PCNA expression. Cancer Res 2012;72:3217–27.

31. Kogure M, Takawa M, Cho HS, Toyokawa G, Hayashi K, Tsunoda T, et al.Deregulation of the histone demethylase JMJD2A is involved in humancarcinogenesis through regulation of the G/S transition. Cancer Lett 2013;336:76–84.

32. Cho HS, Toyokawa G, Daigo Y, Hayami S, Masuda K, Ikawa N, et al. TheJmjC domain-containing histone demethylase KDM3A is a positive reg-ulator of the G1/S transition in cancer cells via transcriptional regulation ofthe HOXA1 gene. Int J Cancer 2012;131:E179–89.

33. Kang D, Cho HS, Toyokawa G, Kogure M, Yamane Y, Iwai Y, et al. Thehistone methyltransferase Wolf-Hirschhorn syndrome candidate 1-like 1(WHSC1L1) is involved in human carcinogenesis. Genes ChromosomesCancer 2013;52:126–39.

34. Kogure M, Takawa M, Saloura V, Sone K, Piao L, Ueda K, et al.The oncogenic polycomb histone methyltransferase EZH2 methylateslysine 120 on histone H2B and competes ubiquitination. Neoplasia2013;15:1251–61.

35. O'Regan L, Fry AM. The Nek6 and Nek7 protein kinases are required forrobust mitotic spindle formation and cytokinesis. Mol Cell Biol2009;29:3975–90.

36. Fry AM, O'Regan L, Sabir SR, Bayliss R. Cell cycle regulation by the NEKfamily of protein kinases. J Cell Sci 2012;125:4423–33.

37. Curtin JF, Cotter TG. JNK regulates HIPK3 expression and promotesresistance to Fas-mediated apoptosis in DU 145 prostate carcinoma cells.J Biol Chem 2004;279:17090–100.

38. Boivin A, Hanot M, Malesys C, Maalouf M, Rousson R, Rodriguez-Lafrasse C, et al. Transient alteration of cellular redox buffering beforeirradiation triggers apoptosis in head and neck carcinoma stem and non-stem cells. PLoS ONE 2011;6:e14558.

39. Moniz L, Dutt P, Haider N, Stambolic V. Nek family of kinases in cell cycle,checkpoint control and cancer. Cell Div 2011;6:18.

40. Patil M, Pabla N, Ding HF, Dong Z. Nek1 interacts with Ku80 to assistchromatin loading of replication factors and S-phase progression. CellCycle 2013;12:2608–16.

41. Helin K,Dhanak D.Chromatinproteins andmodifications as drug targets.Nature 2013;502:480–8.

42. Wagner T, Jung M.New lysinemethyltransferase drug targets in cancer. NatBiotech 2012;30:622–3.

43. Wee S, Dhanak D, Li H, Armstrong SA, Copeland RA, Sims R, et al.Targeting epigenetic regulators for cancer therapy. Ann N Y Acad Sci2014;1309:30–6.

44. Zagni C, Chiacchio U, Rescifina A. Histone methyltransferase inhibitors:novel epigenetic agents for cancer treatment. Curr Med Chem 2013;20:167–85.

45. McCabe MT, Ott HM, Ganji G, Korenchuk S, Thompson C, Van Aller GS,et al. EZH2 inhibition as a therapeutic strategy for lymphoma with EZH2-activating mutations. Nature 2012;492:108–12.

46. Qi W, Chan H, Teng L, Li L, Chuai S, Zhang R, et al. Selective inhibition ofEzh2 by a small molecule inhibitor blocks tumor cells proliferation. ProcNatl Acad Sci U S A 2012;109:21360–5.

47. Nimura K, Ura K, Shiratori H, Ikawa M, Okabe M, Schwartz RJ, et al. Ahistone H3 lysine 36 trimethyltransferase links Nkx2–5 to Wolf-Hirsch-horn syndrome. Nature 2009;460:287–91.

48. Kim S, Kim S, Rhee K. NEK7 is essential for centriole duplication andcentrosomal accumulationof pericentriolarmaterial proteins in interphasecells. J Cell Sci 2011;124:3760–70.

49. O'Regan L, Blot J, Fry AM. Mitotic regulation by NIMA-related kinases.Cell Div 2007;2:25.

50. Bertran MT, Sdelci S, Regue L, Avruch J, Caelles C, Roig J. Nek9 is a Plk1-activated kinase that controls early centrosome separation throughNek6/7and Eg5. EMBO J 2011;30:2634–47.






51. Wang R, Song Y, Xu X,Wu Q, Liu C. The expression of Nek7, FoxM1, andPlk1 in gallbladder cancer and their relationships to clinicopathologicfeatures and survival. Clin Transl Oncol 2013;15:626–32.

52. de Carcer G, Perez de Castro I, Malumbres M. Targeting cell cycle kinasesfor cancer therapy. Curr Med Chem 2007;14:969–85.

53. Malumbres M, Barbacid M. Cell cycle kinases in cancer. Curr Opin GenetDev 2007;17:60–5.

54. Kokuryo T, Senga T, Yokoyama Y, Nagino M, Nimura Y, Hamaguchi M.Nek2 as an effective target for inhibition of tumorigenic growth andperitoneal dissemination of cholangiocarcinoma. Cancer Res 2007;67:9637–42.

55. Westwood I, ChearyDM,Baxter JE, RichardsMW, vanMontfort RL, FryAM,et al. Insights into the conformational variability and regulation of humanNek2 kinase. J Mol Biol 2009;386:476–85.


Saloura et al.




2015;13:293-304. Published OnlineFirst October 3, 2014.Mol Cancer Res Vassiliki Saloura, Hyun-Soo Cho, Kazuma Kiyotani, et al. and NeckNIMA-Related Kinase-7 in Squamous Cell Carcinoma of the Head WHSC1 Promotes Oncogenesis through Regulation of

Updated version

10.1158/1541-7786.MCR-14-0292-Tdoi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2014/10/04/1541-7786.MCR-14-0292-T.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/13/2/293.full#ref-list-1

This article cites 55 articles, 14 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/13/2/293.full#related-urls

This article has been cited by 4 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/13/2/293To 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-14-0292-T

http://mcr.aacrjournals.org/content/suppl/2014/10/04/1541-7786.MCR-14-0292-T.DC1

http://mcr.aacrjournals.org/content/13/2/293.full#ref-list-1

http://mcr.aacrjournals.org/content/13/2/293.full#related-urls

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/13/2/293


Documents

WHSC1 Promotes Oncogenesis through Regulation of NIMA ... · Regulation of NIMA-Related Kinase-7 in Squamous Cell Carcinoma of the Head and Neck Vassiliki Saloura1, Hyun-Soo Cho1,2,