concurrent alterations in tert kdm6a, and the brca …tert promoter analysis the tert promoter was...

Biology of Human Tumors

Concurrent Alterations in TERT, KDM6A, and the BRCAPathway in Bladder Cancer

Michael L. Nickerson1, Garrett M. Dancik2, Kate M. Im1, Michael G. Edwards3, Sevilay Turan1, Joseph Brown4,ChristinaRuiz-Rodriguez1,CharlesOwens2, JamesC.Costello5, GuangwuGuo6, Shirley X. Tsang6, Yingrui Li6,Quan Zhou6, Zhiming Cai7, Lee E. Moore8, M. Scott Lucia9, Michael Dean1, and Dan Theodorescu2,5,10

AbstractPurpose:Genetic analysis of bladder cancer has revealed a number of frequently altered genes, including

frequent alterations of the telomerase (TERT) gene promoter, although few altered genes have been

functionally evaluated. Our objective is to characterize alterations observed by exome sequencing and

sequencing of the TERT promoter, and to examine the functional relevance of histone lysine (K)–specific

demethylase 6A (KDM6A/UTX), a frequently mutated histone demethylase, in bladder cancer.

Experimental Design:We analyzed bladder cancer samples from54U.S. patients by exome and targeted

sequencing and confirmed somatic variants using normal tissue from the same patient. We examined the

biologic function of KDM6A using in vivo and in vitro assays.

Results: We observed frequent somatic alterations in BRCA1 associated protein-1 (BAP1) in 15% of

tumors, including deleterious alterations to the deubiquitinase active site and the nuclear localization

signal. BAP1 mutations contribute to a high frequency of tumors with breast cancer (BRCA) DNA repair

pathway alterations andwere significantly associated with papillary histologic features in tumors. BAP1 and

KDM6A mutations significantly co-occurred in tumors. Somatic variants altering the TERT promoter were

found in69%of tumors butwerenot correlatedwith alterations in other bladder cancer genes.We examined

the function of KDM6A, altered in 24% of tumors, and show depletion in human bladder cancer cells,

enhanced in vitro proliferation, in vivo tumor growth, and cell migration.

Conclusions: This study is the first to identify frequent BAP1 and BRCA pathway alterations in bladder

cancer, show TERT promoter alterations are independent of other bladder cancer gene alterations, and show

KDM6A loss is a driver of the bladder cancer phenotype. Clin Cancer Res; 20(18); 4935–48. �2014 AACR.

IntroductionBladder cancer is the fifth most common cancer world-

wide (1), with 386,300 new cases and 150,200 deaths in

2008 (1). Bladder cancer is classified into 2 types that arethought to be driven by mutations in different sets of genes(2). The low-grade and nonmuscle invasive (NMI) form issuccessfully treated with transurethral surgery and intrave-sical immunotherapy (2) and accounts for �80% of allcases. Alterations associated with this form include muta-tion or overexpression of HRAS, fibroblast growth factorreceptor 3 (FGFR3), and the KDM6A; refs. 2–4. The remain-ing 20% of cases are muscle invasive (MI) and, despiteaggressive treatment with cystectomy, have a 5-year survivalrate of <50% (2). Overexpression or mutation of humanepidermal growth factor receptor-2 (HER2), epithelial growthfactor receptor (EGFR), TP53, andRB1 are associated withMIbladder cancer (2–4).

Next-generation sequencing (NGS) of the exomes of9 bladder cancer tumors revealed frequent somatic altera-tions in 54 genes, including 16 new bladder cancer genesmutated in �5% of cases (3). Eight were genes encodingchromatin modifying/remodeling enzymes, AT rich inter-active domain 1A (ARID1A), CHD6, CREBBP, EP300, MLL,MLL3, NCOR1, and KDM6A. Recently, analysis of 99 blad-der cancer tumors by whole genome, whole exome, andtranscriptome NGS (5) revealed additional altered genes,

1Cancer and Inflammation Program, National Cancer Institute, NationalInstitutes of Health, Frederick, Maryland. 2Department of Surgery, Univer-sity of Colorado, Aurora, Colorado. 3Division of Pulmonary Sciences andCritical Care, Department of Medicine, University of Colorado, Denver,Aurora, Colorado. 4Ingenuity, Redwood City, California. 5Department ofPharmacology, University of Colorado, Aurora, Colorado. 6BGI-Shenzhen,Shenzhen, China. 7ShenzhenSecond People's Hospital, Shenzhen, China.8Division of Cancer Epidemiology and Genetics, National Cancer Institute,National Institutes of Health, Rockville, Maryland. 9Department of Pathol-ogy, University of Colorado, Aurora, Colorado. 10University of ColoradoComprehensive Cancer Center, Aurora, Colorado.

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

Current address for G.M.Dancik: Department ofMathematics andComputerScience, Eastern Connecticut State University, Willimantic, Connecticut.

Corresponding Author: Dan Theodorescu, University of Colorado Com-prehensive Cancer Center, MS F-434, 13001 East 17th Place, Aurora, CO80045. Phone: 303-724-3152; Fax: 303-724-3162; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-14-0330

�2014 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org 4935

on February 1, 2021. © 2014 American Association for Cancer Research. clincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

including stromal antigen 2 (STAG2) and ESPL1, encodingproteins involved in the spindle checkpoint of the cell cycle,and a recurrent fusion involving FGFR3 and TACC3, anoth-er spindle checkpoint gene (6).

All tumors analyzed to date by NGS were from Chinesepatients. Thus, findings may reflect the regional ethnicity ofthe patients, differences in exposure to damaging agents, orlifestyle factors. Here we determine by exome sequencingof 14 tumors if novel alterations could be found in a cohortof non-Asian patients diagnosed with bladder cancer in theUnited States.Given the recent report ofTERT genepromotermutations in bladder cancer (7), we examined the TERTpromoter using targeted sequencing in the 14 tumors and inan additional 40 tumors. Healthy, noncancerous tissueDNAwas available to determine variants that were tumor specific.Finally, given the frequent alteration of KDM6A found bythis study and in Chinese patient bladder cancer tumors,we examined whether KDM6A expression has functionalin vitro, in vivo, and clinical prognostic impact.

Materials and MethodsHuman subjects, tumor and normal tissue samples,and cell lines

The study was approved by the University of ColoradoDenver (UCD), Institutional Review Board (ColoradoMul-tiple Institutional Review Board, CB F490, protocols 10-1365 and 09-913). Clinical details on patients and tumorsused for exome and targeted sequencing are summarized inSupplementary Table S1. Tissue samples were snap frozenin liquid nitrogen and stored at �80�C until DNA wasisolated using proteinase K digestion followed by phe-nol/chloroformextraction.Hematoxylin and eosin–stained(H&E) sections of tumors were prepared by standard meth-ods and evaluated by a board certified genitourinary pathol-ogist (S.M. Lucia). Only tumor samples with �85% puritywere processed for DNA. DNA was isolated from normaladjacent bladder tissue that was manually reviewed by apathologist (S.M. Lucia) to confirm samples were free ofurothelial carcinoma in situ and low-grade urothelial dys-plasia. Functional experiments utilized T24T and MGHU3,2 human bladder cancer cell lines that have been usedextensively were selected. PCR and Sanger sequencing ofT24T DNA revealed a KDM6A homozygous c.G2683T(p.E895X, NM_021140) introducing a stop codon. Thistruncates 506 C-terminus amino acid (aa) (36% of 1401AA total), including the catalytic jumonji C domain (aa

1099–1241). Because T24T was derived from a femalepatient, the wild-type (WT) copy of KDM6A was also lost.The KDM6A protein coding sequences and splice junctions(SJ) were WT in MGHU3. T24T cells were cultured inDulbecco’s Modified Eagle Medium/F12 with 2.5 mmol/LL-glutamine solution adjusted with 2.4 g/L sodium bicar-bonate and 5% fetal bovine serum. MGHU3 were culturedin minimal essential medium with 2 mmol/L L-glutamine,Earle’s balanced salt solution (2.2 g/L sodiumbicarbonate),and 10% fetal bovine serum. Cell lines have been tested andauthenticated by single nucleotide polymorphism analysis.

Exome capture, sequencing, and analysisBladder tumor genomic DNA (3 mg, quantitated by

fluorometer and agarose gel) was fragmented and theexome captured using probes for�180,000 protein-codingexons and microRNA loci based on curated genes in theconsensus coding sequence database (SureSelect HumanAll Exon Kit; Agilent). Libraries were sequenced on a HiSeq2000 platform (Illumina) and base calls on paired-end,100 bp reads were generated using the Genome AnalyzerPipeline, v. 1.3, and standard parameters. Additional detailsare available in the Supplementary Methods.

Variant validationSelected variants identified in tumor exomes were exam-

ined in DNA from the tumor and normal adjacent bladdertissue. Primers were designed using Primer 3 and ExonPri-mer (UCSC Genome Browser) as described in the Supple-mentary Methods. PCR utilized FastStart reagents (Roche)and GeneAmp 9700 thermal cyclers (ABI) at a 58�C anneal-ing temperature for 30 cycles preceded by a 10 cycle, 5�Ctouchdown. PCRproductswere evaluated by2%agarose gelelectrophoresis and sequenced using Big Dye v.3.1 reagents(ABI) and a 3730 Genetic Analyzer (ABI). Mutation Sur-veyor (Softgenetics) and Sequencher, v.4.8 (GeneCodes)were used for sequence analysis.

Copy number assessmentCopy number variation (CNV) was assessed using the

relative signal intensity (RSI) of somatic and heterozygousgermline variants in Sanger chromatograms (8, 9), which isan estimate ofmutant/mutantþWTnucleotide peakheightsin sequencing chromatograms. RSI are semiquantitativemeasures of relative allelic fractions in heterogeneous can-cer samples comparable to allele frequencies obtained aftersubcloning. High RSI scores indicate allelic imbalancebecause of somatic loss of WT alleles or amplification ofoncogenic mutant alleles in cancer samples.

Annotation of CRM-DA proteins and BAP1Chromatin remodeling-DNA associated (CRM-DA) pro-

teins were annotated using UniProtKB/Swiss-Prot, Gene-cards, PhosphoSitePlus, Modbase, and The Protein ModelPortal databases for selected isoforms: 729 aa BAP1 adaptedfrom ref. 10; ARID1A,2285 aa;KDM6A, 1401 aa; and STAG2,1231 aa. Phosphorylation sites are recurrent observationsby mass spectrometry (PhosphoSitePlus). The BAP1 ribbon

Translational RelevanceSequencing of bladder cancers reveals hitherto unap-

preciated mutations in BAP1 and BRCA pathway genes,while finding that somatic TERT promoter alterationswere independent of somatic alterations in other genes.These data suggest novel strategies with poly (ADP-ribose) polymerase inhibitors combined with DNAdamaging agents may be effective in specific patients.

Nickerson et al.

Clin Cancer Res; 20(18) September 15, 2014 Clinical Cancer Research4936



diagram includes residues 5-291 (UniProtKBQ92560) basedon the 2.5 angstrom crystal structure of homologous ubiqui-tin carboxyl-terminal esterase L3 (UCHL3) (ModBase tem-plate code 3ihrA; ref. 11) and was viewed using The ProteinModel Portal. Aligned BAP1 nuclear localization signal (NLS)residues include orthologous proteins: human, AAH01596;mouse, NP_081364; rat, NP_001100762; and Xenopus,NP_001008206 (adapted from ref. 12).

TERT promoter analysisThe TERT promoter was analyzed similar to a recent

publication (7). Additional primers were designed fornested PCR using 1 mL of a 1:10 dilution of the first roundPCR and sequencing. Forward nested, GATTCGACCT-CTCTCCGCTG; reverse nested, CCTCGCGGTAGTGGCT-GCGC. Transcription factor (TF) binding sites (TFBS) wereanalyzed using Tfsearchv1.3 (http://www.cbrc.jp/research/db/TFSEARCH.html) and 15 bases on either side of eachvariant. Multiple variants in the same sample were testedindividually and in combination.

Gene expression datasets and network analysisGene expression datasets are described in Supplementary

Information. The genes with somatic mutations confirmedby Sanger sequencing in a recent work (5) and the currentmanuscript were imported into Ingenuity Pathway Analysis(IPA; Ingenuity Systems) for bioinformatic analysis. IPAwas used to group these mutated genes into gene-limitednetworks (35 genes maximum) based on evidence of director indirect relationships between genes according to the IPAKnowledge Base. The IPA network algorithm seeks to max-imize the interconnectivity within a group of selected genesand scores networks based on a right tailed Fisher’s exact testthat calculates the probability that the given relationshipscan be explained by a randommodel. The networks do notinclude all possible relationships for eachmember, becauseof size constraints placed on the network, and specific genesmay appear in multiple networks. We identified differen-tially expressed genes in bladder cancer (relative to normalbladder samples) using publicly available microarray data-sets. P values were calculated by the nonparametric Wil-coxon rank-sum test and each P value adjusted to obtain thefalse discovery rate. The expression patterns of genes con-sistently up- or downregulated in 2 or more cohorts (falsediscovery rate < 5%) were then overlaid on the networksconstructed by IPA from the mutation data.The comutational network was based on a gene matrix

table where total mutations are found at a gene’s intersec-tion with itself and gene comutations in the remainingcolumns/rows. From this table we parse into Cytoscape(www.cytoscape.org) compatible, tab-delimited tablesusing Python. The output from the script is imported intoCytoscape 3.0.2, generating the bladder cancer interactionnetwork. For the network in Fig. 4B, we removed genes withfewer than 5 total mutations and less than 2 comutationinteractions bringing the count from 423 genes with 9,317interactions down to 46 genes with only 596 interactions.Edge-weights based on comutation rates between genes

were used to organize the network, pulling genes withhigher rates of comutation closer together. Node size cor-responds to a gene’s total mutations and color to a node’srelative centrality score (genes in the center, red; periphery,green). The raw count data, script, script output, additionsamplemetadata, and saved Cytoscape session are available(https://github.com/brwnj/interaction_network).

A KDM6A mutation signature and bladder cancerdiagnosis

We previously identified a 26-gene signature reportingon KDM6A mutation status that consists of genes differen-tially expressed between KDM6A WT and mutant tumors(13). The ability of this signature to discriminate betweennormal urothelial and bladder cancer cells was assessed bycalculating a signature score. Microarray probes were match-ed to gene symbols according to Affymetrix or Illuminaannotation. When multiple probes mapped to a singlesignature gene, the probe with the highest mean expressionwas selected. The expression values of all genes were nor-malized tohavemean0 and variance 1 across all samples in acohort. The signature score is the sum of the (normalized)expression of all genes upregulated in mutants minus thesumof the expressionof all genes downregulated inmutants.Therefore, a high signature score indicates a sample is moresimilar to a tumor with a mutated gene. The relationshipbetween signature score and cancer diagnosis (tumor vs.normal) was evaluated by the area under the receiver oper-ating characteristic curve (AUC), with AUC > 0.50 indicatingthat the signature score is higher in tumor samples (com-pared with normal urothelial cells). TheWilcoxon rank-sumtest was used to calculate a P value comparing AUC to 0.50(i.e., what would be predicted by a random model).

KDM6A constructs, transfections, RT-PCR, andfunctional assays

Short hairpin RNA (shRNA) targeting human KDM6A(shKDM6A) and a scrambled shRNA control (shCTL) innontargeting plasmid pLKO.1-puro (Sigma-Aldrich) wasused to examine KDM6A in MGHU3 cells. Mammalianexpression vectors allowing reexpression of KDM6A(FLAG-KDM6A) or empty control vector (FLAG) were con-structed using a modified Gateway Multisite Recombina-tion system (Life Technologies) and transfected into T24Tcells. KDM6A depletion and reexpression was validated byquantitative reverse transcriptase-PCR (qRT-PCR). Anchor-age-dependent and -independent growth and cellmigrationwas assessed as previously described (14). For in vivo assess-ments, 5-week-old male NCrnu/numice were injected with2� 106MGHU3 cells stably expressing shKDM6A or shCTLand were assessed as described in the SupplementaryMethods.

Statistical analysisThe Pearson’s correlation coefficient was used to assess

pairs of altered genes in 54 U.S. tumors (ARID1A, BAP1,STAG2, KDM6A, and TERT). We used a Fisher exact test orlogistic regression for univariate analysis and logistic

BAP1 and KDM6A Bladder Cancer Mutations

www.aacrjournals.org Clin Cancer Res; 20(18) September 15, 2014 4937



regression for multivariate analysis to investigate associa-tion of clinical factors with gene mutation status. Nonsy-nonymous, coding insertions and deletions (indels), UTR,and intronic variants within 10 bp of the intron/exonborder were included in the mutation correlation and theclinical association analyses. Statistical analysis was per-formed using R software (http://cran.r-project.org).

ResultsExome sequencing of 14 bladder tumors from U.S.patients

We subjected 14 primary urothelial bladder tumors toexome sequencing to characterize genomic variants in U.S.

patients with bladder cancer (Supplementary Table S1).This yielded an average of >7.6 gigabases per sample andan average mean NGS read depth of 92� (SupplementaryTable S2). Grouping NGS-predicted variants by the type ofnucleotide change revealed significant enrichment in C>TandG>A changes, similar to previous reports in bladder andother cancers (Fig. 1A, top; refs. 3 and 15). We identifiedputative somatic variants by removing variants observed at afrequency >1% in the 1000Genomes Project and variants inknown segmental duplications (16). This yielded 9,312missense, 348 nonsense, 216 SJ, and 653 coding inser-tion/deletion (indel) candidate variants.

We analyzed matched tumor and normal tissue DNA(Supplementary Table S1) using PCR and Sanger sequencing

Figure 1. Altered bladder cancergenes. A, 41 bladder cancer geneswith somatic nonsynonymous andSJalterations. Tophistogram,NGSpredicted variants by nucleotidechange for NMI (stage Ta and T1)and MI (stage T2–T4) tumors.Central panel, genes (left) andCRM-DA functions (teal color, seetext); asterisk, a proven or likelyoncogene; black dot, tumor has�1somatic mutation; histogram, rightside, percent of tumors withsomatic gene mutation; bold orhatched box, CNV; bottom, tumorIDs (1–14). B, somatic alterations in5 CRM-DA bladder cancer genes(top) in 40 validation tumors (leftside), annotated as in A.

Nickerson et al.




to confirmNGS-predicted variants and identify somatic fromgermline alterations.Nonsynonymous and SJ variants select-ed for validation were in genes with >1 NGS-predictedalteration, knowncancer and related genes, geneswith clearlydeleterious variants (frameshift and nonsense), and severalrandomly selected genes. We examined 316 variants andconfirmed 228 (72%), 112 were somatic (SupplementaryTable S3) and 116 were germline, including 43 novel germ-line variants (Supplementary Table S4).

Somatic alterations in 4 novel bladder cancer genesValidated somatic variants were observed in 67 genes,

including 10 indels, 22 nonsense, 78 missense, and 2variants in SJs. Twenty genes were mutated in >1 tumor,including genes known to be mutated in bladder cancersuch as FGFR3 (17), TP53 (18), and TSC1 (ref. 19; Fig. 1A,middle; Supplementary Table S3). Eight genes were alteredin �3 tumors and are referred to as frequently mutatedgenes (FMG), including previously identified bladder can-cer genes, KDM6A and ARID1A, altered in 4 and 3 tumors,respectively (3). Twenty additional genes with a somaticalteration are also included in Fig. 1A (bottom) based onprevious identification as a cancer or cancer-related gene.This selection is further supported by the types of somaticalterations. For example, MLL3 is altered in tumor 6 by asomatic c.C8695T (p.Q2899X) and was altered in 25% ofTCGAbladder cancer tumors, and EP300 is altered in tumor2 by a somatic c.G3052T (p.E1018X) and was altered in17% of TCGA bladder cancer tumors (20, 21).To our knowledge, 4 altered genes are novel bladder

cancer FMGs when compared with other studies (5):BAP1 located on chromosome (chr) 3p21, chromodomainhelicase DNA binding protein 1 (CHD1, chr. 5q21), chro-modomain helicase DNA binding protein 1-like (CHD1L, chr.1q21), and GCN1 general control of amino-acid synthesis 1-like 1 (GCN1L1, chr. 12q24). These genes are of particularinterest as they encode proteins that have distinct rolesin chromatin remodeling (Fig. 1A, gene names in teal).These FMGs increase the number of bladder cancer genesencoding proteins with CRM-DA functions functions. Anaverage of 4.7 somatic sequence changes to CRM-DAgenes occurred in each tumor, range 2–13, indicating asubstantial contribution from altered CRM-DA genes tobladder cancer.

Somatic CHD1, CHD1L, and GCN1L1 alterations inbladder cancerFour alterations inCHD1were confirmed as somatic (Fig.

1A). Recurrent somatic C>T variants in tumors 3 and 5altered a highly conserved proline 1684 to serine (p.P1684S) and a somatic G>T in tumor 7 introduced anonsense codon (p.E272X), truncating most of the 1,710amino acid (aa) protein (NM_001270). CHD1 encodes aCHD family-related protein that alters chromatin structureand influences transcription via SNF2-related helicase/ATPase and chromo (chromatin organization modifier)domains (22). Somatic alterations were observed in genesrelated to CHD1, including a truncation (p.Q728X) of

CHD1L and a p.Q1833E in CHD4. Altered expression ofCHD1L has been implicated in bladder cancer (23) andgermline alterations are associated with congenital anom-alies of the kidney and urinary tract (24), suggestingCHD1Lis a valid bladder cancer gene. A total of 4 of 14 tumors(29%) showed somatic alteration of CHD genes. CHD2,CHD4, CHD5, and CHD6were altered in 4%, 3%, 4%, and7%, respectively, of Chinese patient tumors in a total of 15of 99 tumors (15%; ref. 5).

Four somatic alterations of GCN1L1 were observedand these may lead to loss of function, as indicated bya single-base deletion in tumor 10 causing a frameshiftand truncation after codon 1777 (2671 aa full length,NM_006836; Fig. 1A). Two somatic, synonymous changeswere observed previously in 99 Chinese patient tumors(5). The current data indicate, to our knowledge, the firstevidence of somatic nonsynonymous GCN1L1 altera-tions in bladder cancer. GCN1L1 may link AA metabolismto chromatin remodeling and gene expression throughroles in regulating GCN2 kinase and as a transcriptionalcoregulator in mediator complexes (25). Future analysesof CHD1, CHD1L, CHD-related genes, and GCN1L1 inadditional bladder cancer tumors are planned to deter-mine accurate somatic mutation frequencies in bladdercancer.

BAP1 and BRCA pathway defects in bladder cancerExome sequencing revealed 4 somatic missense variants

in BAP1 in 3 tumors (Fig. 1 and Supplementary Table S3).To confirm BAP1 as a bladder cancer gene, PCR and Sangersequencing was performed on an additional 40 bladdertumors with matched normal tissue from U.S. patients(Supplementary Table S1). Five additional somatic altera-tionswere observed and, in total,BAP1was altered in8of 54tumors (15%).

The likely functional effects of BAP1 mutations are pre-dominantly loss of function, such as introduction of a stopcodon (p.E257X) and 2 somatic changes altering histidine169 to an arginine (p.H169R) or a glutamine (p.H169Q; Fig.2A). Histidine 169 has a critical function as the predictedproton donor residue in the ubiquitin hydrolase active site(Uniprot DB). A comparative 3-dimensional (3D)model ofBAP1 was generated at ModBase and the structure confirmsthat p.H169 occupies a critical position directly across fromthe active sitenucleophile, cysteine91 (Fig. 2B, adapted fromModBase; ref. 11). Both substitutions alter the active site:histidine to an arginine introduces a significantly longer sidechain and histidine to a glutamine introduces an unchargedfor a positively charged residue. BAP1 was also altered in 2tumors by recurrent, nonconservative substitutions of aglycine for arginine 718 (p.R718G). Arginine 718 is 1 of 4charged arginine residues in a highly conserved NLS con-sisting of AAs 717-722 (Fig. 2C; refs. 11 and 12). Previouslypublished alteration of arginines to alanines in the NLS,including p.R718, completely blocked cytoplasmic to nucle-ar shuttling of the protein and abolished BAP1 deubiquiti-nase activity. These observations are consistent with delete-rious somatic alterations targeting a tumor suppressor in





Nickerson et al.




bladder cancer, similar to previous observations of BAP1 inother cancers (26).BAP1 protein binds breast cancer 1 (BRCA1; ref. 27) and

loss of function BAP1 mutations may target the BRCADNA repair pathway, including BRCA1 and breast cancer2 (BRCA2) (28). This is supported by 16 somatic altera-tions in BAP1 and 4 BRCA DNA repair pathway genes in 9of 14 tumors (64%; Fig. 1A and Supplementary Table S3).A tenth tumor was homozygous for a truncating BRCA2germline allele, totaling a surprising 10 of 14 (71%)bladder cancer tumors with BRCA pathway alterations(Supplementary Table S4). Eight tumors were altered byclearly deleterious mutations, including 4 truncations,and approximately 1 pathway gene per tumor was inacti-vated (6 tumors). Ataxia telangiectasia mutated (ATM)was altered by 6 somatic sequence changes in 4 tumors,including alteration of a conserved SJ (c.5497-1G>C)and a p.Q1636X in tumors 9 and 12, respectively. BRCA1was altered by somatic missense p.P364A and p.S1286T intumors 3 and 8, respectively. BRCA2 was altered by ahomozygous germline p.K3326X and a somatic p.Q3066Xin tumors 7 and 11, respectively. Finally, partner andlocalizer of BRCA2 (PALB2) was altered by somaticp.S535C and p.Q613X in tumors 6 and 10, respectively.These alterations define a majority of bladder cancertumors with a common DNA repair deficiency that mightbe exploited by targeted therapies such as poly (ADP-ribose) polymerase (PARP) inhibitors (28).BAP1 mutation in kidney tumors is associated with

rhabdoid morphology (26) and we examined this pheno-type in bladder cancer tumors. Blinded to the BAP1 status,one author (S.M. Lucia) classified a set of 16 H&E-stainedtumor sections, 8 with a BAP1 mutation and 8 WT. Thisrevealed papillary features in 6 tumors (Fig. 2D and Sup-plementary Table S5), of which 5 had a somatic BAP1alteration. Three tumors with BAP1 mutations did notexhibit papillary features in the tumor section analyzed,although additional tumor sections were not available toexamine other regions of these tumors. Our data suggestthat papillary features in some bladder tumors are associ-ated with somatic alteration of BAP1.We compared BAP1 somatic mutations in 54 U.S. patient

tumors to 99 Chinese patient tumors (5) and found BAP1was altered at a significantly greater frequency in the U.S.cohort (15% vs. 1%, P ¼ 0.003). This remained marginallysignificant after accounting for differences in tumor gradeand stage (P¼ 0.037). To determine if other CRM-DA genes

were altered at different frequencies between these cohorts,we sequenced the 3 most frequently mutated CRM-DAbladder cancer genes, ARID1A, KDM6A, and STAG2(3, 5, 29) in the same set of 40 validation tumors (Fig.2A and Supplementary Table S1). ARID1A, KDM6A, andSTAG2 were altered by somatic sequence changes in 17%,24%, and 17%, respectively, of 54 U.S. patient tumors(Supplementary Table S3). Cumulatively, the 4 genes werealtered in 25 of 54 (46%) of U.S. tumors. Guo and collea-gues (5) observed ARID1A, KDM6A, and STAG2 alterationsin 15%, 30%, and 11%, respectively, of 99 Chinese patienttumors. The somatic mutation frequencies of ARID1A,KDM6A, and STAG2 were similar and only BAP1 wassignificantly different. Thus,BAP1 seems tobe preferentiallyaltered in U.S. patients with bladder cancer, perhapsbecause of patient ethnicity, exposure, or lifestyle factors.

Copy number alterations in bladder cancer genesExome and Sanger sequencing revealed allelic imbalance

at sites of somatic and heterozygous germline sequencevariants in tumor versus normal tissue DNA, which couldbe classified as copy-number variation CNV (Fig. 1, boldboxes). Somatic and heterozygous germline variants withhigh RSI scores � 0.7 (mutant allele signal/WT þ mutantallele signal) are reliable indicators of CNVs in cancer (8, 9).The 41 bladder cancer genes in Fig. 1A were altered by 27CNVs with an average of 2 per tumor, range 0 to 5. Thesegeneswere altered by100 somatic sequence changeswith anaverage of 7.1/tumor, range 4 to 16. In total, an average of9.1 somatic alterations/tumor, range 4 to 20,were observed,which were predominantly sequence changes (78%) com-pared with CNVs (21%). Two or more alterations poten-tially affecting both alleles of a bladder cancer gene weredetected in 30 genes with an average of 2.1 multiply-alteredgenes/tumor, range 0 to 4. Similar results were observed for5 genes analyzed in B (Fig. 1B). Somatic alterations (n¼ 77)were composed of 64 sequence changes (83%) and 13CNVs (17%), and 17 tumors (43%) displayed at least 1gene with >1 alteration. These data indicate homozygousloss of function of tumor suppressor genes or homozygousactivation of oncogenes affects 2 genes/tumor and indicatespotential high priority therapeutic targets.

Rare, deleterious germline alleles in bladder cancerExome and targeted sequencing of 54 tumors identified

11 deleterious germline variants resulting in nonsense,frameshift truncation, and SJ alterations. Seven alleles

Figure 2. Chromatin remodeling genes frequently altered in bladder cancer. A, locations of somatic alterations on annotated proteins. BAP1: ubiquitin Cterminal hydrolase domain (UCH, purple); HCF-1 binding motif (orange line); BRCA1 binding region (BRCA1 Domain, red); UCH37-like domain (ULD, gray);NLS, aa 717–722 (green); active site residues (vertical arrows). Adapted from refs. 26 and 10). ARID1A: LXXLL motif (black line); AT-interaction domain(ARID, red); HIC1 binding domain (yellow); Pfam homology domain (orange); B/C box (green); acetylation site (þ); region omitted in isoform B, aa 1366–1583(left bracket); GR binding domain (glucocorticoid receptor, right bracket). KDM6A: alanine rich region (green); 8 TPR repeats (black lines); Jumonji Cdomain (JmjC, red); iron (black triangles) and zinc (orange triangles) binding sites. STAG2: STAG domain (STAG, green); SCD domain (SCD, red); serine richregion (black line). In all proteins, recurrent phosphorylation sites are shown (�). B, a ribbon diagram of BAP1 residues 5–291 (UniProtKB Q92560) isbased on the crystal structure of homologous UCHL3-UbVME (11). Active site residues, H169 (proton donor, green) and cysteine 91 (nucleophile, yellow) areshown (inset at highermagnification). C, Alignment of NLS residues inBAP1orthologues. Residuenumbers are shown for humanBAP1 (top), p.R718mutation(asterisk). Adapted from ref. 12. D, papillary features in H&E-stained tumor sections from bladder tumors with somatic BAP1 mutations. Top: tumor 48,a low-grade papillary urothelial carcinoma; bottom: tumor 9, a low-grade urothelial carcinoma, inverted papillary type.





caused frameshift truncations, 3 introduced nonsensecodons, and 1 altered a highly conserved SJ nucleotide(bold, Supplementary Table S4). Two previously identifiedcancer genes were altered, BRCA2 (p.K3326X, rs11571833)and RB1 (p.Q846fs, novel). One novel allele in tumor9 truncated the FGF binding protein 1 (FGFBP1; c.6delG,p.K2fs), potentially indicating a new (likely rare) disease-associated allele in the FGF signaling pathway. Finally, wedescribe tumors with germline TERT promoter variants,including identical variants confirmed as either germlineor somatic in distinct tumors. These data suggest a greaterthan expected contribution to these supposed sporadiccases of bladder cancer from rare germline variants.

Somatic and germline alteration of the TERT promoterSomatic TERT promoter alterations were recently

reported in melanoma, glioma, and in small sample setsfrom other cancers, including bladder (7, 30, 31). Weexamined the proximal promoter of TERT in 54 bladdertumors and identified 93 single nucleotide substitutions in48 of 54 tumors, composed of 29 unique variants (Figs. 1, 3and Supplementary Table S6). Of these, 42 nucleotidesubstitutions, consisting of 20 unique variants, were germ-line in 30 of 54 tumors (56%) and 19 were novel(asterisk; Fig. 3B). The most common variant, c.�245T>C(rs2853669), was observed in 32 of 54 tumors (59%). Thisvariant was confirmed as germline in tumors from 23patients and as a novel somatic (tumor-specific) change in9 tumors. Novel variants at 2 other sites (c.�113C>T, andc.�212C>G and c.�212C>T) were also confirmed as bothgermline and somatic in distinct tumors.

We observed 51 somatic alterations in 37 of 54 tumors(69%) consisting of 11 unique variants, including 4 novelvariants, c.�111C>T, �113C>T, �133C>G, and �212C>G(asterisk, Fig. 3B and Supplementary Table S6). Eleventumors (20%) had >1 somatic alteration and 1 sample(tumor 17) had 4 variants in a cluster, c.�111, c.�112,c.�113, and c.�124. Three somatic variants previouslyreported in other cancers (7, 30, 31) were observed inmultiple bladder tumors, c.�124C>T in 27 tumors(50%), c.�245T>C in 9 tumors (17%) tumors, andc.�146C>T in7 tumors (13%), suggesting thesenucleotidesare recurrent targets in bladder cancer. Cytosine to thyminealterations comprised 39 of 51 (76%) of the observedsomatic TERT promoter alterations.

Similar to a previous report (31), a subset of somaticand germline variants were predicted to create newTFBS (Fig. 3C). Variant c.�113C>T creates a potential c-Relbinding site; variants c.�124C>T, c.�138C>T, andc.�139C>T potential Ets binding sites; variant c.�228A>G,a potential Sp1 binding site; variant c.278G>A, a potentialHSF2 binding site, and variant c.�284C>T, a potentialElk-1 binding site. In addition, variants altered sequencesshown to bind TFs by chromatin immunoprecipitationfollowed by NGS (ENCODE data tracks in the UCSCGenome Browser; Fig. 3A), including Pol2, TAF1, EGR-1,Myc, Max, Sin3A, and CTCF. Variants c.�110A>T,c.�111C>T, c.�112C>T, c.�113C>T, c.�126C>T, and

c.�133C>G altered multiple Sp1 binding sites causing anexpected loss of TF binding (32). Finally, 3 promotervariants altered CpG dinucleotides or neighboring residues(�1 bp), including a somatic c.�269G>A, germlinec.�306G>A, and germline c.�329C>T. These may affectregulation of promoter methylation.

Association of mutations with patient and tumorvariables

Next we examined 5 genes for which we had somaticmutation status in 54 tumors, ARID1A, BAP1, KDM6A,STAG2, and TERT, for correlations in somatic mutationsand for associations between mutation status and clinicalfactors. We found only BAP1 and KDM6Amutations werecorrelated (P ¼ 0.017; Supplementary Table S7). Somaticalteration of the TERT promoter occurred in tumors bothwith and without mutations in each of the other genes.These data indicate likely independent contributions ofthe mutations in these 5 genes to the tumor cell pheno-type. No significant associations between the mutationstatus of the 5 genes and clinical factors were observed(not shown).

Network analysis reveals the importance of KDM6A inbladder cancer

To better understand the potential interactions betweenaltered bladder cancer genes, we performed networkanalysis on genes with confirmed somatic mutations inthis study and a recent study of 99 Chinese bladder tumorexomes (5). This identified several robust interactingnetworks and biologic functional groups (SupplementaryTable S8). The highest scoring network (P < 10�68) wascomposed of 34 mutated genes known to bind UbiquitinC (UBC; Supplementary Fig. S1). The second highestscoring network (P < 10�46) contained proteins encodedby the third through the fifth most mutated genes in theChinese cohort (ARID1A, CREBBP, and EP300; ref. 5;Supplementary Fig. S2). Genes encoding enzymes asso-ciated with the deubiquitinylating pathway, includingBAP1, which is mutated in this study (BAP1,DUB,USP21,USP26, USP31, USP34, USP36, USP38, and USP48), andDNA methylation (ARID4B, CHD4, CHD3, MTA, andSIN3A), were well represented in this network. The thirdhighest scoring network (P < 10�42) contained TP53along with 4 genes involved in DNA methylation(DNMT1, DNMT3A, MLL3, and MLL5) and multiplegenes associated with chromosomal structural elementsand posttranslational modifiers of histones H3 and H4(Fig. 4A).

KDM6A is frequently mutated in both United States andChinese tumors and alterations are significantly correlatedwith somatic BAP1 alterations. KDM6A has limited pro-tein–protein interaction data, however, it could be includedin the third highest scoring network because this network isthe only 1 of the 3 that contains proteins known to interactwith KDM6A. Inclusion of KDM6A in a network differentfrom BAP1 makes sense because mutations in these genesare correlated and therefore they likely contribute different

Nickerson et al.




selective advantages to bladder cancer cells. We also includ-ed the RB1 tumor suppressor in this network, based onpreviously described relationships with TP53 and KDM6A(33). We found that KDM6A interacts with several other

proteins encoded by genes mutated in Chinese tumors.KDM6A directly binds to MLL3, CSPG4, and SMARCA4(BRG1; refs. 34 and 35) and regulates RB1 expression (33).MLL3, RB1, and SMARCA4 are known tumor suppressors

Figure 3. TERT promoter variants in bladder cancer. A, The proximal TERT locus in the UCSC Genome Browser is shown with HG19 genomic coordinates(Hg19), TERT isoforms (TERT transcripts), TFBS fromENCODEChIP-seq, the amplicon location (Amplicon BLAT), and aCpG-rich region (CpG island). B, TheTERT promoter amplicon (50, top left, to 30, bottom right) is shown with germline variants (above) and somatic variants (below) the genomic referencesequence (solid line). Variant annotation is based on the TERT coding strand. Frequencies of variants (n ¼ 54, in parentheses), novel variants (asterisk), andvariants in new TFBS (underlined, see details in Fig. 3C) are indicated. C, TFBS in the TERT promoter (bold horizontal line, center) are shown relative to

the transcription start site (TSS) and the protein-coding start codon (ATG):^, Sp1;Ο, E-box; and$, Ets TFBS. New TFBS created by novel (single underline)

and previously identified variants (double underline) are shown in the enlarged regions (boxed).





(36–38) whereas CSPG4 is thought to be oncogenic(35, 39). Using 3 publicly available microarray datasets(Supplementary Table S9), we identified genes that wereconsistently up- or downregulated in tumors in at least 2datasets (FDR < 5%) and highlighted themon themutationnetwork (Fig. 4A). In addition to being a known tumorsuppressor interacting with KDM6A, SMARCA4 expressionwas upregulated in tumors in all 3 cohorts. KDM6A indi-rectly interacts with several proteins consistently deregu-lated in bladder cancer, such as histone H3, which isrequired for EGF induced transformation (40). Networkanalysis revealed that KDM6A interacts with the products ofgenes that harbor bladder cancer driver mutations and withgenes consistently differentially expressed in bladder can-cer, suggesting its central importance as a driver of bladdercancer. We also used Cytoscape (www.cytoscape.org) toplot these genes in 3Dspaceusing the comutational rate andthe frequency of mutation as parameters in our model. ThisputsKDM6A as themost central andwell-connected gene inthis mutational network (Fig. 4B).

Functional relevance of KDM6A in bladder cancerKDM6A is frequently mutated in bladder cancer and is

important in signaling networks as we show above. There-fore, we developed a gene mutation signature for KDM6A(13), where a high KDM6A signature score corresponded to

a likely somatic mutation. We examined 3 patient cohorts(Supplementary Table S9) and found higher scores corre-sponded to NMI bladder cancer stratified patients with amore favorable prognosis. KDM6A signature scores arehigher in bladder tumor samples compared with normalurothelial cells (Fig. 5), indicating the signature likelydetected altered KDM6A-related pathways in bladder can-cer. Although these studies suggest thatKDM6A is a driver ofbladder cancer and has prognostic value, the direct func-tional role of KDM6A in bladder cancer has not yet beenexperimentally established.

Hence we determined the role of KDM6A in shaping thecancer cell phenotype using 2 well-established humanbladder cancer models, T24T (14) andMGHU3 (41). Thesetumor-derived cell lines were selected based on KDM6Amutation status. MGHU3 was WT, and T24T had a homo-zygous G>T nonsense mutation introducing a prematurestop codon (p.E895X, NM_021140, 1401 aa full length).The stop codon truncated 506 C-terminus AAs (36% of thefull length protein) including the catalytic jumonji Cdomain (see Materials and Methods). shRNA and cDNAconstructs were used to deplete andoverexpressWTKDM6Ain MGHU3 and T24T, respectively. The effect of thesetreatments on KDM6A expression was verified by qPCR(Fig. 6A). KDM6A depletion in MGHU3 enhanced anchor-age independent growth (P ¼ 0.04) and cell migration

Figure 4. Biologic and comutational connections of KDM6A to a network of genes with somatic mutations in bladder cancer. A, the third most significantnetwork (Fisher's exact test, P < 10�42) constructed by IPA from genes with confirmed somatic mutations in bladder cancer, including KDM6A and RB1(highlighted in blue). The number of independently identified mutations is listed above each respective gene. Evidence of direct (solid lines) or indirect(dashed lines) relationships between gene products is shown. Colored genes indicate decreased (green) or increased (red) gene expression in tumorscomparedwith normal bladder samples in at least 2of 3patient cohorts (FDR<5%).B, the networkwas created using the comutation rate and the frequencyofmutation to plot the mutations in 3D space. The size of each gene (node) is representative of the times it was mutated in different tumors. Theconnection lines (edges) in the network indicate whether the genes were foundmutated together in the same tumor, with the length and thickness of each lineshorter and wider with an increasing comutation rate. The nodes and edges of our model have been filtered to show only those genes found mutatedin 5 or more tumors and edges between genes that weremutated together more than once. Node color corresponds to a gene's relative centrality score, withgenes closer to the center of the network red and genes farther away turning to yellow then green at the periphery.

Nickerson et al.




(P < 0.001) but notmonolayer growth (Fig. 6B–D). In T24Tcells, KDM6A overexpression diminished anchorage inde-pendent growth (P ¼ 0.02) and cell migration (P ¼ 0.055)but notmonolayer growth (Fig. 6B–D). In vivo, depletion ofKDM6A in MGHU3 led to significantly enhanced tumorgrowth (P¼ 0.002), confirming the functional relevance ofintact KDM6A as a tumor growth inhibitor (Fig. 6D). Thedeleterious mutations observed in tumors in this and otherstudies (3, 5) and the functional data presented here indi-cate KDM6A is a frequentlymutated gene encoding a tumorsuppressor in bladder cancer.

DiscussionThis study contributes to the catalog of genes altered in

human bladder cancer by evaluating 54 tumors from U.S.patients. We identified 41 genes altered by somatic variantsthat may be relevant to disease, including 4 genes, CHD1,CHD1L, GCN1L1, and BAP1, not previously reported asmutated in bladder cancer. Interestingly, all 4 impact chro-matin remodeling and/or regulation.CHD-family gene alterations may play an important role

in bladder cancer because of their influence on chromatinstructure and transcription (22). Overexpression of CHD1Lhas been implicated in bladder cancer (23) and germlinealterations are associated with congenital anomalies of thekidney and urinary tract (24), indicating a likely role inbladder development. CHD1 somatic alterations have beenreported in prostate (42) and other cancers (21, 43) and areview of the COSMIC DB shows CHD1L is altered inendometrial (4%), stomach (2%), and lung cancers(2%). Alteration of GCN1L1 may target its role in AAmetabolism and excretion (44). The COSMIC DB showsGCN1L1 is mutated in 4% to 7% of endometrial, cervical,urinary tract (bladder), colon, lung, and stomach cancers.We plan to examine additional bladder cancer tumors toconfirm accurate mutation frequencies for CHD1, CHD1L,and GCN1L1.BAP1 acts to remodel chromatin via an ubiquitin

hydrolase catalytic function that removes ubiquitin fromhistone H2A. Somatic and germline BAP1 alterations

have been observed in melanoma, mesothelioma, andkidney cancers (45–47). Somatic alterations of BAP1define a subtype of clear cell kidney cancer characterizedby papillary features (26). Interestingly, review of bladdertumor samples revealed that papillary features were morecommon in bladder tumors with mutant BAP1, indicatingsimilar phenotypic characteristics between BAP1-mutantbladder and kidney tumors. BAP1 mutations are corre-lated with KDM6A alterations, suggesting these altera-tions may provide complementary advantages to tumorcells, similar to co-occurrence of PBRM1 and SETD2mutations in renal cell carcinoma (48). Interestingly,BAP1 mutations were complementary to alterations ingenes encoding proteins of the BRCADNA repair pathway(ATM, BRCA1, BRCA2, and PALB2). Components of thispathway are altered in Fanconi anemia, an inherited,autosomal recessive disease characterized by leukemiaand an increased susceptibility to multiple types of can-cer. The pathway regulates the cellular responses to DNAdamage and, via ATM, associated cell-cycle checkpoints.To our knowledge, frequent alteration of the BRCA path-way in bladder cancer has not been previously reported.Mutations in BRCA pathway genes may indicate tumorsthat are deficient in DNA repair and therefore vulnerableto DNA lesions created by chemotherapeutic drugs (28),especially if combined with PARP (49) inhibitors.

A high frequency of somatic TERT promoter alterationswas recently reported in several cancers (7, 30, 31).Although we found germline and somatic alterations inbladder cancer that had not been previously identified,somaticTERTpromoter alterationswere not associatedwithalterations in other bladder cancer genes or with tumorstage or grade. The variants observed may affect the TERTpromoter in several ways. TERT expression is known to beregulated by methylation and hydroxymethylation of pro-moter CpGs (50, 51). Several variants altered CpGs orneighboring nucleotides, suggesting potential disruptionof epigenetic regulation of TERT. In addition, ENCODEdata show germline and somatic variants altered DNAsequences known to bind TFs. TF binding may also bealtered by variants predicted to create new TF binding sites,

–10

–10

–50

510

0

Sig

nat

ure

sco

re

Sig

nat

ure

sco

re

–30

–20

–10

010

20

Sig

nat

ure

sco

re

Normal

CNUHAUC = 0.31, P < 0.005

A B CLindgrenAUC = 0.98, P < 0.005

MSKCCAUC = 0.84, P < 0.005

Tumor Normal Tumor Normal Tumor

1020

Figure 5. Relationship between aKDM6A signature score andbladder cancer. The KDM6Asignature score is plotted fornormal and tumor samples in A,CNUH (n ¼ 197). B, Lindgren(n ¼ 156) and C, MSKCC (n ¼ 129)cohorts. AUC, area under receivingoperator characteristics curve,withAUC>0.50 corresponding to asignature score higher in tumorthan normal samples. P valueswere calculated by the Wilcoxonrank-sum test.





such as c.�113C>T, which creates a consensus sequence forc-Rel. Hence, germline and somatic variants may work inconcert to alter TERT gene expression.

It is not clear why certain cancer genes vary in mutationfrequency across patient cohorts with the same disease.When compared with bladder tumors from Chinesepatients, BAP1 is mutated at a higher frequency in tumorsfrom U.S. patients, although the mutation frequencies ofKDM6A, ARID1A, and STAG2 were similar between thesecohorts. Thus, BAP1 and perhaps BRCA pathway genesmay represent genes targeted for alteration because ofspecific ethnic, lifestyle, or geographic differences. Givenits frequent and similar mutation frequency in both

patient cohorts, we sought to determine the functionaland prognostic role of KDM6A in human bladder cancer.In vitro and in vivo experiments examining KDM6A deple-tion and overexpression in bladder tumor cells support arole for KDM6A as a suppressor of tumor growth and cellmigration.

Taken together, our work provides further insights on thegenomic landscape of bladder cancer while providing cluesto possible prognostic (KDM6A status signature score) andcompanion biomarkers (BAP1 and BRCA pathway genemutations), suggesting novel therapeutic strategies (PARPinhibitors combined with DNA damaging agents) that maybe effective against bladder cancer.

Figure 6. KDM6A loss drives the bladder cancer phenotype. KDM6AWTMGHU3 cells were treated with shRNA targeting KDM6A (shKDM6A) or a scrambledshRNA (shCTL). These were compared with KDM6A-mutant T24T cells transiently reexpressing FLAG-tagged KDM6A (FLAG-KDM6A) or empty FLAGvector (FLAG). A, relative KDM6A mRNA expression by qPCR. B, anchorage-independent growth assay (n ¼ 6 wells/line). C, transwell cell migration(n¼ 4 wells/line). D, monolayer growth of cells (left) assessed by CYquant fluorescence assay (n¼ 4 wells/line) and subcutaneous tumor growth (right, n¼ 20mice/line). Results are shown as the mean� SEM. P values were calculated using a Student's t test on the final day of the assay. Assay details are describedor referenced in Materials and Methods.

Nickerson et al.




Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design:M.L. Nickerson, S.X. Tsang, L.E. Moore, M. Dean,D. TheodorescuDevelopment of methodology:M.L. Nickerson, M.G. Edwards, S.X. Tsang,Z. Cai, M. DeanAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): M.L. Nickerson, S. Turan, C. Owens, Y. Li, M.S.Lucia, M. DeanAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): M.L. Nickerson, G.M. Dancik, K.M. Im,M.G. Edwards, S. Turan, J. Brown, C. Owens, J.C. Costello, G. Guo, S.X.Tsang, Q. Zhou, L.E. Moore, M.S. Lucia, M. Dean, D. TheodorescuWriting, review, and/or revisionof themanuscript:M.L.Nickerson,G.M.Dancik, K.M. Im, J.C. Costello, L.E. Moore, M. Dean, D. TheodorescuAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): M.L. Nickerson, S. Turan, C. Ruiz-Rodriguez, C. Owens, S.X. Tsang, L.E. Moore, D. TheodorescuStudy supervision: M.L. Nickerson, D. TheodorescuOther (pathway analysis on the paper, as well as the co-mutationcytoscape model with collaborator J. Brown.): M.G. EdwardsOther (PCRandSanger sequencingof bladder tumor samples):C. Ruiz-Rodriguez

AcknowledgmentsThe authors thank the Theodorescu and Dean lab members and Dr.

Berton Zbar (National Cancer Institute) for comments; and Dr. SilviaJim�enez Morales (Instituto Nacional de Medicina Gen�omica, Mexico), Dr.Dom Esposito, David Wells, Alana Ebert-Zavos, Andrew Warner, and AllenKane (National Cancer Institute), and Dr. Jean-Noel Billaud (Ingenuity) fortechnical assistance.

Grant SupportThis work is supported in part by NIH grants CA075115 and CA104106

(D. Theodorescu) and by the Intramural Research Program of the NIH,the National Cancer Institute, Center for Cancer Research (M.L. Nickerson,M. Dean).The content of this publication does not necessarily reflect theviews or policies of the Department of Health andHuman Services, nor doesmention of trade names, commercial products, or organizations implyendorsement by the U.S. Government. The funders had no role in studydesign, data collection and analysis, decision to publish, or preparation ofthe manuscript.

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received February 7, 2014; revised May 5, 2014; accepted May 22, 2014;published online September 15, 2014.

References1. JemalA,BrayF,CenterMM,Ferlay J,WardE, FormanD.Global cancer

statistics. CA Cancer J Clin 2011;61:69–90.2. WuXR.Urothelial tumorigenesis: a tale of divergent pathways. Nat Rev

Cancer 2005;5:713–25.3. Gui Y, Guo G, Huang Y, Hu X, Tang A, Gao S, et al. Frequentmutations

of chromatin remodeling genes in transitional cell carcinoma of thebladder. Nat Genet 2011;43:875–8.

4. Hurst CD, Platt FM, Taylor CF, KnowlesMA. Novel tumor subgroups ofurothelial carcinoma of the bladder defined by integrated genomicanalysis. Clin Cancer Res 2012;18:5865–77.

5. Guo G, Sun X, Chen C, Wu S, Huang P, Li Z, et al. Whole-genome andwhole-exome sequencing of bladder cancer identifies frequent altera-tions in genes involved in sister chromatid cohesion and segregation.Nat Genet 2013;45:1459–63.

6. SolomonDA,Kim JS, Bondaruk J, Shariat SF,WangZF, Elkahloun AG,et al. Frequent truncating mutations of STAG2 in bladder cancer. NatGenet 2013;45:1428–30;45:1459–1463.

7. Killela PJ, Reitman ZJ, Jiao Y, Bettegowda C, Agrawal N, Diaz LAJr.,et al. TERT promoter mutations occur frequently in gliomas and asubset of tumors derived fromcells with low rates of self-renewal. ProcNatl Acad Sci U S A 2013;110:6021–6.

8. NickersonML, Im KM,Misner KJ, TanW, Lou H, Gold B, et al. Somaticalterations contributing tometastasis of a castration resistant prostatecancer. Hum Mutat 2013;34:1231–41.

9. Nickerson ML, Jaeger E, Shi Y, Durocher JA, Mahurkar S, Zaridze D,et al. Improved identification of von Hippel-Lindau gene alterations inclear cell renal tumors. Clin Cancer Res 2008;14:4726–34.

10. Farley MN, Schmidt LS, Mester JL, Pena-Llopis S, Pavia-Jimenez A,Christie A, et al. A novel germline mutation in BAP1 predisposesto familial clear-cell renal cell carcinoma. Mol Cancer Res 2013;11:1061–71.

11. Burgie SE, Bingman CA, Soni AB, Phillips GN Jr. Structural charac-terization of human Uch37. Proteins 2012;80:649–54.

12. Ventii KH, Devi NS, Friedrich KL, Chernova TA, Tighiouart M, Van MeirEG, et al. BRCA1-associated protein-1 is a tumor suppressor thatrequires deubiquitinating activity and nuclear localization. Cancer Res2008;68:6953–62.

13. Dancik GM, Owens CR, Iczkowski KA, Theodorescu D. A cell of origingene signature indicates human bladder cancer has distinct cellularprogenitors. Stem Cells 2014;32:974–82.

14. Gildea JJ, Golden WL, Harding MA, Theodorescu D. Genetic and phe-notypic changes associated with the acquisition of tumorigenicity inhuman bladder cancer. Genes Chromosomes Cancer 2000;27:252–63.

15. Greenman C, Stephens P, Smith R, Dalgliesh GL, Hunter C, Bignell G,et al. Patterns of somatic mutation in human cancer genomes. Nature2007;446:153–8.

16. Kumar A, White TA, MacKenzie AP, Clegg N, Lee C, Dumpit RF, et al.Exome sequencing identifies a spectrum of mutation frequencies inadvanced and lethal prostate cancers. Proc Natl Acad Sci U S A2011;108:17087–92.

17. Cappellen D, De Oliveira C, Ricol D, de Medina S, Bourdin J, Sastre-Garau X, et al. Frequent activating mutations of FGFR3 in humanbladder and cervix carcinomas. Nat Genet 1999;23:18–20.

18. Cordon-Cardo C, Dalbagni G, Saez GT, Oliva MR, Zhang ZF, Rosai J,et al. p53 mutations in human bladder cancer: genotypic versusphenotypic patterns. Int J Cancer 1994;56:347–53.

19. Knowles MA, Habuchi T, Kennedy W, Cuthbert-Heavens D. Mutationspectrum of the 9q34 tuberous sclerosis gene TSC1 in transitional cellcarcinoma of the bladder. Cancer Res 2003;63:7652–6.

20. Kandoth C,McLellanMD, Vandin F, Ye K, Niu B, Lu C, et al. Mutationallandscape and significance across 12 major cancer types. Nature2013;502:333–9.

21. Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA Jr.,Kinzler KW. Cancer genome landscapes. Science 2013;339:1546–58.

22. Woodage T, Basrai MA, Baxevanis AD, Hieter P, Collins FS. Charac-terization of the CHD family of proteins. Proc Natl Acad Sci U S A1997;94:11472–7.

23. Tian F, Xu F, Zhang ZY, Ge JP, Wei ZF, Xu XF, et al. Expression ofCHD1L in bladder cancer and its influence on prognosis and survival.Tumour Biol 2013;34:3687–90.

24. Brockschmidt A, Chung B, Weber S, Fischer DC, Kolatsi-Joannou M,Christ L, et al. CHD1L: a new candidate gene for congenital anomaliesof the kidneys and urinary tract (CAKUT). Nephrol Dial Transplant2012;27:2355–64.

25. EbmeierCC, TaatjesDJ. Activator-mediator binding regulatesmediator-cofactor interactions. Proc Natl Acad Sci U S A 2010;107:11283–8.

26. Pena-Llopis S, Vega-Rubin-de-Celis S, Liao A, Leng N, Pavia-JimenezA,WangS, et al. BAP1 loss defines a newclass of renal cell carcinoma.Nat Genet 2012;44:751–9.

27. Jensen DE, Proctor M, Marquis ST, Gardner HP, Ha SI, Chodosh LA,et al. BAP1: a novel ubiquitin hydrolase which binds to the BRCA1RING finger and enhances BRCA1-mediated cell growth suppression.Oncogene 1998;16:1097–112.

28. O'Donovan PJ, Livingston DM. BRCA1 and BRCA2: breast/ovariancancer susceptibility gene products and participants in DNA double-strand break repair. Carcinogenesis 2010;31:961–7.





29. Solomon DA, Kim T, Diaz-Martinez LA, Fair J, Elkahloun AG, Harris BT,et al. Mutational inactivation of STAG2 causes aneuploidy in humancancer. Science 2011;333:1039–43.

30. Horn S, Figl A, Rachakonda PS, Fischer C, Sucker A, Gast A, et al.TERT promoter mutations in familial and sporadic melanoma. Science2013;339:959–61.

31. Huang FW, Hodis E, Xu MJ, Kryukov GV, Chin L, Garraway LA. Highlyrecurrent TERT promoter mutations in human melanoma. Science2013;339:957–9.

32. Thiesen HJ, Bach C. Target Detection Assay (TDA): a versatile pro-cedure to determine DNA binding sites as demonstrated on SP1protein. Nucleic Acids Res 1990;18:3203–9.

33. Terashima M, Ishimura A, Yoshida M, Suzuki Y, Sugano S, Suzuki T.The tumor suppressor Rb and its related Rbl2 genes are regulated byUtx histone demethylase. Biochem Biophys Res Commun 2010;399:238–44.

34. Lee MG, Villa R, Trojer P, Norman J, Yan KP, Reinberg D, et al.Demethylation of H3K27 regulates polycomb recruitment and H2Aubiquitination. Science 2007;318:447–50.

35. Issaeva I, Zonis Y, Rozovskaia T, Orlovsky K, Croce CM, Nakamura T,et al. Knockdown of ALR (MLL2) reveals ALR target genes and leadsto alterations in cell adhesion and growth. Mol Cell Biol 2007;27:1889–903.

36. ClemBF,Chesney J.Molecular pathways: regulation ofmetabolismbyRB. Clin Cancer Res 2012;18:6096–100.

37. Marschalek R. Mixed lineage leukemia: roles in human malignanciesand potential therapy. FEBS J 2010;277:1822–31.

38. Medina PP, Romero OA, Kohno T, Montuenga LM, Pio R, Yokota J,et al. Frequent BRG1/SMARCA4-inactivatingmutations in human lungcancer cell lines. Hum Mutat 2008;29:617–22.

39. Wang X,Wang Y, Yu L, Sakakura K, Visus C, Schwab JH, et al. CSPG4in cancer: multiple roles. Curr Mol Med 2010;10:419–29.

40. Choi HS, Choi BY, Cho YY, Mizuno H, Kang BS, Bode AM, et al.Phosphorylation of histone H3 at serine 10 is indispensable for neo-plastic cell transformation. Cancer Res 2005;65:5818–27.

41. Williams RD. Human urologic cancer cell lines. Invest Urol 1980;17:359–63.

42. GrassoCS,WuYM,RobinsonDR,CaoX,DhanasekaranSM,KhanAP,et al. The mutational landscape of lethal castration-resistant prostatecancer. Nature 2012;487:239–43.

43. Garraway LA, Lander ES. Lessons from the cancer genome. Cell2013;153:17–37.

44. McGregor RF, Crawford R, Johnson DE, Brown B, Sharon MS, John-stonD.Urinary amino acid excretion: comparison of normal individualsand patients with bladder cancer. Urology 1977;9:538–42.

45. Scheuermann JC, de Ayala Alonso AG, Oktaba K, Ly-Hartig N, McGintyRK, Fraterman S, et al. Histone H2A deubiquitinase activity of thePolycomb repressive complex PR-DUB. Nature 2010;465:243–7.

46. Testa JR, Cheung M, Pei J, Below JE, Tan Y, Sementino E, et al.GermlineBAP1mutations predispose tomalignantmesothelioma. NatGenet 2011;43:1022–5.

47. Wiesner T, Obenauf AC, Murali R, Fried I, Griewank KG, Ulz P, et al.Germline mutations in BAP1 predispose to melanocytic tumors. NatGenet 2011;43:1018–21.

48. Pena-Llopis S, Christie A, Xie XJ, Brugarolas J. Cooperation andantagonism among cancer genes: the renal cancer paradigm. CancerRes 2013;73:4173–9.

49. Fong PC, Yap TA, Boss DS, Carden CP, Mergui-Roelvink M, GourleyC, et al. Poly(ADP)-ribose polymerase inhibition: frequent durableresponses in BRCA carrier ovarian cancer correlating with platinum-free interval. J Clin Oncol 2010;28:2512–9.

50. Guilleret I, Benhattar J. Demethylation of the human telomerasecatalytic subunit (hTERT) gene promoter reduced hTERT expressionand telomerase activity and shortened telomeres. Exp Cell Res2003;289:326–34.

51. PettigrewKA, Armstrong RN, Colyer HA, Zhang SD, Rea IM, Jones RE,et al. Differential TERTpromotermethylation and response to5-aza-20-deoxycytidine in acute myeloid leukemia cell lines: TERT expression,telomerase activity, telomere length, and cell death. Genes Chromo-somes Cancer 2012;51:768–80.


Nickerson et al.



2014;20:4935-4948. Clin Cancer Res Michael L. Nickerson, Garrett M. Dancik, Kate M. Im, et al. Bladder Cancer

, and the BRCA Pathway inKDM6A, TERTConcurrent Alterations in

Updated version

http://clincancerres.aacrjournals.org/content/20/18/4935

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2014/10/30/20.18.4935.DC1

Access the most recent supplemental material at:

Cited articles

http://clincancerres.aacrjournals.org/content/20/18/4935.full#ref-list-1

This article cites 51 articles, 19 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/20/18/4935.full#related-urls

This article has been cited by 5 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://clincancerres.aacrjournals.org/content/20/18/4935To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/content/suppl/2014/10/30/20.18.4935.DC1

http://clincancerres.aacrjournals.org/content/20/18/4935.full#ref-list-1

http://clincancerres.aacrjournals.org/content/20/18/4935.full#related-urls

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

mailto:[email protected]



concurrent alterations in tert kdm6a, and the brca …tert promoter analysis the tert promoter was...

Documents