casc15-s is a tumor suppressor lncrna at the 6p22 ... · tumor and stem cell biology casc15-s is a...

Tumor and Stem Cell Biology

CASC15-S Is a Tumor Suppressor lncRNA at the6p22 Neuroblastoma Susceptibility LocusMike R. Russell1, Annalise Penikis1, Derek A. Oldridge1, Juan R. Alvarez-Dominguez2,3,Lee McDaniel1, Maura Diamond1, Olivia Padovan4, Pichai Raman1,5, Yimei Li1, Jun S.Wei6,Shile Zhang6, Janahan Gnanchandran7, Robert Seeger7, Shahab Asgharzadeh7,Javed Khan6, Sharon J. Diskin1,8,9, John M. Maris1,8,9, and Kristina A. Cole1,8,9

Abstract

Chromosome 6p22 was identified recently as a neuroblastomasusceptibility locus, but its mechanistic contributions to tumor-igenesis are as yet undefined. Here we report that the most highlysignificant single-nucleotide polymorphism (SNP) associationsreside within CASC15, a long noncoding RNA that we define as atumor suppressor at 6p22. Low-level expression of a shortCASC15 isoform (CASC15-S) associated highly with advancedneuroblastoma and poor patient survival. In human neuroblas-

toma cells, attenuating CASC15-S increased cellular growth andmigratory capacity. Gene expression analysis revealed downregu-lation of neuroblastoma-specific markers in cells with attenuatedCASC15-S, with concomitant increases in cell adhesion andextracellular matrix transcripts. Altogether, our results point toCASC15-S as a mediator of neural growth and differentiation,which impacts neuroblastoma initiation and progression. CancerRes; 75(15); 3155–66. �2015 AACR.

IntroductionNeuroblastoma, a cancer of the developing autonomic nervous

system, is themost commonmalignancydiagnosed in the first yearof life and accounts for approximately 10% of all pediatric cancermortality (1–3). While the majority of low-risk neuroblastomapatients are cured with surgery alone, 50% of patients have thehigh-risk form of the disease, and only about half of these childrensurvive despite highly intensive therapy (1). Neuroblastomas arethought to develop from cells derived from the neural crestcommitted to the sympathicoadrenal lineage (peripheral auto-nomic nervous system; refs. 1–4). Because malignant transforma-tion can occur at any point during sympathetic development,tumors may arise throughout the developing sympathetic nervous

system (most commonly the adrenal gland), contributing to thehallmark heterogeneity observed in this disease (4). To address theetiology of sporadic neuroblastoma, we conducted the first pedi-atric cancer genome-wideassociation study (GWAS), leading to theidentification of numerous validated susceptibility loci in severalpopulations (5–12). Moreover, we showed that many of thesesusceptibility alleles are specifically associated with disease phe-notype and patient outcomes. Themajority of these SNPs act in cisto influence expression of protein coding genes at these loci, andseveral of these transcripts, such as LMO1, BARD1, and LIN28B,appear to play an oncogenic role in established tumors (5–12).

The first identified neuroblastoma susceptibility locus identi-fied byGWAS, and the one that remainsmost significant, mappedto chromosome 6p22.3 and robustly replicated in three indepen-dent cohorts [rs6939340: P¼ 9.33� 10�15; allelic OR 1.97, 95%confidence interval (CI), 1.58–2.45; ref. 5]. Like other subsequent-ly identified loci, we observed a highly significant associationwithneuroblastoma susceptibility and clinically aggressive presenta-tion. Theminor allele (G)was over represented in neuroblastomacases compared with controls, and presence of the G allele wasfurther enriched in the high-risk subset of neuroblastoma (P ¼0.007), tumors with MYCN amplification (P ¼ 0.002) and stageIV disease (P ¼ 0.025), implying the risk alleles were associatedwith amoremalignant neuroblastomaphenotype.On thebasis ofHapMap data available at the time of this initial discovery, theassociated SNPs tagged a 94.2 kb linkage disequilibrium (LD)block; this LD block overlapped two hypothetical genes(FLJ22536 and FLJ44180; ref. 5). However, both FLJ22536 andFLJ44180 lacked protein-coding potential, impeding further char-acterization of this region in neuroblastoma initiation.

Recent data obtained from whole-genome sequencing studiesof neuroblastomahave illustrated far fewer recurrentmutations inprotein-coding genes than previously predicted (13–16); howev-er, it is now clear that asmuch as 70%of the genome is transcribedinto products other than traditional protein-coding mRNAs (17,

1Division of Oncology, The Children's Hospital of Philadelphia, Phila-delphia, Pennsylvania. 2Whitehead Institute for Biomedical Research,Massachusetts Institute of Technology, Cambridge, Massachusetts.3Department of Biology, Massachusetts Institute of Technology, Cam-bridge, Massachusetts. 4Department of Physics and Astronomy, Uni-versity of Pennsylvania, Philadelphia, Pennsylvania. 5Center for Bio-medical Informatics,The Children's Hospital of Philadelphia, Philadel-phia, Pennsylvania. 6Oncogenomics Section, Genetics Branch, Centerfor Cancer Research, NCI, Bethesda, Maryland. 7Department of Pedi-atrics, Division of Hematology-Oncology, Children's Hospital LosAngeles and Saban Research Institute, University of Southern Cali-fornia, Los Angeles, California. 8Department of Pediatrics, Universityof Pennsylvania School of Medicine, Philadelphia, Pennsylvania. 9TheAbramson Family Cancer Research Institute, University of Pennsylva-nia School of Medicine, Philadelphia, Pennsylvania.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

CorrespondingAuthor:KristinaA. Cole, Children'sHospital of Philadelphia, 3501Civic Center Boulevard, Philadelphia, PA 19104. Phone: 267-426-2285; Fax: 267-426-0685; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-14-3613

�2015 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 3155

on October 21, 2020. © 2015 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst June 22, 2015; DOI: 10.1158/0008-5472.CAN-14-3613

http://cancerres.aacrjournals.org/

18). Althoughmany of these transcriptionally active loci produceRNA species involved in translation (i.e., ribosomal and transferRNAs), several other RNA classes have been functionally validatedas bona fide regulatory molecules. The recently identified longnoncoding RNAs (lncRNA), defined as RNA species >200 nt inlength that lack a functional open reading frame, have beenincreasingly implicated in a wide variety of cellular functions(19). LncRNAs share several transcriptional features in commonwith mRNAs, they are often spliced, demonstrate RNA polymer-ase II occupancy, contain a 50 methylguanosine cap, and arecommonly (though not always) polyadenylated (20, 21).Although lncRNA function is highly context dependent, theycommonly play a prominent role in the spatiotemporal regula-tion of gene expression during developmental processes (22–24),and therefore exhibit a tendency to be located throughout thegenome, in sites proximal to developmentally critical protein-coding genes (25). Indeed, several lncRNAs reside near protein-coding genes known to regulate lineage commitment in neuralcrest cells (26), serving as an attractive hypothesis to explain theetiology of embryonal cancers such as neuroblastoma.

As might be expected, lncRNAs have been increasingly impli-cated in a variety of oncogenic processes through association withepigenetic complexes and modification of chromatin accessibil-ity, ultimately influencing gene expression (27–30). To date, thereare few reports concerning the role of lncRNAs in the initiationand progression of solid pediatric neoplasms, despite the fact thatmany childhood cancers are fundamentally defects of normalhuman development (31). Here we describe the identificationand characterization of a novel lncRNA, CASC15, which contri-butes to the GWAS association signal on 6p22.3 by functioning asa tumor suppressor in neuroblastoma.

Materials and MethodsGWAS and imputation

In an effort to refine the association signal and search for acausal variant at the 6p22 locus, we performed genotype impu-tation in apreviously describedEuropean ancestry cohort of 2,101neuroblastoma cases and 4,202 controls (10) using the 1000Genomes Phase I Release 3 as a reference. Genotyping and qualitycontrol methods have been previously published (5). GWASimputation and statistical tests are detailed in SupplementaryMaterials and Methods.

Data sourcesThe human February 2009 (GRCh37/hg19) genome assembly

was used throughout the study. Transcript structures and annota-tions were obtained from GENCODE version 19. Details on thevarious data sources used are available in Supplementary Materi-als and Methods.

Neuroblastoma data. The neuroblastoma RNAseq, SNP profilingand HuEx datasets are part of the Therapeutically ApplicableResearch to Generate Effective Treatments (TARGET) initiative,supported by NCI Grant U10 CA98543. The low-level sequencedata have been deposited in the Sequence Read Archive at theNational Center for Biotechnology Information, and are furtheraccessible through the database of genotypes and phenotypes(dbGAP, accession number phs000218). The gene expression andcopy number data, as well as clinical information on theNBL casesstudied, is available via the TARGET Data Matrix (32).

50/30 Rapid amplification of cDNA ends50 and 30 Rapid amplification (RACE) was performed via the

First Choice RLM-RACE Kit (Ambion) using 10 mg of RNAobtained from fetal brain or NB69 neuroblastoma cells followingthe manufacturer's protocol. Specificity for CASC15-S wasachieved by nested PCR using the following gene-specific primer(GSP) pairs:

50 RACE: Outer GSP: 50-CTAGCCCATCAGTTCCTTCG-30

50 RACE: Inner GSP: 50-TTCACCCTGTCCTCCAAGTC-30

30 RACE: Outer GSP: 50-TGGTTACCTGAGCTGCTCCT-30

30 RACE: Inner GSP: 50-CTCAGCCAGTGCAACACAAC-30

Gene products were cloned into a pCR4-TOPO vector forsequencing.

RNA sequencingNeuroblastoma. PolyA selected RNA libraries obtained from 108high-risk neuroblastoma patients as part of the NCI TARGETproject were prepared using TruSeq v3 (Illumina) for RNAsequencing on Illumina HiSeq2000 sequencers. The 101 bppaired-end reads generated were aligned to the hg19 build of thehuman reference genome using TopHat v2.2.0, yielding amedianof 115 million total aligned reads per patient sample (range, 48–250million total aligned reads). The HTSeq package (v0.6.1) wasused to map aligned reads to the VISTA-annotated enhancerregion, hs1335, as well as all transcripts annotated in Ref-Seq(v66) and/or the UCSC Genome Browser. Transcript expressionvalues were normalized and quantified using the metric of readsper kilobase per million reads (RPKM). For isoform-specificquantitation of CASC15 and CASC15-S, only exons and exon–exon junctions that were unique to each isoform were used incomputing RPKM.

Other.Mapped RNA-seq reads from 16 primary tissues (IlluminaHuman Body Map) and 51 cell lines (Human ENCODE)were used to quantify gene- and transcript-level expression, basedon GENCODE v19 annotations, using Cufflinks v.2.2.0 withdefault parameters and "–min-frags-per-transfrag 0 –compati-ble-hits-norm –min-isoform-fraction 0.0." Expression data heat-maps were generated using the heatmap.2 function of the gplotsR package.

RNA FISHWe performed RNA FISH and counted RNA in single cells as

described in ref. 33. Fluorescently labeled, nonoverlapping oli-gonucleotide probes (20 mers) were designed to tile RNA tran-scripts, including CASC14 (LOC729177), hs1335, CASC15, andCASC15-S. Probes were then divided into odd and even pools andhybridized to neuroblastoma cells. Details of image acquisitionare available in Supplementary Materials and Methods.

Quantification of CASC15-S in neuroblastoma cellsQuantification of RNA transcripts was performed on a panel of

neuroblastoma cancer lines using TaqMan RT-PCR. For quanti-fication of CASC15-S, we developed a custom primer/probesetspanning exons 1–2 and consisting of the following primers:

Forward: GCTGTCGACGAAGGAACTGATReverse: GTCCAAGTCAAAAGTCTCATCCAAGAReporter: CCTGTGAGCCTAGCCC

Primer/probe sets for CASC15 (Hs01371949_m1) and CASC14(Hs04275511_s1) are commercially available (Life Technologies).

Russell et al.

Cancer Res; 75(15) August 1, 2015 Cancer Research3156




Quantification was normalized to the geometric mean of house-keeping genes TBP and GUSB.

Cell growth and siRNA assaysON-TARGETplus SmartPool siRNAs containing four constructs

per target (Thermo Scientific) were used for GAPDH and PLK-1knockdown. ON-TARGETplus Non-targeting Control Pool (con-taining four constructs)was used as a control. Two constructswereused for each lncRNA, and were purchased as Silencer SelectsiRNA or custom siRNAs from Life Technologies.

siRNA constructs were transfected into neuroblastoma cells intriplicate, using 50 nmol/L of siRNA in 0.1%–0.2%DharmaFECT1 (Thermo Scientific). Knockdown was assessed by TaqMan RT-PCR, comparing levels 72 hours after transfection to nontargetingcontrol–transfected cells. For all siRNA experiments, the mini-mum knockdown achieved was 71.5% (average ¼ 78.6, maxi-mum ¼ 92.0).

For growth assays, while both tested, the siRNA with the bestknockdown efficiency for each target is shown. Cell growth assayswere conducted using the xCELLigence real-time growth (RT-CES,ACEABiosciences) and/or Cell-TiterGlo (Promega) assays accord-ing to manufacturer protocols.

CASC15-S add-back experimentsFor CASC15-S addback experiments, expression plasmids con-

taining the cDNA for either CASC15-S or GFP were transfected intriplicate in a 96-well plate using 250 ng of DNA and 3 mL ofLipofectamine 2000 per well. These experiments were conductedusing the growth assays described above, and the expression levelsof theCASC15-S transcript were confirmed by TaqMan RT-PCR tobe highly expressed in the addback condition. Transfection effi-ciency was estimated to be approximately 90% based on the GFP-transfected condition.

shRNA expression and lentiviral constructsThe siRNA construct for CASC15-S (targeting exon 1) was used

to create a double-stranded shRNA construct andwas subsequent-ly cloned into the pLenti-GFP DEST lentiviral vector (Addgene). Atotal of 5�106293T cellswere transfectedwith 15mgof thepLentitransfer vector, 15 mg of pLP1, 6 mg of pLP2, and 3 mg of pVSV-Gvectors. Lentiviral particles were collected at 48 and 72 hours aftertransfection. Lentiviral transduction of neuroblastoma cells wascarried out overnight at 37�C using 3 mg/mL of polybrene. Foreach neuroblastoma line used, serial dilutions of lentiviral par-ticles in basal media were carried out in a 6-well plate, and cellswere assessed at 48 hours after transduction for GFP expression.The minimum lentiviral concentration resulting in >95% GFPexpression was chosen for expansion and knockdown was vali-dated by TaqMan PCR. For all shRNA experiments, knockdown ofat least 70.3% (average ¼ 82.6, maximum ¼ 91.8) was achieved.Each neuroblastoma line with confirmed knockdown was thenfrozen at low passage (<5) and thawed as needed for subsequentexperiments.

Luciferase reporter assayThe respective risk (A_A) andnon-risk (T_T) genotypes of Lan-5

and CHP-134 cells at rs9295534 were verified by PCR and Sangersequencing. Subsequently, a 1,500 bp fragment was cloned fromeach cell line (750 bp on either side of the rs9295534) and waspurified using the Qiaquick Gel Extraction Kit (Qiagen). Thefragment was inserted into the pGL4.23 vector upstream of a

minimal promoter used to drive firefly luciferase expression.Fragment-containing vectors were then transfected into HEK-293 cells, along with a Renilla luciferase vector (pGL4.75) tocontrol for transfection efficiency. Seventy-two hours after trans-fection, luciferase activity was assayed by Dual-Glo reporter assay(Promega) using the GloMaxMulti Detection System (Promega).Firefly luciferase was normalized to Renilla luciferase and a min-imum of three replicates was used for each condition.

In vitro transcription/translation assayCASC15-S was cloned into a T7-promoter containing plasmid

(T7-CFE-Chis) and subsequently verified by Sanger sequencing.Coding potential of CASC15-S was assayed using the TnT QuickCoupled Transcription/Translation System according to theman-ufacturer's protocol. Incorporation of biotinylated lysine residueswas visualized via Western blot analysis using a 1:10,000 anti-biotin HRP–labeled antibody (Cell Signaling Technology) andchemiluminescent detection (Thermo Scientific).

Wound-healing assaysScratch assays were carried out on SK-N-BE2 and SK-N-SH cells

stably depleted of CASC15-S plated at 85% confluence in 60 mmdishes, and were scratched using a sterile 200 mL pipette tip. Cellswere photographed at regular intervals using a previously cali-brated 5� lightmicroscope (Nikon). Assessment of cellmigrationwas carried out by measuring scratch closure as a percentage ofinitial scratch size in ImageJ, and was compared with control cellsusing a linear regression function in GraphPad Prism 6.

Cell cultureNeuroblastoma cell lines were obtained from the neuroblas-

toma cell line bank maintained at the Children's Hospital ofPhiladelphia, Philadelphia, PA. Cell line identity is routinelyconfirmed via AmpFLSTR Identifier (Applied Biosystems), lastdone in November 2013. Non-neuroblastoma cell lines werepurchased from ATCC and all cell lines are routinely tested formycoplasma. All cell lines are maintained in basal media (eitherRPMI1640 or DMEM) supplemented with 10% FBS and 1%gentamycin and cultured at 5% CO2.

Statistical analysisAll group comparisons were conducted using nonparametric

testing (Mann–Whitney U) in GraphPad Prism.For Kaplan–Meier analysis: optimal cutoff was determined by

employing a scanning approach to the Kaplan–Meier method byiteratively splitting the ordered genes expression values acrosssamples into two groups and calculating the P value by theMantel–Haenszel log-rank test. The lowest P value correspondsto the optimal breakpoint. A Benjamini–Hochberg correctionwasapplied to reflect the presence of multiple hypotheses testing.

For multivariate analyses, a Cox Proportional Hazard modelwas used to evaluate the effect of each gene expression on overallsurvival, adjusting for clinical factors such as age (as a continuousvariable), MYCN amplification, and 1p/11q LOH status. Propor-tional hazard assumption was evaluated by the log–log plot andthere was no evidence of violation of this assumption.

ResultsFinemapping of the 6p22.3 neuroblastoma susceptibility locus

The initial GWAS study illustrating 6p22.3 as a neuroblastomasusceptibility locus identified three common polymorphisms

CASC15 Is a Neuroblastoma Suppressor Gene

www.aacrjournals.org Cancer Res; 75(15) August 1, 2015 3157




clustered on chromosome 6p22.3 that were highly associatedwith aggressive disease and a significantly increased risk of neu-roblastoma development (5). These SNPs were shown to tag a94.2 kb LDblock overlapping two hypothetical genes and flankedupstream by SOX4. To map this region with finer detail, andidentify putative causal/functional SNPs, we performed genotypeimputation in our recently published discovery cohort of 2,101cases and 4,202 controls (10). We first applied SHAPEIT (34) toinfer haplotypes, and then utilized IMPUTE2 (35) with defaultparameters and Ne ¼ 20,000, along with a multipopulationreference panel from the worldwide 1000 Genomes Project Phase1 release to impute genotypes across the region. Genotyped andimputed variants were tested for association with neuroblastomausing the frequentist association test under the additive modelusing the "score" method implemented in SNPTEST (36). Var-iants with minor allele frequency < 1% and/or IMPUTE2-infoquality score < 0.8were excluded for quality control purposes. Weidentified 32 polymorphisms with P values less than 1� 10�10 (P¼ 8.26� 10�10

–1.88� 10�15), 12 of which were in high LD (r2 >0.8) and localized to an intronic regionof an annotated gene locusformerly titled FLJ22536/LINC00340 and more recently renamed

cancer-associated susceptibility candidate 15 (CASC15, Fig. 1A;Supplementary Table S1). These 32 SNPs identify a 34.9-kblinkage disequilibrium (LD) block based on the 1000 GenomesNorthern and Western European (CEU) population (Fig. 1B),significantly refining this relatively small locus.

Identification of CASC15 isoform expressionCASC15 is annotated as an lncRNA, prompting us to exam-

ine the difference in lncRNA transcript levels between 220high-risk and 30 low-risk primary tumors, for which wepossessed microarray expression data. This analysis revealedthat only 3.5% of annotated lncRNAs (hg19: 8/229) appearsignificantly differentially regulated between high- and low-risk tumors (Table 1). Furthermore, high-risk neuroblastomasexhibit a 4.4-fold lower level of CASC15 expression as com-pared with low-risk disease, the most highly significant (P ¼3.60 � 10�17) differentially expressed lncRNA we observed.Several computationally predicted lncRNAs map to the 6p22.3locus, including multiple CASC15 isoforms and an antisense-strand lncRNA, CASC14 (formerly LOC729177) (Supplemen-tary Fig. S1A).

A

B

CASC15 CASC15-S

CASC14

22.05 22.1 22.15 22.2 Position on chr6 (Mb)

Annotation key

15

10

5

0

100

80

60

40

20

0

−log

10 (

P v

alue

)

Recom

bination rate (cM/M

b)

r2

0.8

0.6

0.4

0.2

Figure 1.Fine mapping of 6p22 identifies CASC15-S as a candidate cis-acting neuroblastoma susceptibility gene. A, regional association plot of SNPs generated withLocusZoom software using genome-wide imputation data from 2,817 neuroblastoma cases and 7,473 controls. The pair-wise LD (r2) for each SNP isdenoted by color, and the log-transformed P values for each SNP are shown on the y-axis. Thirty-two SNPswith highly significant P values (<1� 10�10) ranging from4.67 � 10�10 to 4.81 � 10�17 were found to cluster as a narrow peak within a genomic region containing the lncRNAs CASC14 and CASC15. The exons andtranscribed regions of genes are shown as solid vertical lines. B, the haplotype structure of this region in Northern European (CEU) population demonstrates a signalthat overlaps with these SNPs, refining and initial 94.2 kb LD block (boundaries denoted by black vertical dashed lines) down to 34.9 kb, demonstrated bythe dashed red vertical line and rightmost black vertical dashed line. SNP annotations: color, representative of LD r2 value (purple, reference SNP); &, predictedcoding or 30UTR region; D, nonsynonymous; �, tfbccons, conserved motif at transcription factor binding site; x, mcs44placental, highly conservedregion in placental mammals.

Russell et al.





We therefore first utilized RNA sequencing (RNASeq) andRNA-paired end tagged (RNA-PET) data available from theENCODE project to identify the products transcribed from thislocus in SK-N-SH and SK-N-BE2 neuroblastoma cell lines.These data demonstrated the existence of two capped andpolyadenylated nuclear CASC15 transcripts, a long (hg19, chr6:21,666,675-22,194,616; Ensembl: CASC15-003) and a shortisoform (hg19, chr6: 22,146,883-22,194,616; Ensembl:CASC15-004; Fig. 2A and Supplementary Figs. S1B and S2A).These nuclear noncoding transcripts are highly conserved invertebrates and readily detected in several brain regions andneuroblastoma cell lines (Supplementary Fig. S2) with putativepromoter regions separated by 480 kb indicative of indepen-dent transcriptional regulation.

We next examined RNA expression data from a panel of 16primary human tissues as part of the Illumina Human Body Mapproject, where we found that the short CASC15 isoform wasexpressed in abundance in the brain, but at modest to low levelsin most other tissues (Fig. 2B). We subsequently generated RNAsequencing data from 108 primary neuroblastoma tumors usingnonoverlapping, transcript-specific reads to investigate the expres-sion of lncRNAs transcribed from this locus. We augmented thesefindings with our own strand-specific RNA sequencing data fromtwoprimary tumors, confirmingnearly complete alignment to theplus strand (average ¼ 92.6%, minimum ¼ 86.9%, maximum ¼98.6%). Together, these RNA sequencing data identify the shortCASC15 isoform as the predominant transcript expressed fromthis locus in neuroblastoma, with expression values 20- to 40-foldhigher than CASC14 and full-length CASC15, respectively (Fig.2C). Finally, RNA sequencing data identified a third unsplicedtranscript, spanning part of exon 1 of the short CASC15 isoformthrough a downstream noncoding element (hs1335) with a val-idated enhancer function in the developing murine neural tube,further supporting the role of this locus in neural development(37).

To validate the isoforms identified from RNA sequencingdata, we performed 50 and 30 RACE for CASC15, subsequentlycloning and sequencing these transcripts from two neuroblas-toma cell lines and fetal brain tissue. We verified the sequenceof both the 12-exon 1.9 kb CASC15 transcript (NR_015410.1,Ensembl: CASC15-003) and the 4-exon short (1.2 kb) variant(Ensembl: CASC15-004), hereafter referred to as CASC15-S.CASC15-S resembles a known cDNA clone (GenBank:AK094718), containing a unique first exon, yet sharing itsremaining sequence with the last three exons of CASC15 (Fig.2A and Supplementary Fig. S3A). Despite several attempts, wewere unable to isolate a 50-capped product for the intron-less

transcript (Ensembl: CASC15-006) predicted to overlap exon 1of CASC15-S and extend to the noncoding enhancer element(hs1335). Indeed, although the presence of this isoform isindicated on the SK-N-SH RNASeq track (Fig. 2A), it is absentin RNA-PET data (Supplementary Fig. S1B), supporting thisresult. Finally, RNA-FISH experimentally confirmed the pres-ence, relative abundance, and cellular localization of thesetranscripts in both NB-69 and NGP neuroblastoma cells; usingnonoverlapping, strand-specific probes to label CASC15,CASC15-S, CASC14, and hs1335. These studies revealed exclu-sively nuclear CASC15-S and hs1335 transcripts (Fig. 2D) andvirtually no CASC14 or CASC15 expression (SupplementaryFig. S3B and S3C). Taken together, these experimental resultsconfirm our predictive bioinformatic data identifying CASC15-S as a bona fide lncRNA transcript in neuroblastoma.

Identification of candidate functional SNPs at 6p22.3Because risk alleles at several other neuroblastoma suscep-

tibility loci have been shown to function in cis to influencemRNA expression levels of nearby transcripts (6, 9, 10), weattempted to associate our previously published, highly signif-icant polymorphism, rs6939340 (P ¼ 1.67 � 10�14; OR, 1.80;95% CI, 1.55–2.10; ref. 5) with expression of the CASC14,CASC15, and CASC15-S gene products at this locus. However,we did not observe a significant correlation between risk allelesand transcript expression levels in either a set of 250 primaryneuroblastomas (Supplementary Fig. S4A) or a representativepanel of 20 neuroblastoma cell lines (Supplementary Fig. S4B).Furthermore, all three of our previously published polymorph-isms lie in regions devoid of DNase hypersensitivity or otherepigenetic marks indicative of transcriptional activity. In fact,eQTL analysis of all imputed genotypes with CASC15-S expres-sion failed to reach significance after multiple comparisontesting, leading us to postulate that polymorphisms contribut-ing marginal effects may aggregately impact function at thislocus.

We therefore took advantage of our genome-wide imputationdata to identify other highly significant polymorphisms lyingwithin putative regulatory regions. As demonstrated in otherpost-GWAS follow-up studies (38), we employed the followingworkflow (Fig. 3A) to narrow the field of potentially functionalpolymorphisms: (i) we first chose polymorphismswith aGWAS Pvalue < 1 � 10�10, resulting in 32 candidates; (ii) we furtherrefined this list by selecting only those SNPs within regions ofDNaseI hypersensitivity (indicating open chromatin) andH3K27acetylation marks (indicative of enhancer activity) leaving uswith four candidates: rs1543310, rs6905441, rs9295534, and

Table 1. Differential expression analysis of lncRNAs between high- and low-risk neuroblastomas

Gene Ratio Disease state P Location

LINC00340 4.4-fold lower in High risk 3.60E�17 chr6:21666675-22194616LINC00174 2.4-fold lower in High risk 2.28E�15 chr7:65841031-65865395LINC01296 8.9-fold higher in High risk 5.80E�15 chr14:19880209-19925329LINC00260 2.8-fold lower in High risk 1.75E�13 chr1:203699705-203700979LINC00221 2.4-fold higher in High risk 7.04E�12 chr14:106938445-106951529LINC00478 2.5-fold higher in High risk 1.74E�04 chr21:17442842-17982094LINC00514 2.4-fold lower in High risk 3.70E�04 chr16:3039055-3044510LINC00355 2.9-fold higher in High risk 5.01E�04 chr13:64560504-64650144

NOTE: Differential gene expression analysis (>2.0-fold; P < 0.05; FDR < 0.05) was conducted using 220 high- and 30 low-risk primary neuroblastoma tumors,focusing on differences in lncRNA expression. The top differentially regulated lncRNA was CASC15 (annotated here as LINC00340), which was significantly lower inhigh-risk disease (4.4-fold decrease; ANOVA, P ¼ 3.60 � 10�17).






rs9368402; (iii) Finally, we looked for SNPs with evolutionaryconservation, resulting in a single candidate polymorphism,rs9295534 (P ¼ 3.51 � 10�12; OR, 1.63; 95% CI, 1.4–1.89)upstream of CASC15-S, and this variant localizes to an expanse ofregulatory chromatin and dense transcription factor binding sitesin several cell lines (Fig. 3B). Moreover, this region exhibitsenhancer activity, evidenced by H3K27Ac CHIP-Seq data inseveral fetal tissues available from the NIH Roadmap Epige-nomics Mapping Consortium (Fig. 3C).

We next verified rs9295534 genotypes in Chp134 (homozy-gous non-risk) and Lan5 (homozygous risk) neuroblastomacells by Sanger sequencing of a 1.5 kb region encompassing thisSNP, and subsequently cloned risk and non-risk fragmentsfrom these lines. To assess the impact of rs9295534 genotype

on transcriptional activity, we inserted risk (A/A) and non-risk(T/T) fragments into luciferase reporter constructs upstream ofa minimal promoter. Results from these experiments demon-strated significantly decreased transcriptional reporter activityfollowing insertion of the risk genotype fragment (Fig. 3D),suggesting this region possesses an enhancer-like function thatis disrupted following the inclusion of a neuroblastoma riskallele.

CASC15-S is differentially expressed in neuroblastoma andhighly associated with disease outcome

Because the rs9295534 homozygous risk genotype, by virtueof its linkage with rs6939340, is associated with aggressive neu-roblastoma, poor survival (5), and demonstrates decreased

CASC15 (refseq)

CASC15-S CASC140

10

20

30

RN

ASe

q re

ads

per k

iloba

s pe

r mill

ion

read

s (R

PKM

)

hs1335CASC15-S

Merge Merge + DAPI

CB

A

D

Figure 2.CASC15-S is the predominant lncRNA isoform expressed in neuroblastoma. A, graphical representation of the lncRNA transcripts observed to originate from the6p22.3 locus via RNA sequencing. This locus includes two predominant transcripts on the positive strand: a 1.9 kb CASC15 transcript (current RefSeq forthis lncRNA) and a novel shorter (1.2 kb) transcript,CASC15-S, that shares the last three exonswith the long isoformofCASC15. These transcriptswere validated using50 and 30 RACE. Epigenetic marks support active transcription (RNA PolII occupancy) and an enhancer-like function (H3K27Ac, H3K4Me3) of this region,including a proximal VISTA enhancer element (hs1335). RNA sequencing, RNA PolII, DNAse, and P300 ChIP-Seq data were obtained from SK-N-SH neuroblastomacells (via ENCODE), whereas enhancer track ChIP-Seq data was taken from MGG8 human glioblastoma stem cells (48). B, expression levels of CASC15, CASC15-S,CASC14, and hs1335were normalized andquantified across a panel of 16 normal primary tissues (HumanBodyMap 2.0 project), indicating predominant expression ofthe CASC15-S isoform in brain. C, RNA sequencing performed on 108 primary neuroblastoma tumors, analyzed for unique isoform expression, exhibited apredominance of the short CASC15 isoform. This short isoform was expressed at 20-fold and 40-fold higher levels than CASC14 and or full-length CASC15,respectively. D, RNA-FISH was conducted in NGP neuroblastoma cells for several transcripts known to exist at this locus. To ensure probe specificity, odd and evenpools of fluorescently labeled oligonucleotide probes were used to tile CASC15-S and VISTA hs1335. The yellow fluorescence observed in the "merge"panel was obtained from overlap of the hs1335 and CASC15-S fluorescent signals and indicate colocalization of these two transcripts. Consistent with themajority ofdescribed lncRNAs, these transcripts appear to be predominately nuclear as evidenced in the "merge þ DAPI" panel. (magnification, �40)

Russell et al.





transcriptional activity in a minimal promoter assay, we postu-lated that patients with high-risk disease would have reducedCASC15-S expression. Indeed, we observed significantly lowerCASC15-S expression in high-risk patient tumors (n ¼ 230)compared with low-risk patient tumors (n ¼ 30; Fig. 4A). Thisfinding appeared independent of MYCN amplification, a knownoncogenic driver in neuroblastoma (P¼ 0.29, Supplementary Fig.S5A). In addition, patients with tumors enriched in CASC15-Sexpression exhibited superior survival when compared withpatients harboring neuroblastoma with low CASC15-S expres-sion, both when low- and high-risk patients were included in theanalysis (adj. P¼ 3.2� 10�06, Fig. 4B) but also when comparingexpression from only high-risk patients (adj. P ¼ 0.002, Supple-mentary Fig. S5B). Furthermore, this finding remained significant(P ¼ 0.0084) after multivariate analysis adjusting for clinicalfactors such as age, MYCN amplification, and 1p/11q deletionstatus, all known prognostic variables in this disease (Supple-mentary Table S2). Although expressed at much lower levels than

CASC15-S, the long isoform of CASC15 demonstrated a similarpattern in patient tumors (Supplementary Fig. S5C–S5E). Takentogether, these data indicate that low CASC15-S expression cor-relates with a more aggressive phenotype in neuroblastoma anddecreased overall survival probability.

To understand the contribution of CASC15-S in the initiationand progression of neuroblastoma, we performed gene set enrich-ment analysis (GSEA) restricted to high-risk neuroblastoma sam-ples (n ¼ 220), where we utilized the same 1.9-fold difference inmedian CASC15-S expression as the Kaplan–Meier analysis todefine 146 low- and 74 high-expressing CASC15-S samples(Fig. 4C). The topdifferential gene expression profile (normalizedenrichment score ¼ �2.69, nominal P value < 0.0001, FDR Qvalue < 0.0001) that emerged from these analyses consists of a 55gene signature previously shown by Asgharzadeh and colleaguesto be downregulated in neuroblastoma patients with poor out-comes (39). This result indicates that tumors with highCASC15-Sexpression are enriched in the expression of these genes, and again

H3K27acFetal Adrenal

Fetal LungHUVEC

Weak transcription

Enhancer

Strong enhancer

22112900 22113100 22113300 22113500 22113700

Chromosome 6p22.3

rs9295534chr6:22113390 P = 3.5139e−12

A

0.0

1.0×10-3

2.0×10-3

3.0×10-3

Nor

mal

ized

luci

fera

seac

tivity

RiskNon-risk

*** P < 0.0001

rs9295534 genotype

C

B

D

3. DNase hypersensi�vityH3K27 acetyla�on

1. Regional imputa�on

2. GWAS P < 1x10-10

4. Evolu�onary conserva�on

Filtering strategy Candidate SNPs

2,627

32

4

rs9295534

Figure 3.Identification of rs9295534 as a functional polymorphism. A, filtration strategy employed to refine list of imputed SPSs at the 6p22.3 locus for functional variants.Weidentified SNPs with robust P values (<1 � 10�10) that fell within regions of regulator chromatin (DNase hypersensitivity) and putative enhancer activity(H3K27 acetylation) in SK-N-SH neuroblastoma cells. This resulted in four highly significant polymorphisms: rs1543310, rs6905441, rs9295534, and rs9368402.Further refinement based on evolutionary conservation identified rs9295534 as the only SNP to fit all criteria (a full table of SNP attributes is given in SupplementaryTable S1). B, evidence from glioblastoma cells revealing that rs9295534 (light blue vertical line) maps to the most proximal enhancer site to CASC15-S asevidenced by DNaseI sensitivity (DS), H3K4me1, and H3K27ac marks (49). C, further characterization of rs9295534 was indicated due to a robust P value (P ¼ 3.51� 10�12), as well as the inclusion of this SNP within a region of predicted enhancer activity (H3K27Ac peaks) in several fetal tissues, including fetal adrenalgland. D, functional demonstration of rs9295534 was accomplished by insertion of 1500-bp risk (A_A) or non-risk (T_T) fragments (cloned from Lan5 and Chp134neuroblastoma cell lines, respectively) into a luciferase reporter vector (pGL4.23) upstream of a minimal promoter. Luciferase activity was normalized to acontransfectedRenilla luciferase vector (pGL4.75). Expressionof the risk fragment resulted in significantly attenuated transcriptional activity thanwas observedwiththe non-risk fragment following quantification of luciferase activity in HEK-293 cells (P < 0.0001).






support a protective role forCASC15-S, resulting in less aggressivedisease within even this subgroup of high-risk cases (Fig. 4D).

CASC15 depletion in neuroblastoma cell lines enhancesproliferation and invasive capabilities

Having demonstrated a clinically relevant association withpatient outcome forCASC15-S in neuroblastoma patients, we nextsought functional validation for a role in tumorigenesis. We firstassessed CASC15-S levels by qRT-PCR across a well-characterizedpanel of neuroblastoma cell lines (n ¼ 21) where we observeddifferential expression similar to our primary tumor panel (Fig.5A). To investigate the functional role of CASC15-S in neuroblas-toma, we utilized an siRNA construct targeting exon 4 near the 30

end of the gene. Depletion of CASC15-S resulted in a highlyreproducible increase in neuroblastoma proliferation as evidencedby real-time cell growth and viability assays (Fig. 5B), and we wereable to recapitulate these results by subsequently targetingCASC15-S within its unique first exon in several neuroblastomalines (Fig. 5C and Supplementary Fig. S6A and S6B). Depletion of

full-length CASC15 or CASC14 had no observable impact on cellgrowth or viability, supporting our initial findings of CASC15-S asthe functional isoform in neuroblastoma (Supplementary Figs.S6C and S6D and S7A–S7D). We next derived neuroblastoma celllines stably depleted of CASC15-S (Supplementary Fig. S7E) andfound that these cells also exhibited a substantial increase incellular proliferation identical towhatweobserved inour transientsiRNA-based CASC15-S depletion experiments (Fig. 4D and Sup-plementary Figs. S6E and S6F and S8). Furthermore, rescue experi-ments, conducted by ectopically expressing CASC15-S in thesecells, were able to revert the accelerated growth (Fig. 5E). Micro-scopic examination revealed overt morphologic changes inshCASC15-S SK-N-BE2 cells, including striking changes in cellshape and size, with a resultant 3-fold increase in cell area(659.5� 50.2 mm2 in control vs. 2024� 211.1 mm2 in shCASC15cells, P < 0.0001; Fig. 5F andG). Furthermore, both SK-N-BE2 (Fig.6A and B) and SK-N-SH neuroblastoma cells (Fig. 6C) stablydepleted of CASC15-S exhibited an increased migratory capacityand invasiveness as evidenced bywound-healing assays (SK-N-SH:

DC

CASC15-Slow

CASC15-Shigh

Number of patients 146 74

Median expression 435.7 820.1

MYCN Amplified 30.1% 28.9%

Median diagnosis (y) 3.01 3.16

Median survival (y) 3.31 6.38 ***

Downregulated genes associated with poor outcome in neuroblastoma

Enr

ichm

ent s

core

0.0

−0.1

−0.2

−0.3

−0.4

−0.5

−0.6

−0.7

NES = −2.69Nominal P < 0.0001FDR q < 0.0001FWER P < 0.0001

A

0 2,000 4,000 6,0000

50

100

Overall survival

Time (days)

Per

cen

t su

rviv

al

Low CASC15-S (n = 163)

High CASC15-S (n = 87)

Adj. P = 3.2 × 10−06

BRisk group

Low risk High risk0

500

1,000

1,500

2,000

***

(n = 30) (n = 221)

CA

SC

15-S

exp

ress

ion

Figure 4.Low CASC15-S expression correlates with poor clinical prognosis. A, microarray expression data from clinically annotated primary neuroblastoma tumors (n¼ 250)obtained at diagnosis reveals that high-risk (stage IV) neuroblastomas (n ¼ 220) demonstrate significantly lower expression of CASC15-S than low-risk(stage 1) tumors (n ¼ 30). B, Kaplan–Meier analysis demonstrates significantly poorer overall survival for children with tumors expressing low levels ofCASC15-S (n¼ 163 for group "low"; n¼ 87 for group "high"; adj. P¼ 3.2� 10�06). C, relevantmetrics for selection of high-risk patient tumors used for differential geneexpression analyses between patients with high (n ¼ 74) and low (n ¼ 146) CASC15-S levels. CASC15-S expression was increased 1.9-fold in the "high"group, and these patients exhibited a significant increase in overall survival. D, the most highly significantly regulated pathway using GSEA was identified to bea set of genes downregulated in poor outcome neuroblastoma (Asgharzadeh neuroblastoma poor survival down). Patients with high CASC15-S expressiondemonstrated enrichment of these genes, suggesting that CASC15-S acts in a protective manner (�� , P < 0.0001).

Russell et al.





P ¼ 0.0006, SK-N-BE2; P < 0.0001). Taken together, these datasuggest thatCASC15-S loss promotes increased cellular growthanda more migratory phenotype in neuroblastoma.

CASC15-S regulates a subset of genes involved in neural crestdevelopment

To better characterize the phenotypic changes we observedfollowing CASC15-S depletion, we surveyed gene expressionsignatures of neuroblastoma cells depleted of CASC15-S. Forinitial studies, SK-N-SH neuroblastoma cells were transfected intriplicate with either a nontargeting construct or siRNA specificfor CASC15-S, and microarray gene expression signatures wereexamined at 48 hours following transfection. We observedsubstantial upregulation of several known cell adhesion genes,

most notably entactin (NID1, P ¼ 3.7 � 10�4) and activatedleukocyte cell adhesion molecule (ALCAM, p ¼ 1.05 � 10�7),after CASC15-S depletion. These gene expression data wereused for Gene Ontology (GO) analysis to examine enrichmenttop level biologic processes. In support of the increased migra-tory phenotype we observed in wound-healing assays, bothlocomotion and cellular adhesion pathways were found amongthe top differentially regulated processes as a result of CASC15-S depletion (Fig. 6D).

We next assayed the result of persistentCASC15-S depletion ongene expression changes by comparing SK-N-BE2 neuroblastomacells stably silenced for CASC15-S compared with control vector–transfected cells. We again examined gene expression signaturesvia microarray analysis, where we observed downregulation of

B

GF

C

A

E

1209672482400

1

2

3

0

1

2

3

Time (h) Time (h)

Empty vector

*** P < 0.0001

***C

ell i

ndex

Cel

l ind

ex

shCASC15-S

Empty vector shCASC15-SshCASC15-SEmpty vector

0

1,000

2,000

3,000

Cel

l siz

e (m

m2)

***

1209672482400

2

4

6

Time (h)

Cel

l ind

ex

siNTC

*** P = 0.0007

***siCASC15-S (exon 4)siPLK1

0 24 48 72 96 1200

1

2

3

4

Time (h)

Cel

l ind

ex

siNTCsiCASC15-S (exon 1)siPLK1

***

*** P < 0.0001

D

Figure 5.CASC15-S depletion induces amore aggressive phenotype in neuroblastoma cells. A,CASC15-S expressionwas investigated by qRT-PCR in a panel of neuroblastomacell lines (n ¼ 21) and normalized relative to the geometric mean of GUSB, HPRT, and TBP housekeeping genes. CASC15-S demonstrated a wide range ofexpression across neuroblastoma cell lines. B and C, SK-N-BE2 neuroblastoma cells transiently transfected with siRNA targeting an exon common to both CASC15and CASC15-S isoforms (exon 12 or exon 4, respectively; B), or specifically targeting only the unique exon of CASC15-S (exon 1; C), show a significant increase inproliferative rate. D, stable depletion of CASC15-S in SK-N-BE2 cells was achieved with lentiviral transduction of shRNA and recapitulated the increased growthobserved with transient knockdown. E, forced ectopic expression of CASC15-S cDNA was able to rescue the growth characteristics of SK-N-BE2 shCASC15-S cells,reverting their growth pattern to that of wild-type cells. F, morphologic observation of SK-N-BE2 cells stably depleted of CASC15-S showed that cells weresubstantially larger than control cells (scale bar, 100 mm). G, cell area was measured in biologic triplicate (n ¼ 10 for each replicate) and quantified in ImageJ,where the area of shCASC15 cells was found to be 3.1-fold increased over controls (�� , P < 0.001). siNTC, nontargeting control (negative control); siPLK1, polo-likekinase 1 (positive control).






several proneural gene family members with known roles inneurogenesis and differentiation such as neurogenic differentia-tion 1 (NEUROD1, P ¼ 1.1 � 10�4), neural precursor cell-expressed, developmentally downregulated gene 9 (NEDD9, P¼ 5.8� 10�4), and neurogenin 2 (NEUROG2, P¼ 5.3� 10�4). Ina manner identical to the SK-N-SH gene expression comparison,we utilized GO analysis of these differentially regulated genesets,with the toppathway, "cellular process" arising due to enrichmentof the cell differentiation pathway node contained within thissubset (Fig. 6E).

Finally, we submitted the differentially expressed gene listsfrom siCASC15-S SK-N-SH and shCASC15-S SK-N-BE2 cells toIngenuity Pathway Analysis (IPA). Neuroblastoma cells deplet-ed of CASC15-S demonstrated highly significant upregulationof pathways involved in cell migration, proliferation, andmetastasis (Fig. 6F, top), and substantial decreases in proneuralgene signatures (Fig. 6F, bottom). Taken together, these datasuggest that loss of CASC15-S shifts the neuroblastoma gene

expression away from a well-differentiated neural phenotypeand promotes increased expression of cellular adhesion andmigratory genes, a finding consistent with our phenotypic andmorphologic observations.

DiscussionHigh-risk neuroblastoma remains a major challenge due to a

relative paucity of somatic mutations hindering the developmentof targeted therapies (1). The identification of mutations in ALKand PHOX2B has helped explain the origin of familial neuro-blastoma; however, an understanding of the basis of sporadicdisease has only recently begun to come into focus (40, 41). Herewe identify and demonstrate CASC15-S as a neuroblastomasuppressor gene via a post-GWAS mechanistic evaluation of acomplex region of the human genome. Despite this robust asso-ciation, however, eQTL analyses correlating risk genotypes withtranscript expression failed to reach statistical significance, a likely

A

Empty vector shCASC15-S

CB

E

FDiseases or Pfunc�ons Ac�va�on z-score # Molecules

Cell movement 01x56.2 -14 5.283 113Migra�on of cells 01x48.2 -14 4.71 105Homing of cells 01x70.2 -10 4.36 42

01x03.2Angiogenesis -11 3.718 50Development of blood vessel 3.92x10-13 3.657 59

01x66.2Vasculogenesis -14 3.338 57Cell movement of tumor cell lines 1.63x10-11 2.972 55Prolifera�on of connec�ve �ssue cells 2.39x10-12 2.762 48

01x44.1sisatsateM -12 2.081 51Prolifera�on of tumor cells 1.07x10-12 0.44 59Organiza�on of cytoskeleton 7.55x10-7 -2.187 37Organiza�on of cytoplasm 2.06x10-6 -2.195 38Extension of cellular protrusions 2.42x10-4 -2.202 9Migra�on of neural stem cells 5.06x10-8 -2.219 5Differen�a�on of neurons 8.31x10-4 -2.303 13Migra�on of neurons 5.65x10-8 -2.369 15Flux of Ca2 01x53.5+ -3 -2.395 9Migra�on of stem cells 1.75x10-6 -2.433 6Ion homeostasis of cells 2.71x10-3 -2.577 15Outgrowth of neurites 1.35x10-4 -2.801 15

D15.30% locomotion15.11% response to stimulus14.71% single-organism process11.82% multicellular organismal process7.95% developmental process6.79% biologic adhesion5.27% cellular process5.26% biologic regulation3.70% cellular component morphogenesis3.44% immune system process2.90% growth2.76% reproduction4.99% other

SK-N-SHsiCASC15-S

16.05% cellular process14.79% single-organism process12.26% developmental process10.31% cellular component organization10.08% multicellular organismal process9.53% biologic regulation5.53% response to stimulus5.16% localization3.29% signaling3.22% metabolic process2.90% locomotion2.11% multi-organism process1.84% reproductive process1.14% growth1.79% other

SK-N-BE2shCASC15-S

423630241812600

50

100

Time (h)

% W

ound

hea

led

SK-N-BE2 EVSK-N-BE2 shCASC15

P = 0.0006

242016128400

50

100

Time (h)

% W

ound

hea

led

SK-N-SH shCASC15-S

SK-N-SH EV

P < 0.0001

Figure 6.CASC15-S regulates a subset of genes involved in neural differentiation and neuroblastoma tumorigenesis. A, SK-N-BE2 cells constitutively depleted ofCASC15-S demonstrated an increased migratory capacity in wound-healing assays (t ¼ 24 hours). B and C, linear regression comparison of wound closure atregular intervals demonstrates a clear enhancement of migration in silenced SK-N-BE2 (B) or SK-N-SH (C) cells (EV, empty vector). D, GO analysis of SK-N-SHcells following CASC15-S depletion via siRNA at 48 hours. The top-level biologic processes are shown by percentage of enrichment signal, with bothlocomotion and cellular adhesion gene sets exhibiting over representation. E, a similar analysis was carried out for SK-N-BE2 cells stably expressing an shRNAconstruct targeting CASC15-S. The top enrichment observed in these cells, "cellular process," was primarily the result of the cell differentiation gene subset. F,most significantly altered pathways in SK-N-SH and SK-N-BE2 neuroblastoma cells following depletion of CASC15-S and subjected to Ingenuity PathwayAnalysis. Pathways shown were the top gene signatures to arise from differential analysis and indicate activation of cellular programs of proliferation,migration, and metastasis (red), as well as downregulation of several pathways known to modulate neural-specific development (green).

Russell et al.





consequence of an underpowered patient data set and/or addi-tional mechanisms capable of impacting expression (such asadditional SNPs affecting expression, posttranslational modifica-tion/degradation, etc.). To address this shortcoming, we finemapped this locus, using orthogonal methodologies to identifythe likely disease causal SNP, rs9295534, which localizes to theclosest upstream enhancer of CASC15-S. Moreover, expression ofthe rs9295534 risk allele disrupts the enhancer function of thisregion, proving that genotype can indeed impact transcriptionalability at this locus. More importantly, we provide evidence for apotent effect of this lncRNAonneuroblastomadifferentiation andmigratory capacities, yielding mechanistic insights into why theGWAS signal at this locus is associatedwithmetastatic disease andpoor survival probability.

A growing body of work supports a defined role of lncRNAsas spatiotemporal-specific regulators of gene expression criticalfor ensuring proper differentiation during development. Thus,predominant expression of CASC15-S in brain (but not othertissues), the derivation of several cDNA clones from brainregions, and its proximity to a validated enhancer element(hs1335) strongly suggest that this lncRNA is uniquelyinvolved in neural tube development. The role of lncRNA-mediated tumorigenesis in embryonal cancers provides a log-ical hypothesis to contribute to the understanding of theetiology of neuroblastoma tumors, which are typically devoidof activating somatic mutations (1).

A preliminary working model of how this transcript functionsin neuroblastoma biology can be proposed from the functionaland expression data we have demonstrated.CASC15-S expressionstrongly correlates with disease stage and overall survival, andpatient tumors with high CASC15-S levels are enriched in genestypically lost in poor outcome neuroblastoma, demonstrating aprotective role forCASC15-S. Furthermore, ablation of CASC15-Sin neuroblastoma cell lines results in increased proliferative andmigratory capacities, upregulation of adhesion and migrationgene pathways, and a concomitant decrease in neural-specifictranscripts. Therefore, reduced CASC15-S expression as the resultan inherited polymorphism would impact neural crest cellularlineage commitment and predispose these cells to undergomalig-nant transformation.

The phenotypic and gene signatures changes we observedsignify that CASC15-S is responsible for maintaining a moredifferentiated and benign cell state, with CASC15-S loss leadingto a poorly differentiated phenotype and expression of genesassociated with transformed cells. It has been recently shown thatCASC14 (renamed NBAT-1), although very lowly expressed,exerts a similar phenotype (42), suggesting that many lncRNAsin this region may cooperate. The precise mechanism by whichCASC15-S (andpotentially other transcripts near this locus) exertsits effect is currently under investigation, although it is likely thatthe observed transcriptional changes may be the result of mod-ulation of cis elements. One possibility, given its proximity to avalidated enhancer (hs1335) with neural tube expression and itsposition downstream of the developmental regulatory gene

SOX4, is that CASC15-S functions as an enhancer RNA (43).Future studies to identify direct interaction partners ofCASC15-S will undoubtedly strengthen our understanding oflncRNA function and yield key insights into neuroblastomatumorigenesis.

In summary, our findings support a recent and growing body ofevidence that convincingly demonstrates involvement of thenoncoding genome in the tumorigenesis of pediatric cancers ingeneral, and neuroblastoma in particular (42, 44–47). While wecertify CASC15-S as neuroblastoma suppressor gene, we still havenot elucidated all of the stochastic and/or epigenetic events thatselect for CASC15-S repression. Future studies focused on engi-neering CASC15-S depletion in vivo will further explore mechan-isms for tumor initiation and progression as well as screen forsynthetic lethal interactions to be leveraged therapeutically. Final-ly, this work provides the first highly significant GWAS-supportedidentification of lncRNAs in neuroblastoma, thus further stimu-lating exploration of this class of regulatory RNAs in humancancer etiology and clonal evolution.

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

Authors' ContributionsConception and design: M. Russell, J.R. Alvarez-Dominguez, J.M. Maris,K.A. ColeDevelopment of methodology: M. Russell, A. Penikis, D. Oldridge, J.S. Wei,R.C. Seeger, S.J. Diskin, J.M. Maris, K.A. ColeAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): M. Russell, A. Penikis, O. Padovan, J.S. Wei,R.C. Seeger, S. Asgharzadeh, J. Khan, J.M. MarisAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): M. Russell, D. Oldridge, J.R. Alvarez-Dominguez,L. McDaniel, P. Raman, Y. Li, J. Gnanachandran, S. Zhang, S. Asgharzadeh,S.J. Diskin, J.M. Maris, K.A. ColeWriting, review, and/or revision of the manuscript: M. Russell, D. Oldridge,J.R. Alvarez-Dominguez, Y. Li, S. Asgharzadeh, S.J. Diskin, J.M. Maris, K.A. ColeAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): M. Russell, A. Penikis, M. Diamond, J.S. Wei,J. Gnanachandran, J.M. MarisStudy supervision: J.M. Maris, K.A. Cole

AcknowledgmentsThe authors acknowledge the Children's Oncology Group (U10-CA98543)

for providing blood and tumor specimens from neuroblastoma patients.

Grant SupportThis work was supported by grant K08CA136979 (K.A. Cole) and Alex's

Lemonade Stand Foundation (K.A. Cole and M.R. Russell).The costs of publication of this article were defrayed in part by the

payment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received December 10, 2014; revised May 4, 2015; accepted May 5, 2015;published OnlineFirst June 22, 2015.

References1. Maris JM. Recent advances in neuroblastoma. N Engl J Med 2010;362:

2202–11.2. Brodeur GM. Neuroblastoma: biological insights into a clinical enigma.

Nat Rev Cancer 2003;3:203–16.

3. Cheung N-K V, Dyer M. Neuroblastoma: developmental biology, cancergenomics and immunotherapy. Nat Rev Cancer 2013;13:397–411.

4. Takahashi Y, Sipp D, Enomoto H. Tissue interactions in neural crest celldevelopment and disease. Science 2013;341:860–3.






5. Maris JM, Mosse YP, Bradfield JP, Hou C, Monni S, Scott RH, et al.Chromosome 6p22 locus associated with clinically aggressive neuroblas-toma. N Engl J Med 2008;358:2585–93.

6. CapassoM, DevotoM,HouC, Asgharzadeh S, Glessner JT, Attiyeh EF, et al.Common variations in BARD1 influence susceptibility to high-risk neu-roblastoma. Nat Genet 2009;41:718–23.

7. Diskin SJ, Hou C, Glessner JT, Attiyeh EF, Laudenslager M, Bosse K, et al.Copy number variation at 1q21.1 associated with neuroblastoma. Nature2009;459:987–91.

8. Nguyen le B, Diskin SJ, CapassoM,Wang K, DiamondMA, Glessner J, et al.Phenotype restricted genome-wide association study using a gene-centricapproach identifies three low-risk neuroblastoma susceptibility loci. PLoSGenet 2011;7:e1002026.

9. Wang K, Diskin SJ, Zhang H, Attiyeh EF, Winter C, Hou C, et al. Integrativegenomics identifies LMO1 as a neuroblastoma oncogene. Nature 2011;469:216–20.

10. Diskin SJ, Capasso M, Schnepp RW, Cole KA, Attiyeh EF, Hou C, et al.Common variation at 6q16 within HACE1 and LIN28B influences sus-ceptibility to neuroblastoma. Nat Genet 2012;44:1126–30.

11. Latorre V, Diskin SJ, Diamond MA, Zhang H, Hakonarson H, Maris JM,et al. Replication of neuroblastoma snp association at the BARD1 locus inAfrican-Americans. Cancer Epidemiol Biomarkers Prev 2012;21:658–63.

12. Capasso M, Diskin SJ, Totaro F, Longo L, De Mariano M, Russo R, et al.Replication of GWAS-identified neuroblastoma risk loci strengthens therole of BARD1 and affirms the cumulative effect of genetic variations ondisease susceptibility. Carcinogenesis 2013;34:605–11.

13. Molenaar JJ, Koster J, Zwijnenburg DA, van Sluis P, Valentijn LJ, van derPloeg I, et al. Sequencing of neuroblastoma identifies chromothripsis anddefects in neuritogenesis genes. Nature 2012;483:589–93.

14. Sausen M, Leary RJ, Jones S, Wu J, Reynolds CP, Liu X, et al. Integratedgenomic analyses identify ARID1A and ARID1B alterations in the child-hood cancer neuroblastoma. Nat Genet 2013;45:12–7.

15. Pugh TJ, Morozova O, Attiyeh EF, Asgharzadeh S, Wei JS, Auclair D, et al.The genetic landscape of high-risk neuroblastoma. Nat Genet 2013;45:279–84.

16. Cheung N, Zhang J, Lu C, Parker M. Association of age at diagnosis andgeneticmutations in patients with neuroblastoma. JAMA2012;307:1062–71.

17. Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, Hunter DJ,et al. Finding the missing heritability of complex diseases. Nature2009;461:747–53.

18. Bertone P, Stolc V, Royce TE, Rozowsky JS, Urban AE, Zhu X, et al. Globalidentification of human transcribed sequences with genome tiling arrays.Science 2004;306:2242–6.

19. Carninci P, Kasukawa T, Katayama S,Gough J, FrithMC,MaedaN, et al. Thetranscriptional landscape of the mammalian genome. Science 2005;309:1559–63.

20. Mercer T, Dinger M, Mattick J. Long non-coding RNA insights into func-tions. Nat Rev Genet 2009;10:155–9.

21. Rinn JL, Chang HY. Genome regulation by long noncoding RNAs. AnnuRev Biochem 2012;81:145–66.

22. Lee JT. Epigenetic regulation by long noncoding RNAs. Science 2012;338:1435–9.

23. Mercer TR, Mattick JS. Structure and function of long noncoding RNAs inepigenetic regulation. Nat Struct Mol Biol 2013;20:300–7.

24. Chu C, Qu K, Zhong F, Artandi S, Chang H. Genomic maps of lincRNAoccupancy reveal principles of RNA chromatin interactions. Mol Cell2011;44:667–78.

25. Ponjavic J, Oliver PL, Lunter G, Ponting CP. Genomic and transcriptionalco-localization of protein-coding and long noncoding RNA pairs in thedeveloping brain. PLoS Genet 2009;5:e1000617.

26. Knauss JL, Sun T. Regulatory mechanisms of long noncoding RNAs invertebrate central nervous system development and function. Neurosci-ence 2013;3:200–14.

27. Gibb EA, Brown CJ, LamWL. The functional role of long non-coding RNAin human carcinomas. Mol Cancer 2011;10:38.

28. Prensner JR, Chinnaiyan AM. The emergence of lncRNAs in cancer biology.Cancer Discov 2011;1:391–407.

29. Prensner JR, Iyer MK, Sahu A, Asangani IA, Cao Q, Patel L, et al. The longnoncoding RNA SChLAP1 promotes aggressive prostate cancer and antag-onizes the SWI/SNF complex. Nat Genet 2013;45:1392–8.

30. Gutschner T, H€ammerle M, Eissmann M, Hsu J, Kim Y, Hung G, et al. Thenoncoding RNAMALAT1 is a critical regulator of themetastasis phenotypeof lung cancer cells. Cancer Res 2013;73:1180–9.

31. Federico S, Brennan R, Dyer M. Childhood cancer and developmentalbiology: a crucial partnership. Top Dev Biol 2011;901:1–10.

32. http://target.nci.nih.gov/dataMatrix/TARGET_DataMatrix.html33. Raj A, van den Boogard P, Rifkin SA, van Oudenaarden A, Tyagi S. Imaging

individual mRNA molecules using multiple singly labeled probes. NatureMethods 2008;5:877–79.

34. DelaneauO,Marchini J, Zagury JF. A linear complexity phasingmethod forthousands of genomes. Nat Methods 2011;9:179–81.

35. Howie BN, Donnelly P, Marchini J. A flexible and accurate genotypeimputation method for the next generation of genome-wide associationstudies. PLoS Genet 2009;5:e1000529.

36. Marchini J, Howie B, Myers S, McVean G, Donnelly P. A new multipointmethod for genome-wide association studies by imputation of genotypes.Nat Genet 2007;39:906–13.

37. Visel A,Minovitsky S,Dubchak I, Pennacchio LA. VISTA Enhancer Browser-a database of tissue-specific human enhancers. Nucleic Acids Res 2007;35:D88–92

38. Zhang X, Bailey SD, Lupien M. Laying a solid foundation for Manhattan–'setting the functional basis for the post-GWAS era'. Trends Genet 2014;30:140–9.

39. Asgharzadeh S, Pique-Regi R, Sposto R, Wang H, Yang Y, Shimada H,et al. Prognostic significance of gene expression profiles of metastaticneuroblastomas lacking MYCN amplification. J Natl Cancer Inst 2006;98:1193–203.

40. Moss�e YP, Laudenslager M, Longo L, Cole KA, Wood A, Attiyeh EF, et al.Identification of ALK as a major familial neuroblastoma predispositiongene. Nature 2008;455:930–5.

41. Mosse YP, Laudenslager M, Khazi D, Carlisle AJ, Winter CL, Rappaport E,et al. Germline PHOX2B mutation in hereditary neuroblastoma. Am JHum Genet 2004;75:727–30.

42. Pandey GK, Mitra S, Subhash S, Hertwig F, Kanduri M, Mishra K, et al. Therisk-associated long noncoding RNA NBAT-1 controls neuroblastomaprogression by regulating cell proliferation and neuronal differentiation.Cancer Cell 2014;26:722–37.

43. �romUA,Derrien T, BeringerM,Gumireddy K,Gardini A, Bussotti G, et al.Long noncoding RNAs with enhancer-like function in human cells. Cell2010;143:46–58.

44. Barnhill LM,Williams RT, CohenO, Kim Y, Batova A,Mielke JA, et al. HighexpressionofCAI2, a 9p21-embedded longnoncodingRNA, contributes toadvanced-stage neuroblastoma. Cancer Res 2014;74:3753–63.

45. Liu PY, ErriquezD,Marshall GM, Tee AE, Polly P,WongM, et al. Effects of anovel long noncoding RNA, lncUSMycN, on N-Myc expression and neu-roblastoma progression. J Natl Cancer Inst 2014;106:1–11.

46. Tee A, Ling D, Nelson C, Atmadibrata B, Dinger M, Xu N, et al. The histonedemethylase JMJD1A induces cellmigration and invasion by up-regulatingthe expression of the long noncoding RNA MALAT1. Oncotarget 2014;5:1793–804.

47. Atmadibrata B, Liu PY, Sokolowski N, Zhang L, Wong M, Tee AE, et al. Thenovel long noncoding RNA linc00467 promotes cell survival but is down-regulated by N-Myc. PLoS ONE 2014;9:e88112.

48. Suv�a ML, Rheinbay E, Gillespie SM, Patel AP, Wakimoto H, Rabkin SD,et al. Reconstructing and reprogramming the tumor-propagating potentialof glioblastoma stem-like cells. Cell 2014;157:580–94.

49. Rheinbay E, Suv�a ML, Gillespie SM, Wakimoto H, Patel AP, Shahid M,et al. An aberrant transcription factor network essential for Wntsignaling and stem cell maintenance in glioblastoma. Cell Rep 2013;3:1567–79.


Russell et al.




2015;75:3155-3166. Published OnlineFirst June 22, 2015.Cancer Res Mike R. Russell, Annalise Penikis, Derek A. Oldridge, et al. Neuroblastoma Susceptibility LocusCASC15-S Is a Tumor Suppressor lncRNA at the 6p22

Updated version

10.1158/0008-5472.CAN-14-3613doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2015/06/23/0008-5472.CAN-14-3613.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/75/15/3155.full#ref-list-1

This article cites 48 articles, 8 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/75/15/3155.full#related-urls

This article has been cited by 2 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://cancerres.aacrjournals.org/content/75/15/3155To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-14-3613

http://cancerres.aacrjournals.org/content/suppl/2015/06/23/0008-5472.CAN-14-3613.DC1

http://cancerres.aacrjournals.org/content/75/15/3155.full#ref-list-1

http://cancerres.aacrjournals.org/content/75/15/3155.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/75/15/3155


casc15-s is a tumor suppressor lncrna at the 6p22 ... · tumor and stem cell biology casc15-s is a...

Documents