Open Full Page

Published OnlineFirst October 21, 2014; DOI: 10.1158/1078-0432.CCR-14-1324

Biology of Human Tumors Clinical Cancer Research Molecular Characterization of Choroid Plexus Tumors Reveals Novel Clinically Relevant Subgroups Diana M. Merino1, Adam Shlien1, Anita Villani1, Malgorzata Pienkowska1, Stephen Mack1, Vijay Ramaswamy1, David Shih1, Ruth Tatevossian2, Ana Novokmet1, Sanaa Choufani1, Rina Dvir3, Myran Ben-Arush4, Brent T. Harris5, Eugene I. Hwang6, Rishi Lulla7, Stefan M. Pﬁster8, Maria Isabel Achatz9, Nada Jabado10, Jonathan L. Finlay11, Rosanna Weksberg1, Eric Bouffet1, Cynthia Hawkins1, Michael D. Taylor1, Uri Tabori1, David W. Ellison2, Richard J. Gilbertson2, and David Malkin1

Abstract

Purpose: To investigate molecular alterations in choroid plexus were observed in 60% of CPCs. Investigating the number of tumors (CPT) using a genome-wide high-throughput approach to mutated copies of p53 per sample revealed a high-risk group of identify diagnostic and prognostic signatures that will refine patients with CPC carrying two copies of mutant p53, who tumor stratification and guide therapeutic options. exhibited poor 5-year event-free (EFS) and overall survival (OS) Experimental Design: One hundred CPTs were obtained compared with patients with CPC carrying one copy of mutant from a multi-institutional tissue and clinical database. Copy- p53 (OS: 14.3%, 95% confidence interval, 0.71%–46.5% vs. number (CN), DNA methylation, and gene expression signa- 66.7%, 28.2%–87.8%, respectively, P ¼ 0.04; EFS: 0% vs. tures were assessed for 74, 36, and 40 samples, respectively. 44.4%, 13.6%–71.9%, respectively, P ¼ 0.03). CPPs and aCPPs Molecular subgroups were correlated with clinical parameters exhibited favorable survival. and outcomes. Discussion: Our data demonstrate that differences in CN, gene Results: Unique molecular signatures distinguished choroid expression, and DNA methylation signatures distinguish CPCs plexus carcinomas (CPC) from choroid plexus papillomas (CPP) from CPPs and aCPPs; however, molecular similarities among the and atypical choroid plexus papillomas (aCPP); however, no papillomas suggest that these two histologic subgroups are indeed significantly distinct molecular alterations between CPPs and a single molecular entity. A greater number of copies of mutated aCPPs were observed. Allele-specific CN analysis of CPCs revealed TP53 were significantly associated to increased tumor aggres- two novel subgroups according to DNA content: hypodiploid and siveness and a worse survival outcome in CPCs. Collectively, hyperdiploid CPCs. Hyperdiploid CPCs exhibited recurrent these findings will facilitate stratified approaches to the clinical acquired uniparental disomy events. Somatic mutations in TP53 management of CPTs. Clin Cancer Res; 21(1); 184–192. 2014 AACR.

Introduction described: choroid plexus papilloma (CPP, WHO grade I), atypical choroid plexus papilloma (aCPP, WHO grade II), and choroid Choroid plexus tumors (CPT) are rare intraventricular neo- plexus carcinoma (CPC, WHO grade III). Long-term survival of plasms accounting for up to 20% of brain tumors in children CPPs is favorable with surgical resection alone (>90%; ref. (3)). under 2 years of age (1, 2). Three histologic subgroups have been Conversely, CPCs exhibit a dismal prognosis, with an overall survival of about 30% (4–6). Despite aggressive treatment pro- 1The Hospital for Sick Children, Toronto, Ontario, Canada. 2St. Jude Children's Research Hospital, Memphis, Tennessee. 3Tel Aviv Sourasky tocols, including surgical resection and combination of chemo- Medical Center, Tel Aviv, Israel. 4Rambam Health Care Campus, Haifa, and radiation therapy (6), the clinical behavior of CPCs is variable Israel. 5Georgetown University Medical Center,Washington, DC. 6Chil- and most of the few survivors exhibit long-term cognitive and dren's National Medical Center, Washington, DC. 7Ann & Robert H. ﬁ Lurie Children's Hospital of Chicago, Chicago, Illinois. 8German Cancer developmental de cits (6). aCPP, a recently described pathologic Research Center (DKFZ), Heidelberg,Germany. 9A.C.Camargo Cancer subgroup, exhibits an intermediate degree of mitotic activity and Center, Sao Paulo, Brazil. 10Montreal Children's Hospital, Montreal, outcome (7, 8); however, some cases may be difﬁcult to distin- 11 Quebec, Canada. Children's Hospital of Los Angeles, Los Angeles, guish from CPC by histology alone (9). California. Over 50% of CPC tumors carry somatic TP53 mutations, and Note: Supplementary data for this article are available at Clinical Cancer TP53 mutant CPCs have been associated with increased genetic Research Online (http://clincancerres.aacrjournals.org/). tumor instability and worse prognosis (5). Germline TP53 Corresponding Author: David Malkin, The Hospital for Sick Children, Division of mutations have also been observed in patients with CPC as Haematoloty/Oncology, 555 University Avenue, Toronto, ON M5V0A4, Canada. CPC is one of the hallmark cancers of the Li–Fraumeni syn- Phone: 416-813-5348; Fax: 1-416-813-5327; E-mail: [email protected] drome (LFS), a familial cancer syndrome in which affected doi: 10.1158/1078-0432.CCR-14-1324 family members harbor a mutant copy of the TP53 tumor 2014 American Association for Cancer Research. suppressor gene.

184 Clin Cancer Res; 21(1) January 1, 2015

Genomic Proﬁling of Choroid Plexus Tumors

from fresh-frozen samples, and the RecoverAll Total Nucleic Acid Translational Relevance Isolation Kit for FFPE (Ambion) from FFPE samples. Total RNA This report is the first to dissect the aberrant complexity in was isolated from fresh-frozen samples using the TRIzol method copy number, methylation, and gene expression of one of the (Invitrogen) according to the manufacturer's instructions. largest cohort of pediatric choroid plexus tumors (CPT). Our findings revealed molecular homogeneity among choroid TP53 sequencing plexus papillomas (CPP) and atypical choroid plexus papil- Sequencing of the coding region of TP53 (exons 2–11) was lomas (aCPP), reflecting the favorable survival of these performed in the molecular diagnostic laboratory at The Hospital patients and suggesting these histologically distinct subgroups for Sick Children (Toronto) by direct Sanger sequencing of whole are a single-tumor entity. Choroid plexus carcinomas (CPC) genome DNA as previously described (5). were significantly different from CPPs and aCPPs. Moreover, Microarray processing and bioinformatics analysis CPCs exhibited molecular heterogeneity, and patient out- Forty RNA samples were hybridized to GeneChip Human Exon comes varied widely. We identified novel CPC subgroups with 1.0ST gene expression microarrays (Affymetrix), and 36 DNA significantly distinct copy-number signatures, suggesting dif- samples were hybridized to Illumina 450 K Infinium methylation ferent mechanisms drive CPC development. We identified that bead arrays (Illumina) as per manufacturer's instructions. An patient overall and event-free survival significantly decreased initial set of 55 tumor DNA samples was hybridized to with an increasing number of mutated copies of p53. By Genome-Wide Human SNP Array 6.0 (Affymetrix), whereas an defining the molecular landscape of CPTs, this study has independent set of 20 tumor DNA samples was hybridized to provided a comprehensive molecular background on which Affymetrix OncoScan FFPE Express 2.0 arrays. One technical to explore mechanisms of tumorigenesis and develop strati- replicate was included in both genotyping platforms and analyz- fied approaches to the clinical management of CPTs. ed for copy-number call consistency. One FFPE-derived sample was removed from further copy number analysis due to poor microarray quality. Cytogenetic studies of central nervous system (CNS) tumors Partek Genomics Suite 6.5 (PGS; Partek Inc.) and BioDiscovery have revealed high chromosomal instability in more than 90% of Nexus Copy Number software (Discovery Edition 7.0; BioDiscovery) CPTs analyzed (Supplementary Table S1). Defining the molecular were used for copy-number analysis as previously described landscape of CPTs and identifying actionable molecular aberra- (5; see Supplementary Appendix). Copy-number changes encom- tions have been challenging due in part to the limited number of passing more than 75% of the chromosome were called whole- patients and high-quality samples available for genome-wide chromosome aberrations. Allele-specific copy number analysis of studies. Here, we use an integrative molecular approach to char- tumors (ASCAT) was performed in R as previously described (10) acterize the genomic, transcriptomic, and epigenomic landscape and verified by Nexus software. Tumor ploidy, where hypodiploid of the largest cohort of CPTs to date. The information derived <1.90 and hyperdiploid >2.10, heterogeneity, and allelic imba- from these analyses creates a molecular foundation on which to lances were inferred from the output. ASCAT failed to resolve the develop approaches to improve the clinical management of this ploidy of two samples with very low aberrant fraction, so these devastating disease. were excluded from further analysis. Clustering of gene expression and methylation data was investigated in R (version 3.0.1) by un- Materials and Methods supervised hierarchical clustering (uHCL), non-negative matrix factorization (NMF), and PVCLUST algorithms. Differential gene Patients and sample preparation expression analysis was conducted with PGS 6.5 (see Supplemen- CPT samples and/or clinical data were collected from institu- tary Appendix). Gene set enrichment analysis (GSEA) was perform- tions in Canada, the United States of America, Brazil, Israel, and ed as previously described (11), and its visualization was obtained Germany (see Supplementary Appendix) in accordance with each by Cytoscape and Enrichment Map using an in-house–curated institution's Research Ethics Board. Informed consent was database containing freely available NCI, Kyoto Encyclopedia of obtained from the parents/legal guardians of all patients. We Genes and Genomes (KEGG), Protein families database (PFAM), studied 100 unique tumor samples (58 CPC, 30 CPP, and 12 Biocarta, and GO databases as described in Witt and colleagues (12). aCPP) from 91 pediatric patients (ages 0.03–16.50 years old) for Differences in DNA methylation status were analyzed with the which TP53 sequence data were available (Supplementary Table Illumina GenomeStudio software (see Supplementary Appendix). S2). Pathologic review of CPTs was conducted by C. Hawkins, D. All probe sets were annotated according to the human genome W. Ellison, and B.T. Harris when samples were available. In all build hg19 (GRCh37). Microarray data can be accessed from GEO other institutions, expert neuropathologists critically examined GSE60886. each case. In 15 CPC cases, immunohistochemical analysis of hHF5/INI1 was conducted and revealed immunopositivity Statistical analysis excluding the diagnosis of atypical teratoid/rhabdoid tumors. Statistical analyses of copy number and gene expression were Nucleic acids were derived from fresh-frozen (n ¼ 75), optimal performed in PGS 6.5, whereas methylation was analyzed in cutting temperature compound (n ¼ 9), and formalin-fixed Genome Studio. Patient survival was calculated in StataSE (ver- paraffin-embedded (FFPE; n ¼ 12) samples. We received isolated sion 12), whereas other statistical analyses were conducted in R tumor DNA from 4 samples. Twenty-two nucleic acid samples (version 3.0.1; see Supplementary Appendix). Survival estimates exhibited suboptimal quality and/or quantity, leaving 78 high- for tumor subgroups, and for CPCs by TP53 and ploidy status quality samples from 73 patients for analysis (Supplementary Fig. were generated using the Kaplan–Meier method, and curves were S1). Detailed clinical data were obtained for 68 patients. Tumor compared using a log-rank test. Overall survival (OS) measured DNA was extracted using standard phenol-chloroform extraction time from initial diagnosis to death from any cause or last follow-

www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 185

Merino et al.

up as of December 1, 2013. Event-free survival (EFS) measured greater than in CPPs and aCPPs (Mann-Whitney test, P < 1.00 time from initial diagnosis to tumor progression, recurrence, or 10 4). Allele-specific copy-number analysis allowed us to inves- death from any cause. tigate the allelic ratios in our samples. This technique revealed a striking pattern of copy-number–neutral LOH. This phenomenon is commonly observed in cancer cells and may also be referred to Results as acquired uniparental disomy (aUPD), wherein a chromosome Genomic, transcriptomic, and DNA methylation profiling of pair is homozygous, thus having two copies of the same allele CPTs reveals significant segregation of CPCs from CPPs and (13). CPCs exhibited frequent aUPD events with an average of aCPPs 2.31 aUPD events per sample, whereas the phenomenon occurred Unsupervised clustering analyses performed with gene expres- less frequently in CPPs and aCPPs (average 0.32 and 1 events per sion and methylation data revealed clear segregation of CPCs sample, respectively). from CPPs and aCPPs (Fig. 1). NMF analysis of gene expression Analysis of clinical variables between the three histologic sub- (Fig. 1A) and methylation (Fig. 1B) data demonstrated greatest groups revealed no significant difference of age at diagnosis difference between two subgroups (FDR-corrected P ¼ 2.54 (Kruskal–Wallis test, P ¼ 0.30) or ratio of males to females 10 7 and P ¼ 1.02 10 34, respectively), segregating CPCs from (two-way ANOVA, P ¼ 0.26; Supplementary Table S2). Survival CPPs and aCPPs. This significant molecular stratification was also outcomes for CPCs were significantly worse than for CPPs and observed using a smaller number of probe sets for gene expression aCPPs. Five-year OS for CPCs was 56.3% [95% confidence inter- and methylation differences analyzed by PVCLUST (Supplemen- val (CI), 36.5%–72.0%], compared with 92.9% (59.1%–99.0%) tary Fig. S2). The concordance between tumors stratified by gene and 100% for CPPs and aCPPs, respectively (P ¼ 0.03; Fig. 3B). expression and methylation was significant (Rand index ¼ 0.73, P Only one patient with CPP died due to complications from a < 1.0 10 4) and revealed consistent molecular segregation of concurrent diagnosis of ependymoma. Five-year EFS for CPCs was CPTs into unique molecular subgroups. 39.7% (95% CI, 22.8%–56.2%), compared with 87.4% (58.1%– Although copy-number analysis revealed widespread chromo- 91.9%) and 70.0% (22.5%–91.8%) for CPPs and aCPPs, respec- somal instability in all tumor subgroups (Fig. 2), a distinct tively (P ¼ 4.90 10 3; Fig. 3C). signature characterized by increased frequency of chromosome- wide gains and losses was observed in CPCs (average 5.31 CPPs and aCPPs share similar molecular signatures, which chromosomes gained and 6.17 lost per CPC), compared with correlate with favorable survival outcomes increased frequency of chromosome-wide gains but very few Analyzing CPPs and aCPPs independently from CPCs revealed losses in CPPs and aCPPs (average 6.92 chromosomes gained a striking molecular similarity between the papilloma subgroups and 0.32 lost per CPP, and 9.64 vs. 0.09 per aCPP; Fig. 2). The (Fig. 4). Unsupervised clustering analysis demonstrated that CPPs frequency of chromosome-wide losses in CPCs was significantly and aCPP did not segregate according to differences in gene

Figure 1. Unsupervised clustering of (A) gene expression normalized intensities and (B) methylation beta values by NMF demonstrates significant segregation of CPCs (red) from CPPs (yellow) and aCPPs (light blue). No segregation was observed between CPPs and aCPPs. NMF was conducted using 5,000 probe sets with the largest median absolute deviation. This clustering algorithm identified the most significant measures of similarity (cophenetic coefficient) when the data were at k ¼ 2 (2 clusters). In the matrix, red represents the highest measure of similarity (1), whereas blue/purple represents the lowest measure of similarity (0). Any other colors within the matrix represent a spectrum of changing measures of similarity, from red to blue/purple. Colors: TP53 mutation status: black, TP53 mutant; white, TP53 wild type.

186 Clin Cancer Res; 21(1) January 1, 2015 Clinical Cancer Research

Genomic Proﬁling of Choroid Plexus Tumors

Figure 2. Characterization of recurrent chromosome-wide gains (n > 2, gray) and losses (n < 2, black) for each tumor subtype (CPC: n ¼ 35, CPP: n ¼ 25, aCPP: n ¼ 11). P values were calculated using the frequency of chromosome-wide copy-number alterations per subgroup and the nonparametric Mann–Whitney test.

expression or methylation (Fig. 4A and B). Supervised analysis Fisher exact test, P < 0.0001; Fig. 3A). Moreover, significant using the Wilcoxon rank-sum test between CPPs and aCPPs enrichment in aUPD was observed in TP53 mutant hyperdiploid revealed no significant differences in gene expression or meth- CPCs compared with hyperdiploid CPCs with wild-type TP53 ylation (Fig. 4C and D). In addition, signatures of chromo- (Fisher exact test, P < 0.0001). aUPD was most frequently somal instability characterized by recurrent chromosome-wide observed in chromosome 17, affecting 29% (10/35) of CPCs. gains and very few losses were observed in both CPPs and We conducted GSEA to identify biologic pathways and pro- aCPPs; no significant differences in the frequency of chromo- cesses that are differentially expressed between hypodiploid and some-wide gains and losses were observed (Mann-Whitney test, hyperdiploid CPCs. GSEA revealed enrichment in RNA proces- P ¼ 0.32 and P ¼ 0.49, respectively; Fig. 4E). There were no sing, DNA replication and repair, and chromosome segregation in differences in age at diagnosis (Mann–Whitney test, P ¼ 0.45) hyperdiploid CPCs. Hypodiploid CPCs exhibited enrichment in or ratio of males to females (Fisher exact test, P ¼ 0.31) between cellular metabolism, signaling, and cell migration pathways, as CPPs and aCPPs. Moreover, survival outcomes for patients with well as leukocyte activation and proliferation (5% FDR, P < 0.05; CPP and aCPP were not significantly different (log-rank test, OS Supplementary Fig. S4). The patterns of enrichment observed P ¼ 0.51, EFS P ¼ 0.30). suggest hyperdiploids are more proliferative than hypodiploid CPCs, and that the latter tumors are undergoing a significant Ploidy analysis reveals novel CPC subgroups with unique immune response. A greater understanding of these distinct molecular alterations enrichment patterns will elucidate the mechanisms underlying Ploidy analysis revealed the presence of aneuploidy in 89% of the progression of these molecularly distinct CPC subgroups tumors (Supplementary Fig. S3). CPPs and aCPPs exhibited (Supplementary Fig. S4). There were no significant differences in ploidy greater than 2 (hyperdiploidy); however, CPCs exhibited DNA methylation between subgroups, although this may be due a wide distribution of ploidy values, with two significantly distinct to the low number of samples in the comparison groups (hypo- subgroups observed: hyperdiploid CPCs (average ploidy 2.76; diploid, n ¼ 4; hyperdiploid, n ¼ 9). range, 2.21–3.34) and hypodiploid CPCs (average ploidy 1.45; range, 1.25–1.71; Mann–Whitney test, P < 1.00 10 4). Only Increased number of copies of mutant TP53 is associated with three of 35 CPCs were diploid. Hypodiploid CPCs exhibited tumor aggressiveness and unfavorable survival outcomes recurrent chromosome-wide losses and very few gains with an Mutations in TP53 were assessed in our cohort by Sanger average of 12.71 chromosomes lost and 0.06 gained per tumor. sequencing. Sixty percent of CPCs (35/58) were mutant for Chromosome 3 was lost in all hypodiploid CPCs, with loss of TP53. Fifteen (15/58, 26%) of these samples belonged to 12 chromosomes 6, 22, 11, and 16 observed in more than 80% of patients with LFS carrying a germline mutation in TP53.Muta- tumors (Fig. 3A). Hyperdiploid CPCs exhibited a high frequency tions in TP53 were observed in both hypodiploid and hyper- of chromosome-wide gains and almost no losses (average 11.33 diploid CPCs; however, in our cohort, the frequency of TP53 and 0.33 chromosomes, respectively). Chromosomes 12, 7, 20, mutations was significantly greater in hypodiploid CPCs and 1 were gained in more than 80% of hyperdiploid CPCs (Fig. (15/17, 88%) than hyperdiploid CPCs (7/15, 47%; Fisher exact 3A). In addition to a high frequency of chromosomal gains, test, P ¼ 0.02). Diploid CPCs (n ¼ 3) were TP53 wild type hyperdiploid CPCs also exhibited aUPD more frequently than (Supplementary Table S3). No significant enrichment for hypodiploid CPCs (average of 4.93 affected chromosomes per patients with LFS was observed in either hypo- or hyperdiploid tumor compared with 0.35 chromosomes in hypodiploid CPCs; subgroups.

www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 187

Merino et al.

Figure 3. A, genome-wide characterization of chromosome-wide gains (red), losses (blue), and aUPD (teal) in 71 unique CPT samples. White squares represent unchanged chromosome-wide copy-number status. Chromosome-wide aberrations were deﬁned as aberrations that encompass more than 75% of the chromosome. Ploidy: green: hyperdiploid, blue: hypodiploid, tan: diploid. TP53 status: black, mutant; white, wild type. Kaplan–Meier curves depicting (B) OS and (C) EFS estimates of patients with CPT by diagnosis. Statistical values were obtained with the log-rank (Mantel–Cox) test.

Unsupervised clustering using gene expression and methyla- homozygous TP53 mutation status in all but one tumor sample tion data segregated CPCs into two significantly distinct clusters with a low aberrant cell fraction (46%), suggesting this sample (P ¼ 0.05; Fig. 5A and B). Although CPCs did not segregate was largely contaminated with normal cells. Seventy-five per- according to ploidy status, we observed two clusters, which were cent of samples with 2 copies of mutated p53 (9/12) exhibited significantly distinct according to TP53 status using DNA meth- aUPD in chromosome 17. Eighty-three percent of CPCs with ylation data (Fisher exact test, P ¼ 0.007), but did not reach 2 copies of mutated p53 (10/12) had missense mutations in the significance using gene expression data (Fisher exact test, P ¼ DNA-binding domain, whereas 1 sample had a missense muta- 0.089). LFS CPCs did not segregate from the spontaneous CPCs, tion in the SH3-like/Proline-rich domain and the other sample, suggesting no unique aberrations were present in tumors arising a splicing mutation. CPCs with 1 copy of mutated p53 had from patients with an inherited TP53 mutation. An in-depth missense mutations in the DNA binding (9/10) and tetrameriza- analysis of the type of TP53 mutations revealed that a few samples, tion (1/10) domains, and carried a single copy of chromosome which appeared to be miscategorized by unsupervised clustering, 17, exhibiting LOH of the entire chromosome. Three samples had an uncharacterized intronic alteration (c.28-28G>A/wt) and exhibited a heterozygous TP53 mutation status by sequencing, mutations outside the DNA-binding domain (i.e., c.290T>A in the which may be a result of normal cell contamination. Gene SH3-like/Proline-rich domain), which may account for differ- expression and methylation analyses revealed no significant dif- ences in the transcriptomic and epigenomic signature of these ferences among CPCs carrying 1 or 2 mutated copies of p53 samples (Fig. 5). because of the limited sample sizes (gene expression: 3 and 6 Combining TP53 sequencing results with allele-specific copy- samples, respectively; methylation: 1 and 6 samples, respectively). number status of chromosome 17 in 35 CPCs, we estimated the Examining clinical variables among CPCs revealed no differ- number of mutated copies of TP53. We found that 34.3% of CPCs ences in the age of diagnosis (Mann–Whitney t test, P ¼ 0.80 (12/35) had 2 copies of mutant p53, 28.6% (10/35) had 1 copy of and P ¼ 0.59) or the ratio of males to females (Fisher exact mutant p53, and 37.1% (13/35) had zero copies of mutant p53 test, P ¼ 1.0 and P ¼ 1.0) according to ploidy nor p53 (wild type). CPCs with 2 mutant copies of p53 exhibited a status, respectively (Supplementary Table S2). No significant

188 Clin Cancer Res; 21(1) January 1, 2015 Clinical Cancer Research

Genomic Proﬁling of Choroid Plexus Tumors

Figure 4. Unsupervised clustering of (A) gene expression normalized intensities and (B) methylation beta values demonstrates no segregation between CPPs (yellow) and aCPPs (light blue). Volcano plot comparing the number of significant differentially expressed genes (C) and significantly methylated regions (D) reveals no significant gene expression or methylation differences between the two subgroups (after FDR adjustment). Signatures of chromosomal instability (E), as measured by the number of chromosome-wide gains and losses in CPPs and aCPPs, were similar between the two subgroups and characterized by extensive chromosomal gains and very few losses. Black squares represent chromosome-wide aberrations, whereas white squares represent unchanged chromosome copy number. Gender was also depicted (pink: female, purple: male).

differences in OS or EFS estimates were observed between OS of patients with CPCs harboring wild-type TP53 (zero copies) patients with CPC exhibiting a hyper- or hypodiploid tumor was 88.9% (95% CI, 43.3%–98.4%) and EFS was 66.5% (32.9%– genome (P ¼ 0.82, P ¼ 0.94, respectively; Supplementary Fig. 86.1%). Patients with CPCs harboring a single copy of mutant S5A and S5B). TP53 status had a significant effect on the OS of TP53 exhibited an estimated OS of 66.7% (28.2%–87.8%) and our CPC cohort (log-rank test, P ¼ 3.8 10 3); however, EFS EFS of 44.4% (13.6%–71.9%), whereas patients with CPCs har- was not significantly different between TP53 mutant and wild- boring two copies of mutant TP53 showed an OS of 14.3% type CPCs (P ¼ 0.07; Fig. 5C and D). (0.71%–46.5%) and EFS of 0% (Fig. 6). Investigating survival differences according to the number of mutant copies of TP53 in CPCs revealed a significant reduction in OS (log-rank test 2 copies vs. 1 copy, P ¼ 0.04; 2 copies vs. 0 Discussion copies, P < 1.0 10 4), and EFS (log-rank test 2 copies vs. 1 Our study is the first and largest comprehensive investigation of copy P ¼ 0.03; 2 copies vs. 0 copies P ¼ 0.003) in patients the molecular alterations found in CPTs, demonstrating that the harboring a greater number of mutant TP53 copies. The estimated molecular profile of CPCs is significantly distinct from that of

www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 189

Merino et al.

Figure 5. Unsupervised hierarchical clustering of (A) gene expression normalized intensities and (B) methylation beta values using the PVCLUST algorithm demonstrates signiﬁcant segregation among CPCs, where subgroups are characterized by differences in TP53 status (mutant ¼ black, wild-type ¼ white). Red rectangles delineate statistically signiﬁcant different groups (P ¼ 0.05). PVCLUST algorithm was conducted using 1,000 probe sets with the largest median absolute deviation. , sample with an uncharacterized intronic alteration; #, samples with mutation in SH3-like/ Proline-rich TP53 domain. Colors: Diagnosis: red, CPC; yellow, CPP; light blue, aCPP; TP53 status: black, TP53 mutant; white, TP53 wild-type; Ploidy: green, hyperdiploid; blue, hypodiploid; tan: Diploid; gray, unknown. Kaplan–Meier curves depicting (C) OS and (D) EFS estimates of patients with CPC by TP53 mutation status. Statistical values were obtained with the log-rank (Mantel–Cox) test.

CPPs and aCPPs, and that the papillomas are not significantly papilloma cells. Analyzing tumor heterogeneity in CPTs will be distinct from each other. In addition, using an innovative allele- necessary to identify benign tumors more likely to recur with an specific approach in combination with TP53 sequencing, we aggressive phenotype. identified a particularly poor prognostic subgroup in patients Wide variability in clinical outcome has been observed with TP53 mutant CPC exhibiting aUPD in chromosome 17, and among patients with CPC despite the use of similar treatment who as a result had an elevated number of mutated copies of p53. protocols (2, 16, 17). Our findings demonstrate that the This study provides evidence for the crucial role of molecular molecular heterogeneity of CPCs may be driving this clinical stratification as a tool to improve the clinical management of variability. patients with CPT. Extensive chromosomal alterations were recurrent in CPCs. Atypical CPPs are currently distinguished from CPPs by An allele-specific copy-number approach allowed us to identify histopathology, where aCPPs exhibit increased mitotic activity aneuploidy in 89% of CPCs, and distinguish between hypo- (14); yet, survival outcomes for both CPPs and aCPPs are diploid CPCs, exhibiting numerous chromosomal losses, and comparably favorable. Standard of care for these tumors hyperdiploid CPCs, exhibiting numerous chromosomal gains consists of surgical resection with very few aCPP cases requir- and concurrent aUPD. Our findings uncovered that chromo- ing adjuvant chemotherapy. In our cohort, all patients with somal instability is a common mechanism involved in CPC aCPP for which we had clinical history (6/11) were treated development; however, further examination of the molecular with surgical resection alone, yet demonstrated favorable differences driving hypo- and hyperdiploid development will survival comparable with CPPs. We suggest that the benign identify distinct mechanisms responsible for tumor progres- phenotype of aCPPs may reflect the molecular characteristics sion. Ploidy was not significantly associated with age at diag- it shares with CPPs, including very few chromosome-wide nosis or patient survival; nonetheless, we identified that hyper- losses, and similar gene enrichment patterns and methylation diploid CPCs were significantly enriched in chromosomes signatures. These data lend support for the conservative man- exhibiting aUPD. A recent study reported similar subgroupings agement of patients with aCPP with surgical resection fol- in CPCs, where a higher frequency of chromosomal losses lowed by observation. were observed in younger children and chromosomal gains in Our findings also revealed that copy-number, gene expres- older children, and loss of 12q was associated with shorter sion, and methylation profiles were significantly distinct survival (18). In our cohort, we found no significant correla- between the papillomas and CPCs, indicating the unlikelihood tions between patient age and CPC subgroups or TP53 thatCPPsoraCPPsprogressuntoCPCsbytheacquisitionofa status. Moreover, survival differences were identified only few additional aberrations. Although a few studies have when TP53 copy number and mutation status were examined reported on progression from papillomas to CPCs (9, 15), we concomitantly. believe this unlikely scenario may have been the result of a Arising from somatic recombination errors during mitosis, heterogeneous tumor sample harboring coexisting CPC and aUPD is an important mechanism leading to LOH with an

190 Clin Cancer Res; 21(1) January 1, 2015 Clinical Cancer Research

Genomic Proﬁling of Choroid Plexus Tumors

Figure 6. Kaplan–Meier curves depicting (A) OS and (B) EFS estimates of patients with CPC by number of mutated copies of p53, as estimated by Sanger sequencing and allele- specific copy-number analysis. Statistical values were obtained with the log-rank (Mantel–Cox) test. unaffected copy number, and is therefore associated with the aggressiveness and may elucidate the different clinical outcomes enrichment of chromosomes or regions harboring preexisting observed. Alterations affecting the p53 pathway, either upstream mutations, specific promoter methylation patterns, and focal or downstream of p53, may generate the molecular background deletion of genes (13). In our study, we observed that aUPD necessary for CPC development, and should be investigated affects entire chromosomes in all tumor subgroups; however, further. aUPD was most frequent in CPCs harboring TP53 mutations. Recurrent lesions, such as the chromosome-wide gains of We identified chromosome 17 to be the most frequent site chromosome 1, which was not only recurrently gained but affected by aUPD, and that 90% of CPCs exhibiting aUPD of also the least frequently lost in CPCs, chromosome 12, and chromosome 17 harbored a mutation in TP53, increasing the chromosome-wide loss of chromosome 3, may also be con- number of mutant p53 copies to 2 in these tumors. Our tributing to CPC's unique genotype and would need to be findings suggest aUPD is a mechanism by which CPCs accu- further investigated to identify unique targets for effective mulate deleterious aberrations, such as TP53 mutations, while therapies. retaining the normal function of other genes due to an unaf- Our study demonstrates that investigating the molecular char- fected chromosome copy number. acteristics of CPTs is crucial to further refine the molecular Focusing on the known association between p53 mutations stratification of patients to improve patient care. We suggest that and CPCs, we identified that the number of mutated copies of the prognostic significance of TP53 mutation and copy-number p53 was significantly associated with patient survival. Our status in CPCs be validated prospectively in future cooperative findings support the concept thatinadditiontothelossof clinical trials. Validation of these data in future prospective studies tumor-suppressive activity of p53, mutant TP53 also acquires will inform risk stratification of patients with CPC, and set the oncogenic activities that promote CPC development. The gain framework for future treatment intensification for high-risk of function (GOF) properties of mutant p53 include cellular patients. invasion, proliferation, genomic instability, and polyploidy, among others (reviewed in ref. 19). Because of increased GOF Disclosure of Potential Conflicts of Interest activity, in addition to a complete loss of the tumor suppressor No potential conflicts of interest were disclosed. functions of p53, an elevated number of mutant p53 copies could result in an aggressive phenotype associated with Authors' Contributions decreased survival as we observed in our high-risk CPC patient Conception and design: D.M. Merino, A. Shlien, M.D. Taylor, U. Tabori, cohort. Because we did not assess the mutation status of other D.W. Ellison, R.J. Gilbertson, D. Malkin cancer genes in chromosome 17, we cannot infer that the Development of methodology: D.M. Merino, A. Shlien, M. Pienkowska, number of mutant copies of p53 is the only aberration on this C. Hawkins, M.D. Taylor, U. Tabori, D.W. Ellison, R.J. Gilbertson, D. Malkin Acquisition of data (provided animals, acquired and managed patients, locus driving CPC development and tumor aggressiveness in provided facilities, etc.): D.M. Merino, A. Shlien, A. Villani, A. Novokmet, fi the high-risk CPC patient cohort. However, the signi cant dose- R. Dvir, B.T. Harris, E.I. Hwang, R. Lulla, M.I. Achatz, N. Jabado, J.L. Finlay, dependent correlation observed with OS and EFS in patients E. Bouffet, C. Hawkins, M.D. Taylor, U. Tabori, D.W. Ellison, R.J. Gilbertson, with CPC indicates a role for p53 GOF activity in CPC D. Malkin aggressiveness. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, TP53 computational analysis): D.M. Merino, A. Shlien, A. Villani, M. Pienkowska, mutations alone do not drive chromosomal instability in fi TP53 S. Mack, V. Ramaswamy, D. Shih, R. Tatevossian, S. Choufani, S.M. P ster, CPCs as wild-type tumors also exhibit high levels of chro- N. Jabado, J.L. Finlay, R. Weksberg, E. Bouffet, M.D. Taylor, D.W. Ellison, R.J. mosome-wide gains and losses. However, we have demonstrated Gilbertson, D. Malkin that TP53 mutations are associated with changes in gene expres- Writing, review, and/or revision of the manuscript: D.M. Merino, A. Shlien, sion and methylation patterns that may result in increased tumor A. Villani, M. Pienkowska, S. Mack, V. Ramaswamy, D. Shih, R. Tatevossian,

www.aacrjournals.org Clin Cancer Res; 21(1) January 1, 2015 191

Merino et al.

A. Novokmet, M. Ben-Arush, B.T. Harris, E.I. Hwang, S.M. Pﬁster, M.I. Achatz, The costs of publication of this article were defrayed in part by the N. Jabado, J.L. Finlay, R. Weksberg, C. Hawkins, M.D. Taylor, U. Tabori, payment of page charges. This article must therefore be hereby marked D.W. Ellison, R.J. Gilbertson, D. Malkin advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate Administrative, technical, or material support (i.e., reporting or organizing this fact. data, constructing databases): D.M. Merino, A. Shlien, A. Villani, M.I. Achatz, D. Malkin Received May 22, 2014; revised September 3, 2014; accepted September 26, Study supervision: U. Tabori, D. Malkin 2014; published OnlineFirst October 21, 2014.

References 1. Ogiwara H, Dipatri AJ Jr, Alden TD, Bowman RM, Tomita T. 10. Van Loo P, Nordgard SH, Lingjñrde OC, Russnes HG, Rye IH, Sun W, et al. Choroid plexus tumors in pediatric patients. Br J Neurosurg 2012; Allele-specific copy number analysis of tumors. Proc Natl Acad Sci U S A 26:32–7. 2010;107:16910–5. 2. Lafay-Cousin L, Keene D, Carret A-S, Fryer C, Brossard J, Crooks B, et al. 11. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette M a, Choroid plexus tumors in children less than 36 months: the Canadian et al. Gene set enrichment analysis: a knowledge-based approach for Pediatric Brain Tumor Consortium (CPBTC) experience. Child's Nerv Syst interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2011;27:259–64. 2005;102:15545–50. 3. Addo NK, Kamaly-Asl ID, Josan VA, Kelsey AM, Estlin EJ. Preop- 12. Witt H, Mack SC, Ryzhova M, Bender S, Sill M, Isserlin R, et al. Delineation erative vincristine for an inoperable choroid plexus papilloma: a of two clinically and molecularly distinct subgroups of posterior fossa case discussion and review of the literature. Case Rep 2011;8: ependymoma. Cancer Cell 2011;20:143–57. 149–53. 13. Tuna M, Knuutila S, Mills GB. Uniparental disomy in cancer. Trends Mol 4. Rickert CH, Paulus W. Tumors of the choroid plexus. Microsc Res Tech Med 2009;15:120–8. 2001;52:104–11. 14. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A, et al. 5. Tabori U, Shlien A, Baskin B, Levitt S, Ray P, Alon N, et al. TP53 alterations The 2007 WHO classification of tumours of the central nervous system. determine clinical subgroups and survival of patients with choroid plexus Acta Neuropathol 2007;114:97–109. tumors. J Clin Oncol 2010;28:1995–2001. 15. Dhillon RS, Wang YY, McKelvie PA, O'Brien B. Progression of choroid 6. Lafay-Cousin L, Mabbott DJ, Halliday W, Taylor MD, Tabori U, plexus papilloma. J Clin Neurosci 2013;20:1775–8. Kamaly-Asl ID, et al. Use of ifosfamide, carboplatin, and etoposide 16. Gozali AE, Britt B, Shane L, Gonzalez I, Gilles F, McComb JG, et al. Choroid chemotherapy in choroid plexus carcinoma. J Neurosurg Pediatr plexus tumors; management, outcome, and association with the Li-Frau- 2010;5:615–21. meni syndrome: The Children's Hospital Los Angeles (CHLA) experience, 7. Jeibmann A, Hasselblatt M, Gerss J, Wrede B, Egensperger R, Beschorner 1991-2010. Pediatr Blood Cancer 2012;58:905–9. R, et al. Prognostic implications of atypical histologic features 17. Bettegowda C, Adogwa O, Mehta V, Chaichana KL, Weingart J, Carson BS, in choroid plexus papilloma. J Neuropathol Exp Neurol 2006;65: et al. Treatment of choroid plexus tumors: a 20-year single institutional 1069–73. experience. J Neurosurg Pediatr 2012;10:398–405. 8. Wrede B, Hasselblatt M, Peters O, Thall PF, Kutluk T, Moghrabi A, et al. 18. Ruland V, Hartung S, Kordes U, Wolff JE, Paulus W, Hasselblatt M. Choroid Atypical choroid plexus papilloma: clinical experience in the CPT-SIOP- plexus carcinomas are characterized by complex chromosomal alterations 2000 study. J Neurooncol 2009;95:383–92. related to patient age and prognosis. Genes Chromosomes Cancer 9. Jeibmann A, Wrede B, Peters O, Wolff JE, Paulus W, Hasselblatt M. 2014;53:373–80. Malignant progression in choroid plexus papillomas. J Neurosurg 2007; 19. Muller PAJ, Vousden KH. Mutant p53 in cancer: new functions and 107:199–202. therapeutic opportunities. Cancer Cell 2014;25:304–17.

192 Clin Cancer Res; 21(1) January 1, 2015 Clinical Cancer Research

Molecular Characterization of Choroid Plexus Tumors Reveals Novel Clinically Relevant Subgroups

Diana M. Merino, Adam Shlien, Anita Villani, et al.

Clin Cancer Res 2015;21:184-192. Published OnlineFirst October 21, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-14-1324

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2014/10/22/1078-0432.CCR-14-1324.DC1

Cited articles This article cites 19 articles, 3 of which you can access for free at: http://clincancerres.aacrjournals.org/content/21/1/184.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/21/1/184.full#related-urls

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

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

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/21/1/184. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.