European Review for Medical and Pharmacological Sciences 2017; 21 (4 Suppl): 13-19 Recent perspectives of (childhood brain tumors)

Z.-D. ZENG, F.-L. LIU, S.-X. LI

Department of Pediatric Surgery, Xuzhou Children’s Hospital, Xuzhou, Jiangsu, P.R. China

Abstract. – Ependymomas are childhood are commonly positive for glial fibrillary acidic brain tumors that occur throughout the cen- protein (GFAP), neural cell adhesion molecule tral , but are most common in protein (NCAM), and epithelial membrane anti- the hindbrain, also known as the posterior fos- gen protein (EMA), which allows for further de- sa (PF). Current standard therapy comprises lineation from other histologically similar brain maximal safe surgery, and there is no scope for 5 further increase in survival. Despite the histo- tumors . The WHO recognizes three grades for logical similarity, ependymomas from through- ependymomas: out the neuroaxis likely comprise multiple in- dependent entities, each with a distinct molec- Grade I – comprising and ular pathogenesis. The present review article myxopapillary ependymomas, which are easi- would discuss both and epigenetics of ly recognizable histological entities associated ependymomas. with a better survival and increased age; Key Words: Grade II – pertains to ependymomas, which lack Ependymomas, Epigenetics, Genetics, Pediatrics. Grade III features described below; Grade III – also described as anaplastic, this classification corresponds to ependymomas with “increased cellularity and brisk mitotic activity. Introduction While the Grade I criteria for Ependymomas are rare, chemo-resistant, cen- are relatively clear, the parameters used to distin- tral nervous system tumors, arising in both chil- guish Grade II and III ependymomas are highly dren and adults1. Ependymomas could arise along debated as prognostic differences have been ob- the entire neuro-axis occurring in the supraten- served in some tumor cohorts, but not others5. torial (ST) brain comprising the cerebral hemi- This discrepancy occurs even after controlling spheres, the posterior fossa (PF) comprising the potentially confounding factors such as age or cerebellum and brain stem, and along the entire tumor location. The lack of reproducibility and (SP)2. In children, 90% of ependymo- reliability of histopathological is demon- mas arise intracranially, with nearly two-thirds strated in a systemic review of three independent occurring in the PF3. The occurrence of relapse European trial cohorts by five leading neuropa- is significantly greater in the pediatric popula- thologists, in which a consensus agreement on tion, and is clinically variable with recurrences the classification of Grade II or III ependymoma observed in some cases 10-15 years following was reached in less than half of 221 total cases treatment of the primary tumor4. examined6. While the prognostic utility of his- topathological grading is still debated, immu- Histopathology nohistochemistry (IHC) and molecular markers The hallmark histological features of ependy- have been proposed as potential solutions e.g. moma include: (1) perivascular pseudorosettes, (TERT) protein expression7, V-erb-b2 composed of ependymal cell processes radially erythroblastic leukemia viral homolog arranged towards blood vessels; (2) true ependy- 2/4 (ERBB2/4) protein expression, or fluores- mal rosettes, consisting of tumor cells arranged cence in situ hybridization (FISH) of recurrent radially surrounding an empty lumen. Regarding chromosomal alterations8,9. These IHC and mo- immunohistochemical staining, ependymomas lecular markers remain to be validated in inde-

Corresponding Author: Fengli Liu, MD; e-mail: [email protected] 13 Z.-D. Zeng, F.-L. Liu, S.-X. Li pendent and prospective ependymoma cohorts, a The Genetic Basis of Ependymoma challenge necessitating multi-center collaborative Efforts to identify driver and tu- efforts. mor suppressor (TSG) of ependymoma began largely with characterization of these Treatment Strategies tumors at a DNA copy number level using cy- To this date, treatment for ependymoma re- togenetic approaches, DNA-based microarrays, mains aggressive surgical intervention and ad- whole-genome and whole exome sequencing15. juvant radiotherapy (10). In spite of the histolog- Despite higher resolution array technologies, the ical challenges in identifying high- vs. low-risk vast majority of recurrent somatic copy number ependymoma patients, the extent of surgical re- alterations (SCNA) are broad and involve losses section is the most frequently and consistently of : 1p, 3, 6q, 9p, 10q, 13q, 16p, 17, reported prognostic indicator of ependymoma 21 and 22q, and gains of chromosome: 1q, 4q, patient survival. In other pediatric brain tumors, 5, 7, 8, 9, 12q, and 2016 (16). The most frequent such as , radiotherapy is typi- and focal SCNA in ependymoma is a homo- cally avoided in children less than 3 years of age, zygous deletion encompassing the CDKN2A/ due to increased risk of long-term neurological Ink4a locus, which is restricted to supratento- and neuroendocrine sequelae. However, the ag- rial ependymomas (ST) (17). In case of poste- gressive nature of ependymoma in infants and rior fossa (PF) ependymomas, recurrent and young children combined with a lack of effective focal copy number alterations pinpointing driver has provided a rationale for the genes have yet to be discovered, highlighting use of conformal and intensity- modulated radio- the difficulty in understanding the biological therapy in infants. Early evidence in prospective basis, and identifying novel therapeutic targets clinical trials has demonstrated that conformal for this anatomical subtype of ependymoma17. radiation is both effective and associated with loss has been shown to be the minimal short-term neurological side effects in most frequent gross copy number alteration in the 5-years following treatment11. Whether these ependymoma with a frequency ranging from outcomes are maintained for long-term is in a 26% to 71%18,19 Further, chromosome 22q loss queue for further evaluation in the near future. has been observed preferentially in spinal vs. Despite improvements in surgical techniques and intracranial ependymoma, and in adult vs. pe- advances in conformal radiation, recurrence rates diatric cases19. The Neurofibromatosis II (NF2) remain high in ependymoma patients, particular- is thought to be the candidate TSG of this ly in the pediatric population12. Although chemo- region, as patients with Neurofibromatosis type therapy has been used extensively in the treat- II disorder develop a variety of central nervous ment of children with intracranial ependymoma, system tumors including ependymoma, schwan- clinical trial response rates to numerous single noma, and meningioma. However, NF2 is mu- agents are less than 12%, with less than 5% of pa- tated exclusively in spinal ependymomas, thus tients experiencing complete responses13. Results suggesting alternate mechanisms of down-regu- from several multi-center ependymoma clinical lation, or another putative chromosome 22q TSG trials suggested that there is a little evidence that in the case of intracranial ependymoma. Anoth- is effective in treatment for this er broad chromosomal abnormality frequently tumor type14. As result, the current standard of observed in ependymoma is monosomy 17, with care for patients with recurrent ependymoma complete or partial loss of both chromosomal p is maximal-safe surgical resection followed by and q arms18,19. Chromosome 1q gain has also re-irradiation. Despite prolonged overall survival been consistently reported as a frequent genom- observed from re-irradiation of recurrent ependy- ic alteration occurring in nearly 22% of cases of moma, the risks of secondary tumors and neu- intracranial ependymoma. Further, an increased rological impairments have yet to be adequately incidence of 1q gain has been observed in pos- assessed. The high recurrence rate of pediatric terior fossa ependymoma and associated with ependymomas, lack of prognostic histological poor clinical outcome20,21. It is thought that the and molecular markers and dearth of chemother- chromosome 1q25 locus harbors bonafide onco- apeutic avenues underscore the importance of gene involved in the initiation, maintenance, or understanding the biological basis of ependymo- progression of ependymoma. Efforts have been ma such that rationale molecular targets could be made to correlate this region of chromosomal identified and rapidly translated into the clinic2. gain with and have identified

14 Recent perspectives of ependymomas (childhood brain tumors)

CHI3L1 and a family of S100 genes as up reg- animal models of this disease26. As novel ependy- ulated and potential driver oncogenes; however, moma targets are discovered, validation and pri- these candidates remain to be functionally val- oritization of candidates would require accurate idated20,21. pre-clinical models of ependymoma. Atkinson et al27 demonstrate the utility and promise of this Inter-Tumoral Heterogeneity and approach in ST ependymoma cultures, generated Putative Cells of Origin by EPHB2 over expression and CDKN2A/Ink4a Using gene expression profiling and unsuper- deletion in forebrain radial glia. vised clustering, ependymomas have been di- vided into three principal molecular subgroups, The Epigenetic Basis of Ependymoma which are separated largely according to an- Aberrant promoter methylation of CpG di-nu- atomical location: (1) Supratentorial (ST); (2) cleotides is a well-recognized feature seen in Posterior Fossa (PF); 3) Spinal cord (SP). These numerous solid and liquid cancers28. TSG is re- three subgroups have been further divided into ported in up to 100% of ependymomas, and oc- molecularly and biologically distinct subtypes curring in all clinical and pathological subtypes29. of ST and PF ependymoma as defined by dis- HIC1 is also commonly methylated, in up to tinct clinical features22,23. The genes distinguish- 83% of ependymomas, with a higher incidence ing supratentorial, posterior fossa, and spinal of intracranial tumors30. Furthermore, the CD- ependymomas involve mainly families of genes KN2A/INK4a locus, which is focally and recur- regulating neural precursor cell proliferation and rently deleted in supratentorial ependymoma18, differentiation. Supratentorial ependymomas has been shown to be hypermethylated in 21% have elevated EPHB-EPHRIN, NOTCH and cell of cases, followed by CDKN2B and cycle related genes, while posterior fossa ependy- in 32% and 33% of tumors, respectively31. To a momas express many inhibitors of differentiation lesser extent, putative TSGs found to be hyper- (ID1/2/4), and the aquaporin family of genes methylated in ependymoma include ZMYD10, (AQP1/3/4). Spinal ependymomas, however, are GSTP1, DAPK, FHIT, MGMT, DNAJC15, characterized by the expression of various ho- RARB, TIMP3, THBS1, TP73, the Tumour Ne- meobox genes including HOXA7/9, HOXB6/7, crosis Factor Related Apoptosis-Inducing Ligand and HOXC6/1024. While these subgroup gene sig- (TRAIL) gene family, CASP8, TNFRSF10C and natures revealed distinct tumorigenic pathways, TNFRSF10D32,33. Despite the frequency of DNA Taylor et al20 proposed potential signatures of methylation of these potential TSGs in ependy- anatomically distinct cells of origin giving rise to moma, their role in tumor development remains different subgroups of ependymoma. They fur- unclear and requires further investigation in ap- ther suggested that ependymoma might originate propriate ependymoma models. from radial glial cells (RGCs), a population of primitive neural and multi-potent precursor cells Aberrant DNA Methylation important for neurogenesis. Further evidence im- In the last 5 years, the expansion of microarray plicating RGCs as cells of origin of ependymoma and next generation sequencing technologies, has was demonstrated by Johnson et al21, in which allowed for genome-wide investigations of DNA ex-vivo over-expression of EPHB2 in /INK4A methylation and histone modifications at unprec- deficient forebrain RGCs led to the formation of edented resolution and throughput. Using the the first mouse model of ST ependymoma. Illumina Golden Gate Methylation Panel 1 (1505 CpG sites), Rogers et al29 profiled a series Cell Line and Animal Models of 73 primary and 25 recurrent ependymomas. of Ependymoma Here they reported that the DNA methylation In the past, lack of clear driver alterations in profiles of ependymoma are distinguished largely ependymoma has hampered the ability to generate according to their location in the central nervous animal models of this disease. Laboratories have system, supporting the notion that ependymomas thus relied upon patient-derived ependymoma arising from different anatomic compartments cultures grown in vitro and orthotopic xenograft are molecularly distinct20,22. Furthermore, they models generated with limited success particular- demonstrated that ST and SP ependymomas, to- ly in the case of PF ependymoma25. Identification gether, exhibited a larger number of hypermethyl- of the putative cell-of-origin of ependymoma and ated and down-regulated genes in comparison to drivers of ST ependymoma, have led to the first PF tumors. These changes in DNA methylation

15 Z.-D. Zeng, F.-L. Liu, S.-X. Li were shown to be associated with alterations in went apoptosis in response to treatment with gene expression of de novo and maintenance the HDACi, Trichostatin-A. Given the rapid of DNA methyltransferases DNMT1, DNMT3A development of novel pharmacologic inhibitors and DNMT3B. Interestingly, genes involved in of epigenetic marks, it raises the question as immune cell response (NOD2, IRF7, IRAK3, to whether these, or at least some, epigenetic OSM and PI3), cell growth and death (MAPK10, modifications are central to ependymoma patho- and TP73), and the c-Jun N-terminal kinases genesis and whether they might represent novel (JNK) pathway were found to be hypermethyl- avenues for therapeutic inhibition39. ated. Understanding the contribution of epigen- etic alterations in these pathways might reveal Future Steps mechanisms of ependymoma tumorigenesis, and Although genomic and transcriptomic pro- potential actionable targets for therapeutic inter- filing efforts have identified distinct molecular vention. In contrast to hypermethylation of CpG subtypes of ependymoma revealing potential island promoters in cancer, global hypometh- drivers of the disease, the vast majority of ylation is a trend observed in numerous tumor ependymal tumors is characterized by either types and is associated with cancer progression. large chromosomal alterations, hampering the A global decrease in methylation has been ob- identification of driver events, or are charac- served predominantly at repetitive elements such terized by very few genomic abnormalities oc- as LINEs, SINEs and LTRs, which are important curring in the youngest patient population40. It for maintaining genomic stability34. To elucidate remains to be seen whether recurrent somatic the contribution of DNA methylation alterations single nucleotide variants (SNVs) or structural at repetitive sequences in ependymoma, Xie et rearrangements (i.e. fusion transcripts) might al35 developed a novel genome-wide approach to contribute to the pathogenesis of ependymoma, generate methylation profiles for thousands of as reported in several other adults and pediatric Alu elements (the most abundant class of repet- CNS neoplasms40-42. DNA methylation profiling itive elements) and their flanking sequences35. efforts have also been important in the molecu- Here they demonstrated that while the majority lar stratification of CNS tumors43. They have al- of Alu elements and flanking sequences remain so shown promise in distinguishing the principle unaltered in ependymoma genesis, a small sub- molecular subgroups of ependymoma as well as set of Alu flanking sequences, with low CpG the identification of pathways targeted by DNA density, exhibited variable methylation patterns. hypermethylation. As a future step, expanding These sequences tended to be hypermethylated in DNA methylation profiling to platforms with ependymoma at regions proximal to CpG islands higher CpG coverage could reveal novel targets, and hypomethylated in intergenic regions. Impor- pathways, and mechanisms of epigenetic alter- tantly, several of these patterns were shown to be ation. Also, given the contributions of aberrant associated with aggressive primary ependymo- methylation near repeat elements in the ependy- mas and tumor relapse. However, the impact that moma epigenome, more global investigations these epigenetic alterations on genomic stability beyond gene promoters might be needed, and and their respective pathways are remains to be could be readily examined with whole-genome elucidated. bisulfite sequencing.

Potential Applications of Epigenetic Modifiers for Ependymoma Treatment Conclusions Characterizing the epigenome of ependymo- ma might hold therapeutic promise, as these The contribution of genetic and epigenetic marks such as CpG DNA methylation and his- changes in ependymoma pathogenesis might tone modification are generally reversible by not only improve our understanding of the bi- pharmacologic inhibition. Importantly, inhibi- ology of this disease, but also reveal actionable tors of DNA methylation (decitabine) and his- pathways that could be rapidly translated to the tone deacetylation (Vorinostat) are FDA ap- clinic. proved and have shown efficacy in hematolog- ical malignancies36,37. These findings were also supported by Rahman et al38, demonstrating Conflict of Interest that the ependymoma cell line nEPN2 under- The Authors declare that they have no conflict of interests.

16 Recent perspectives of ependymomas (childhood brain tumors)

References conformal for pediatric ependy- moma. Int J Radiat Oncol Biol Phys 2014; 88: 1) Frappaz D, Vasiljevic A, Beuriat PA, Alapetite C, Grill 814-821. J, Szathmari A, Faure-Conter C. Pediatric ependy- 12) Bouffet E, Hawkins CE, Ballourah W, Taylor MD, momas: current diagnosis and therapy. Bull Can- Bartels UK, Schoenhoff N, Tsangaris E, Huang A, cer 2016; 103: 869-879. Kulkarni A, Mabbot DJ, Laperriere N, Tabori U. Sur- 2) Chiang JC, Ellison DW. Molecular of vival benefit for pediatric patients with recurrent paediatric tumours. J ependymoma treated with reirradiation. Int J Ra- Pathol 2017; 241: 159-172. diat Oncol Biol Phys 2012; 83: 1541-1548. 3) Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, 13) Bouffet E, Foreman, N. Chemotherapy for intracra- Burger PC, Jouvet A, Scheithauer BW, Kleihues P. nial ependymomas. Childs Nerv Syst 1999; 15: The 2007 WHO classification of tumours of the 563-570. central nervous system. Acta Neuropathol 2007; 14) Venkatramani R, Ji L, Lasky J, Haley K, Judkins A, 114: 97-109. Zhou S, Sposto R, Olshefski R, Garvin J, Tekautz T, 4) Schindler M, Belle FN, Grotzer MA, von der Weid Kennedy G, Rassekh SR, Moore T, Gardner S, Allen NX, Kuehni CE, Swiss Paediatric Oncology Group J, Shore R, Moertel C, Atlas M, Dhall G, Finlay J. (SPOG) . Childhood cancer survival in Switzer- Outcome of infants and young children with new- land (1976-2013): time-trends and predictors. Int ly diagnosed ependymoma treated on the “Head J Cancer 2017; 140: 62-74. Start” III prospective clinical trial. J Neurooncol 5) Godfraind, C. Classification and controversies in 2013; 113: 285-291. pathology of ependymomas. Childs Nerv Syst 15) Pérez-Ramírez M, Hernández-Jiménez AJ, Guerre- 2009; 25: 1185-1193. ro-Guerrero A, Benadón-Darszon E, Pérezpeña-Díaz- 6) Ellison DW, Kocak M, Figarella-Branger D, Felice conti M, Siordia-Reyes AG, García-Méndez A, de León G, Catherine G, Pietsch T, Frappaz D, Massimino M, FC, Salamanca-Gómez FA, García-Hernández N. Ge- Grill J, Boyett JM, Grundy RG. Histopathological nomics and epigenetics: a study of ependymo- grading of pediatric ependymoma: reproducibility mas in pediatric patients. Clin Neurol Neurosurg and clinical relevance in European trial cohorts. J 2016; 144: 53-58. Negat Results Biomed 2011; 10: 7. 16) Thompson YY, Ramaswamy V, Diamandis P, Daniels C, 7) Castelo-Branco P, Choufani S, Mack S, Gallagher D, Taylor MD. Posterior fossa ependymoma: current Zhang C, Lipman T, Zhukova N, Walker EJ, Martin insights. Childs Nerv Syst 2015; 31: 1699-1706. D, Merino D, Wasserman JD, Elizabeth C, Alon N, 17) Mack SC, Witt H, Piro RM, Gu L, Zuyderduyn S, Stütz Zhang L, Hovestadt V, Kool M, Jones DT, Zadeh G, AM, Wang X, Gallo M, Garzia L, Zayne K, Zhang X, Croul S, Hawkins C, Hitzler J, Wang JC, Baruchel S, Ramaswamy V, Jäger N, Jones DT, Sill M, Pugh TJ, Ry- Dirks PB, Malkin D, Pfister S, Taylor MD, Weksberg zhova M, Wani KM, Shih DJ, Head R, Remke M, Bai- R, Tabori U. Methylation of the TERT promoter and ley SD, Zichner T, Faria CC, Barszczyk M, Stark S, Se- risk stratification of childhood brain tumours: an ker-Cin H, Hutter S, Johann P, Bender S, Hovestadt integrative genomic and molecular study. Lancet V, Tzaridis T, Dubuc AM, Northcott PA, Peacock J, Oncol 2013; 14: 534-542. Bertrand KC, Agnihotri S, Cavalli FM, Clarke I, Neth- 8) Gilbertson RJ, Bentley L, Hernan R, Junttila TT, ery-Brokx K, Creasy CL, Verma SK, Koster J, Wu X, Frank AJ, Haapasalo H, Connelly M, Wetmore C, Yao Y, Milde T, Sin-Chan P, Zuccaro J, Lau L, Perei- Curran T, Elenius K, Ellison DW. ERBB ra S, Castelo-Branco P, Hirst M, Marra MA, Roberts signaling promotes ependymoma cell prolifera- SS, Fults D, Massimi L, Cho YJ, Van Meter T, Graj- tion and represents a potential novel therapeutic kowska W, Lach B, Kulozik AE, von Deimling A, Witt target for this disease. Clin Cancer Res 2002; 8: O, Scherer SW, Fan X, Muraszko KM, Kool M, Pome- 3054-3064. roy SL, Gupta N, Phillips J, Huang A, Tabori U, Haw- 9) Korshunov A, Remke M, Werft W, Benner A, Ryzhova kins C, Malkin D, Kongkham PN, Weiss WA, Jabado M, Witt H, Sturm D, Wittmann A, Schöttler A, Fels- N, Rutka JT, Bouffet E, Korbel JO, Lupien M, Aldape berg J, Reifenberger G, Rutkowski S, Scheurlen W, Ku- KD, Bader GD, Eils R, Lichter P, Dirks PB, Pfister SM, lozik AE, von Deimling A, Lichter P, Pfister SM. Adult Korshunov A, Taylor MD. Epigenomic alterations and pediatric are genetically define lethal CIMP-positive ependymomas of in- distinct and require different algorithms for mo- fancy. Nature 2014; 50: 445-450. lecular risk stratification. J Clinical Oncol 2010; 18) Tamiolakis D, Papadopoulos N, Venizelos I, Lam- 28: 3054-3060. bropoulou M, Nikolaidou S, Bolioti S, Kiziridou A, 10) Farr JB, Sabin ND, Hua CH, Pirlepesov F, Shin J, Manavis J, Alexiadis G, Simopoulos C. Loss of chro- Moskvin V, Indelicato DJ, Li Z, Gajjar A, Merchant mosome 1 in myxopapillary ependymoma sug- TE. Dose and dose-weighted linear energy trans- gests a region out of chromosome 22 as critical fer brainstem volumetric thresholds associated for tumour biology: a FISH analysis of four cases with mr signal changes in pediatric patients with on touch imprint smears. Cytopathology 2006; 17: ependymoma treated with proton therapy. Int J 199-204. Radiat Oncol Biol Phys 2016; 96: S229-S230. 19) Mack SC, Taylor MD. The genetic and epigenetic 11) Willard VW, Conklin HM, Boop FA, Wu S, Merchant basis of ependymoma. Child’s nervous system. TE. Emotional and behavioral functioning after Childs Nerv Syst 2009, 25: 1195-1201.

17 Z.-D. Zeng, F.-L. Liu, S.-X. Li

20) Taylor MD, Poppleton H, Fuller C, Su X, Liu Y, Jen- ependymomas display a hypermethylated pheno- sen P, Magdaleno S, Dalton J, Calabrese C, Board J, type, which includes the loss of tumor suppressor Macdonald T, Rutka J, Guha A, Gajjar A, Curran T, genes involved in the control of cell growth and Gilbertson RJ. Radial glia cells are candidate stem death. Acta Neuropathol 2012; 123: 711-725. cells of ependymoma. Cancer Cell 2005; 8: 323- 30) Waha A, Koch A, Hartmann W, Mack H, Schramm J, 335. Sörensen N, Berthold F, Wiestler OD, Pietsch T, Wa- 21) Johnson RA. Cross-species matches ha A. Analysis of HIC-1 methylation and transcrip- driver and cell compartments to model tion in human ependymomas. Int J Cancer 2004; ependymoma. Nature 2010; 466: 632-636. 110: 542-549. 22) Bettegowda C, Agrawal N, Jiao Y, Wang Y, Wood 31) Rousseau E, Ruchoux MM, Scaravilli F, Chapon F, LD, Rodriguez FJ, Hruban RH, Gallia GL, Binder ZA, Vinchon M, De Smet C, Godfraind C, Vikkula M. CD- Riggins CJ, Salmasi V, Riggins GJ, Reitman ZJ, Rasheed KN2A, CDKN2B and p14ARF are frequently and A, Keir S, Shinjo S, Marie S, McLendon R, Jallo G, differentially methylated in ependymal tumours. Vogelstein B, Bigner D, Yan H, Kinzler KW, Papado- Neuropathol Appl Neurobiol 2003; 29: 574- 583. poulos N. Exomic sequencing of four rare central 32) Gonzalez-Gomez P, Bello MJ, Alonso ME, Arjona nervous system tumor types. Oncotarget 2013; 4: D, Lomas J, de Campos JM, Isla A, Rey JA. CpG is- 572-583. land methylation status and analysis of 23) Wani, K, Armstrong TS, Vera-Bolanos E, Raghuna- the RB1 gene essential promoter region and pro- than A, Ellison D, Gilbertson R, Vaillant B, Gold- tein-binding pocket domain in nervous system tu- man S, Packer RJ, Fouladi M, Pollack I, Mikkelsen mours. Br J Cancer 2003; 88: 109-114. T, Prados M, Omuro A, Soffietti R, Ledoux A, Wil- 33) Koos, B, Bender S, Witt H, Mertsch S, Felsberg J, son C, Long L, Gilbert MR, Aldape K; Collaborative Beschorner R, Korshunov A, Riesmeier B, Pfister S, Ependymoma Research Network. A prognostic gene Paulus W, Hasselblatt M. The expression signature in infratentorial ependymo- evi-1 is overexpressed, promotes proliferation, ma. Acta Neuropathol 2012; 123: 727-738. and is prognostically unfavorable in infratentorial 24) Van Vuurden DG, Aronica E, Hulleman E, Wedekind ependymomas. Clin Cancer Res 2011; 17: 3631- LE, Biesmans D, Malekzadeh A, Bugiani M, Geerts 3637. D, Noske DP, Vandertop WP, Kaspers GJ, Cloos J, 34) Cadieux B, Ching TT, VandenBerg SR, Costello JF. Würdinger T, van der Stoop PP. Pre-B-cell leukemia Genome-wide hypomethylation in human glio- homeobox interacting protein 1 is overexpressed blastomas associated with specific copy number in astrocytoma and promotes tumor cell growth alteration, methylenetetrahydrofolate reductase and migration. Neuro Oncol 2014; 16: 946-955. allele status, and increased proliferation. Cancer 25) Milde, T, Kleber S, Korshunov A, Witt H, Hielscher Res 2006; 66: 8469-8476. T, Koch P, Kopp HG, Jugold M, Deubzer HE, Oehme 35) Xie H, Wang M, Bonaldo Mde F, Rajaram V, Stell- I, Lodrini M, Gröne HJ, Benner A, Brüstle O, Gil- pflug W, Smith C, Arndt K, Goldman S, Tomita T, bertson RJ, von Deimling A, Kulozik AE, Pfister Soares MB. Epigenomic analysis of Alu repeats in SM, Martin-Villalba A, Witt O. A novel human human ependymomas. Proc Natl Acad Sci U S A high-risk ependymoma model reveals 2010; 107: 6952-6957. the differentiation-inducing potential of the his- 36) O’Connor OA, Heaney ML, Schwartz L, Richardson tone deacetylase inhibitor Vorinostat. Acta Neu- S, Willim R, MacGregor-Cortelli B, Curly T, Moskow- ropathol 2011; 122: 637-650. itz C, Portlock C, Horwitz S, Zelenetz AD, Frankel 26) Yu L, Baxter PA, Voicu H, Gurusiddappa S, Zhao Y, S, Richon V, Marks P, Kelly WK. Clinical experience Adesina A, Man TK, Shu Q, Zhang YJ, Zhao XM, Su with intravenous and oral formulations of the nov- JM, Perlaky L, Dauser R, Chintagumpala M, Lau CC, el histone deacetylase inhibitor suberoylanilide Blaney SM, Rao PH, Leung HC, Li XN. A clinical- hydroxamic acid in patients with advanced hema- ly relevant orthotopic xenograft model of ependy- tologic malignancies. J Clin Oncol 2006; 24: 166- moma that maintains the genomic signature of 173. the primary tumor and preserves cancer stem 37) Shen L, Kantarjian H, Guo Y, Lin E, Shan J, Huang cells in vivo. Neuro Oncol 2010; 12: 580-594. X, Berry D, Ahmed S, Zhu W, Pierce S, Kondo Y, Oki 27) Atkinson JM, Shelat AA, Carcaboso AM, Kranenburg Y, Jelinek J, Saba H, Estey E, Issa JP. DNA methyla- TA, Arnold LA, Boulos N, Wright K, Johnson RA, tion predicts survival and response to therapy in Poppleton H, Mohankumar KM, Féau C, Phoenix T, patients with myelodysplastic syndromes. J Clin Gibson P, Zhu L, Tong Y, Eden C, Ellison DW, Prie- Oncol 2010; 28: 605-613. be W, Koul D, Yung WK, Gajjar A, Stewart CF, Guy 38) Rahman R, Osteso-Ibanez T, Hirst RA, Levesley J, Kil- RK, Gilbertson RJ . An integrated in vitro and in vivo day JP, Quinn S, Peet A, O’Callaghan C, Coyle B, high-throughput screen identifies treatment leads Grundy RG. Histone deacetylase inhibition atten- for ependymoma. Cancer Cell 2011; 20: 384-399. uates cell growth with associated telomerase in- 28) Hanahan D, Weinberg RA. Hallmarks of cancer: the hibition in high-grade childhood cells. next generation. Cell 2011; 144: 646-674. Mol Cancer Ther 2010; 9: 2568-2581. 29) Rogers HA, Kilday JP, Mayne C, Ward J, Adamo- 39) Witt H, Mack SC, Ryzhova M, Bender S, Sill M, Isser- wicz-Brice M, Schwalbe EC, Clifford SC, Coyle B, lin R, Benner A, Hielscher T, Milde T, Remke M, Jones Grundy RG. Supratentorial and spinal pediatric DT, Northcott PA, Garzia L, Bertrand KC, Wittmann

18 Recent perspectives of ependymomas (childhood brain tumors)

A, Yao Y, Roberts SS, Massimi L, Van Meter T, Weiss Bigner DD, Yan H, Papadopoulos N, Vogelstein B, WA, Gupta N, Grajkowska W, Lach B, Cho YJ, von Kinzler KW, Velculescu VE. The genetic landscape Deimling A, Kulozik AE, Witt O, Bader GD, Hawkins of the childhood cancer medulloblastoma. Sci- CE, Tabori U, Guha A, Rutka JT, Lichter P, Korshunov ence 2011; 331: 435-439. A, Taylor MD, Pfister SM . Delineation of two clin- 42) Pugh TJ, Weeraratne SD, Archer TC, Pomeranz Krum- ically and molecularly distinct subgroups of pos- mel DA, Auclair D, Bochicchio J, Carneiro MO, Car- terior fossa ependymoma. Cancer Cell 2011; 20: ter SL, Cibulskis K, Erlich RL, Greulich H, Lawrence 143-157. MS, Lennon NJ, McKenna A, Meldrim J, Ramos AH, 40) Jäger N, Kool M, Zichner T, Hutter B, Sultan M, Cho Ross MG, Russ C, Shefler E, Sivachenko A, Sogol- YJ, Pugh TJ, Hovestadt V, Stütz AM, Rausch T, War- off B, Stojanov P, Tamayo P, Mesirov JP, Amani V, natz HJ, Ryzhova M, Bender S, Sturm D, Pleier S, Cin Teider N, Sengupta S, Francois JP, Northcott PA, Tay- H, Pfaff E, Sieber L, Wittmann A, Remke M, Witt H, lor MD, Yu F, Crabtree GR, Kautzman AG, Gabriel Hutter S, Tzaridis T, Weischenfeldt J, Raeder B, Avci SB, Getz G, Jäger N, Jones DT, Lichter P, Pfister SM, M, Amstislavskiy V, Zapatka M, Weber UD, Wang Q, Roberts TM, Meyerson M, Pomeroy SL, Cho YJ. Me- Lasitschka B, Bartholomae CC, Schmidt M, von Kalle dulloblastoma exome sequencing uncovers sub- C, Ast V, Lawerenz C, Eils J, Kabbe R, Benes V, van type-specific somatic mutations. Nature 2012; Sluis P, Koster J, Volckmann R, Shih D, Betts MJ, Rus- 488: 106-110. sell RB, Coco S, Tonini GP, Schüller U, Hans V, Graf 43) Sturm D, Witt H, Hovestadt V, Khuong-Quang DA, N, Kim YJ, Monoranu C, Roggendorf W, Unterberg Jones DT, Konermann C, Pfaff E, Tönjes M, Sill M, A, Herold-Mende C, Milde T, Kulozik AE, von Deim- Bender S, Kool M, Zapatka M, Becker N, Zucknick M, ling A, Witt O, Maass E, Rössler J, Ebinger M, Schuh- Hielscher T, Liu XY, Fontebasso AM, Ryzhova M, Al- mann MU, Frühwald MC, Hasselblatt M, Jabado N, brecht S, Jacob K, Wolter M, Ebinger M, Schuhmann Rutkowski S, von Bueren AO, Williamson D, Clifford MU, van Meter T, Frühwald MC, Hauch H, Pekrun SC, McCabe MG, Collins VP, Wolf S, Wiemann S, A, Radlwimmer B, Niehues T, von Komorowski G, Lehrach H, Brors B, Scheurlen W, Felsberg J, Reif- Dürken M, Kulozik AE, Madden J, Donson A, Fore- enberger G, Northcott PA, Taylor MD, Meyerson M, man NK, Drissi R, Fouladi M, Scheurlen W, von Deim- Pomeroy SL, Yaspo ML, Korbel JO, Korshunov A, Ei- ling A, Monoranu C, Roggendorf W, Herold-Mende ls R, Pfister SM, Lichter P. Dissecting the genomic C, Unterberg A, Kramm CM, Felsberg J, Hartmann complexity underlying medulloblastoma. Nature C, Wiestler B, Wick W, Milde T, Witt O, Lindroth 2012; 488: 100-105. AM, Schwartzentruber J, Faury D, Fleming A, Zakrze- 41) Parsons DW, Li M, Zhang X, Jones S, Leary RJ, Lin wska M, Liberski PP, Zakrzewski K, Hauser P, Gara- JC, Boca SM, Carter H, Samayoa J, Bettegowda C, mi M, Klekner A, Bognar L, Morrissy S, Cavalli F, Gallia GL, Jallo GI, Binder ZA, Nikolsky Y, Harti- Taylor MD, van Sluis P, Koster J, Versteeg R, Volck- gan J, Smith DR, Gerhard DS, Fults DW, VandenBerg mann R, Mikkelsen T, Aldape K, Reifenberger G, Col- S, Berger MS, Marie SK, Shinjo SM, Clara C, Phil- lins VP, Majewski J, Korshunov A, Lichter P, Plass C, lips PC, Minturn JE, Biegel JA, Judkins AR, Resnick Jabado N, Pfister SM. Hotspot mutations in H3F3A AC, Storm PB, Curran T, He Y, Rasheed BA, Fried- and IDH1 define distinct epigenetic and biological man HS, Keir ST, McLendon R, Northcott PA, Taylor subgroups of glioblastoma. Cancer Cell 2012; 22: MD, Burger PC, Riggins GJ, Karchin R, Parmigiani G, 425 - 437.

19