Introduction

Embryonal tumors represent the largest group of malignant brain tumors in childhood. The 2007 WHO classification of central nervous system tumors recognizes three main histological entities all of which contain a population of primitive (“embryonal”) cells, i.e., medulloblastoma, atypical teratoid/rhabdoid tumor, and CNS primitive neuroectodermal tumors (PNETs) [8]. In addition, the latter entity is subdivided into CNS neuroblastoma/ganglioneuroblastoma, medulloepithelioma, and ependymoblastoma. Recently, a yet different pediatric embryonal brain tumor has been described by Eberhart and co-workers that occupies an intermediate position between CNS neuroblastoma and ependymoblastoma [4]. Based on its characteristic histopathological findings, this tumor was designated “embryonal tumor with abundant neuropil and true rosettes (ETANTR)”. This tumor entity shows a unique combination of histological, immunohistochemical and ultrastructural features. To date, only 13 cases of these tumors have been reported [3–6]. All patients were very young children (range from 1 to 4 years). In contrast, there were no correlations with regard to gender and tumor localization with tumors occurring in cerebral hemispheres, cerebellum, and brain stem.

Discussions are proceeding on whether these tumors represent a distinct entity or only a variation of parent PNETs. Molecular analysis by fluorescence in situ hybridization (FISH) was only performed in three cases and two of them showed abnormalities of chromosome 17. Therefore, molecular similarity of these tumors to medulloblastoma has been suggested. In the present study, we performed array-CGH analysis and microRNA (miRNA) screening of one ETANTR and identified a novel high-level amplification within chromosomal region 19q13.42 containing a cluster of miRNA coding genes. This finding has not been detected in any other pediatric brain tumor studied with the same array-CGH platform to date (approximately 200 cases), which allows us to hypothesize a unique nosologic position of ETANTR.

Materials and methods

Case report

A 2-year-old girl presented with a 1-month history of personality changes and ataxia. Morning headache, nausea and vomiting appeared 2 weeks before admission to the Burdenko Neurosurgical Institute, Moscow, Russia. In addition, her parents noted appearance of divergent strabism. Neurological examination disclosed ataxia, tetraparesis and central right seventh cranial nerve palsy.

CT and MRI scans with contrast enhancement revealed a homogeneous and poorly contrasted tumor mass located in the middle and upper parts of the vermis cerebelli with protrusion into the cavity of the fourth ventricle. The upper part of the tumor also spread into the pineal region through the tentorial incisure. Occlusive hydrocephalus with dilatation of the third and both lateral ventricles was obvious at presentation. Spinal MRI and CSF cytological analysis at admission were negative.

Subtotal resection of the tumor could be achieved by open neurosurgery. Postoperative MRI disclosed a tumor residue near the tentorium cerebelli, and the large cerebral vein (Galeni) was included in the residual tumor. Postsurgical course was uneventful and some improvement in patient’s neurological status was noted. The child was then treated with high-dose chemotherapy (including stem cell rescue) and thereafter she received local radiotherapy for a total dose of 56 Gy to the tumor bed. Finally, she received six courses of maintenance chemotherapy consisting of cisplatin, lomustine and vincristin. Eighteen months after subtotal tumor resection, the patient is alive and craniospinal MRI performed 17 months after the operation showed no evidence of tumor re-growth with a stable size of the tumor remnants.

Immunohistochemistry

For immunohistochemical analysis, a panel of antibodies directed against the following antigens was applied: glial fibrillary acidic protein (GFAP) (6F2; 1:200: DAKO); synaptophysin (polyclonal; RTU; DAKO); neurofilament protein (2F11; RTU; DAKO); smooth muscle actin (1A4; RTU; DAKO), vimentin (V9; RTU; DAKO); EMA (E29; RTU; DAKO); cytokeratin (AE1/AE3; RTU; DAKO); cytokeratin CAM 5.2 (5D3; 1:50; BioGenex), CD99 (12E7; RTU; DAKO); NeuN (MAB 377; 1:1000;Chemicon); BAF47/INI1 (BAF47; 1:80; BD Biosciences); Ki-67 (SP-6; 1:200; NeoMarkers). For antigen retrieval, the sections were microwaved in antigen unmask solution (BD Bioscience San Jose, CA, USA). Immunostaining was visualized by means of the DAKO Envision system using diaminobenzidine as a chromogen.

Nucleic acid isolation

Extraction of high-molecular-weight DNA and total RNA from frozen tumor samples was carried out by use of cesium chloride ultracentrifugation [13]. Genomic DNA from peripheral blood mononuclear cells of healthy donors (pool of ten donors, age 25–40 years for each gender) was isolated by use of the Qiagen DNA Blood Midi-Kit (Hilden, Germany).

Array-based comparative genomic hybridization (array-CGH)

Array-CGH (matrix-CGH) [12] was carried out as previously described [9, 15]. Selection of genomic clones, isolation of BAC DNA, performance of DOP-PCR, preparation of microarrays, labeling, hybridization, and washing procedures were performed as outlined.

Microarray data analysis

Raw data processing and normalization was performed as previously reported [9] providing log2-ratios of spot intensities. The chromosomal mapping information is based on the University of California at Santa Cruz (UCSC) genome database (March 2006) and the March 2006 (hg18) assembly (NCBI Build 36.1) of the International Human Genome Sequencing Consortium.

Fluorescence in situ hybridization (FISH)

Two-color interphase FISH to the sections was performed using the following fluorochrome-labeled commercial probe sets (Vysis, Inc.): centromere (CEP)7/7p12(EGFR) dual color probe set; CEP9/9p21(CDKN2A) dual color probe set; CEP10/10q23(PTEN) dual color probe set; CEP2/2p24(MYCN) dual color probe set; CEP8 and locus specific 8q24(MYC) probe; locus-specific 17p13(LIS1)/17q21(RARA) dual color probe set; mixed locus-specific 1p36/1q25 and 19p13/19q13 dual color probe sets; orange-labeled 20q13(ZNF217) locus probe, green-labeled 22q11(BCR) locus probe, and subtelomere probes for 19p13.3 and 19q13.4. Pretreatment of slides, hybridization, post-hybridization processing, and signal detection were performed as previously described [1]. Signals were scored in at least 200 non-overlapping, intact nuclei. Metaphase FISH for verifying clone-mapping position was performed using peripheral blood cell cultures of healthy donors as previously outlined [7].

miCHIP analysis

Total RNA (300 ng) was labeled with a Cy3-conjugated RNA linker (Biospring, Frankfurt, Germany) and hybridized to miCHIP as previously described [13]. miCHIP is based on locked nucleic acid (LNA) technology, whereby LNA-modified, Tm-normalized miRCURY capture probes (Exiqon, Denmark) designed to target 550 unique human miRNAs (miRbase v9.2, Wellcome Trust Sanger Institute, http://microrna.sanger.ac.uk/) were printed onto Codelink slides (GE Healthcare, Chalfont St Giles, UK). Array images were generated by using a Genepix 4200AL laser scanner (Molecular Devices, USA) where miCHIP arrays were scanned using the Genepix auto PMT (Photo Multiplayer) algorithm, with pixel saturation tolerance set to 0.2%. Tiff images generated by the Genepix 4200AL laser scanner were processed by the Genepix 6 microarray analysis software (Molecular Devices, USA). Artifact-associated spots were eliminated both by software- and visual-guided flags. Signal intensities were measured according to the local background subtraction method. All probes were spotted in duplicate and the median signal intensity of these was calculated. miCHIP data were median normalized for subsequent analysis and log2-ratios (tumor/control pool of normal cerebellum) were calculated.

Quantitative real-time PCR (QRT-PCR)

To measure PRKCG1 and TFPT mRNA abundance in the tumor sample, a reference of total RNA obtained from normal human cerebellum of 24 individuals (age 16–70; BD Bioscience, USA) was used for tissue-specific normalization. Each cDNA sample was analyzed in triplicate with the ABI PRISM 7700 (Applied Biosystems, USA) using Absolute SYBR Green ROX Mix (ABgene, UK) according to the manufacturer’s instructions. Two endogenous housekeeping genes (PGK1, DCTN2) were used as internal standards. These genes were not regulated in expression profiling experiments comparing pediatric low-grade astrocytomas and medulloblastomas. Furthermore, they were tested in normal tissues from whole brain and cerebellum in which they displayed very little variability. All primers were tested to exclude amplification from genomic DNA. The relative quantification of the RNA of interest in comparison to the housekeeping genes was calculated according to a previously published algorithm with the modification that two housekeeping genes were used in the present study [10]. Oligonucleotide sequences are available upon request.

To validate our miChip findings, QRT-PCR of selected miRNAs was performed using commercially available Taqman Kits and miRNA-specific probes from Applied Biosystems (USA). Taqman analysis was carried out according to the manufacturer’s instructions. RNU66 and RNU6B were used as housekeepers. miRNA abundance in the tumor was normalized against the housekeepers and plotted against normal cerebellum (BD Bioscience, USA, as for the protein-coding genes) or whole brain (pool of donors, Biocat, Germany).

Results

Pathological examination

On light microscopical investigation, the tumor showed a biphasic pattern featuring highly cellular clusters of small cells with round or polygonal nuclei and scanty cytoplasm admixed with large fibrillar and paucicellular neuropil-like areas (Fig. 1a). Among the clusters and aggregates of small cells, numerous true ependymoblastic-like rosettes were identified (Fig. 1b). The latter were mitotically active and multilayered with well-formed central round or slit-like lumina in the absence of an outer membrane. The lumina of the rosettes were either empty or filled with eosinophilic debris. Perivascular and Homer-Wright rosettes were absent. Mitotic figures were frequently observed in highly cellular areas and in the rosettes as well as prominent apoptotic bodies. There were no microvascular proliferates and necrotic areas in this neoplasm. The combination of histopathological patterns mentioned above fulfills all diagnostic criteria of an “embryonal tumor with abundant neuropil and true rosettes” (ETANTR), as previously described by Eberhart and co-workers [4].

Fig. 1
figure 1

Microscopic appearance and immunohistochemical patterns of the ETANTR. a (H&E; ×100) Neurocytic paucicellular areas with numerous rosettes embedded within abundant neuropil. b (H&E; ×200) Clusters of small cells and rosettes; inset of an ependymoblastic-like rosette (H&E; ×400). c Strong expression of synaptophysin in paucicellular areas and neuropil (X100); inset only single cells are synaptophysin-positive in cell-rich areas (×200). d Strong vimentin expression in cell-rich areas (×200); inset vimentin expression in rosette-forming cells (×400). e Single cells stain positive for GFAP (×200). f Expression of Ki-67 was predominately identified in cell-rich areas (×100); inset Ki-67 expression within an ependymoblastic rosette (×400)

Full size image

Immunohistochemical analysis revealed strong positivity for synaptophysin within the neuropil-like tumor part, whereas only some positive cells were observed within the cell-rich tumor part (Fig. 1c). Nuclei of cells within neuropil were also positive for Neu-N antigen. Additionally, the paucicellular tumor part included single ganglioid-like cells that were highlighted upon immunostaining of neurofilament proteins. In contrast, the cell-rich part of the tumor was strongly positive for vimentin; which was also positive in the cells forming true rosettes (Fig. 1d). CD99 immunoreactivity was predominantly membranous or dot-like and was detected within the apical parts of the primitive cells forming true rosettes. Tumor cells were negative for actin, EMA and cytokeratins, whereas single entrapped astrocytes were positive for GFAP (Fig. 1e). Nuclear expression of the BAF47/INI1 protein was found in tumor nuclei throughout both components of the tumor. Ki-67 nuclear expression was found predominantly in the densely cellular areas including rosettes (Fig. 1f), and the mean Ki-67 index was 47%.

DNA-copy number aberrations

Array-based comparative genomic hybridization revealed multiple, predominantly yet unknown DNA copy number imbalances including aberrations affecting entire chromosomes, e.g., trisomy of chromosomes 2 and 19 (Fig. 2a). Most interestingly, a novel high-level genomic amplification of 0.89 Mb at chromosome band 19q13.42 was identified, which has not been described in any other human cancer to date (Fig. 2c). The amplification was found to cover a total of 29 known protein-coding genes (Supplementary Table 1). Based on their reported and putative role in cancer, amplification of the TFPT gene, which acts as a fusion partner of E2A in childhood ALL, and the PRKCG gene, a member of the protein kinase C family of genes, were of particular interest. In addition, 25 genes coding for a cluster of miRNAs at 19q13.42 were affected by this amplification (Supplementary Table 2). Regions of copy-number aberrations smaller than 3 Mb in size are listed in Table 1. Small copy-number aberrations defined possible candidate genes known to be involved in tumorigenesis. Gains included potent oncogenes such as CDK3, and FOXJ1. Losses harbored genes involved in cell adhesion (DNM3, EMCN) among others.

Fig. 2
figure 2

Array-CGH and miRNA-expression profiling results. a Genome-wide array-CGH profile of the ETANTR hybridized against a pool of DNA derived from mononuclear peripheral blood cells from 10 healthy donors. Clones are plotted in order of chromosomal annotation (x-axis). Ratios (ETANTR/control) are given as log2-ratios (y-axis). Large genomic aberrations were detected on chromosomes 2 and 19 (trisomies). As indicated by an arrow, a high-level genomic amplification (>4 copies on average) was observed at 19q13.42. Results of large aberrations were validated by FISH analysis. Labeling of probes was as follows: CEP2 green/2p24 red, 19p13.3 green/19q13.3 red, and 19p13.3 green/19q13.4 red. b Genome-wide miRNA profile as obtained by comparison of miRNA expression in the ETANTR with expression in a pool of normal cerebellum (24 healthy individuals). Clones are plotted in chromosomal order (x-axis) and log2-ratios are given for ETANTR/control. c Magnification of the array-CGH profile at chromosome 19 indicating a low-level gain for the whole chromosome and a high-level amplification (>4 copies on average) at 19q13.42. d Expression of selected miRNAs at chromosome band 19q13.42 was validated by QRT-PCR. All tested miRNAs were highly upregulated in the tumor when compared to either normal cerebellum (light gray bars) or whole brain (dark gray bars). Expression ratios are given as log2-ratios. Normalization was performed against two housekeeping genes (RNU66 and RNU6B) which showed comparable expression in all samples

Full size image
Table 1 Small genomic aberrations <1 Mb in size
Full size table

Confirmation of findings by FISH

FISH analysis of selected loci confirmed a low-level gain of chromosome 2 (>10% of nuclei with more than 3 signals for both 2p24/MYCN and CEP2 probes), as well as chromosomal gains of 19p and 19q identified by using both, subtelomere and locus-specific probes (Fig. 2a). Probes against chromosomes 1, 7, 8, 9 10, 17, 20q and 22q revealed normal dosages (i.e., two copies).

mRNA expression of candidate oncogenes

Messenger-RNA expression of the candidate oncogenes TFPT and PRKCG1 was assessed by quantitative real-time PCR. mRNA abundance in the tumor was normalized against two housekeeping genes (PGK1, DCTN2). When compared to the reference RNA (pool of normal cerebellum from 24 donors, age 16–70 years), no significant up-regulation of these genes was detected (data not shown).

Overexpression of the miRNA cluster located at 19q13.42

To assess whether the cluster of 25 known miRNA genes at 19q13.42 might be a target of the genomic amplification, we performed a genome-wide miRNA-profiling using a LNA-based approach as previously described [2]. Interestingly, the miRNA-genes at 19q13.42 were among the most abundantly expressed miRNAs across the genome (Fig. 2b; Supplementary Table 2). Expression of selected miRNAs was validated by QRT-PCR (Fig. 2d). The function of the miRNAs showing the highest up-regulation in the chip-experiment (miRNA-518b, miRNA-519d, miRNA-520-3p, miRNA-517) is yet unknown. Interestingly, however, two of the most up-regulated miRNAs measured by QRT-PCR were mir-372 and mir-373, which have recently been implicated in tumorigenesis [14].

Discussion

Since the initial description by Eberhart et al. [4], 14 cases of ETANTRs including our case have now been reported (Table 2). According to the statement by Dunham et al. [3], additional rare cases have likely been described in retrospect under different designations. According to the clinical data, ETANTRs are exclusively pediatric neoplasms with various localizations including cerebral hemispheres, cerebellum, and brainstem. Prognosis is regarded as extremely poor; however, one case with an event-free survival up to 34 months after intensive courses of hyper-fractionated radiotherapy and polychemotherapy has been documented [5]. In the case presented here, a cerebellar tumor in a 2-year-old child was subtotally removed, and high-dose chemotherapy followed by irradiation and maintenance chemotherapy resulted in 17 months progression-free survival.

Table 2 Clinical and molecular findings in all cases of ETANTR reported in the literature to date
Full size table

Histopathologically, ETANTRs have a typical biphasic appearance. Usually ependymoblastoma-like rosettes and clusters of primitive neuroepithelial cells are embedded in areas of abundant neuropil. The present case showed predominately neuronal differentiation in both tumor components. Nevertheless, neuronal differentiation was more prominent in the neuropil-like matrix, whereas the primitive cells were intensively positive for vimentin.

Ultrastructural studies of several ETANTR cases describe patterns characteristic for neuronal (processes with numerous microtubules and neurosecretory granules) and ependymal (true rosettes with zonula adherences and basal bodies) cell differentiation. Cytogenetic findings were presented in two studies only. Fuller et al. [5] applied FISH analysis in two cases and described chromosome 17 abnormalities [including i(17q) in both tumors]. MYC/MYCN amplification as well as deletion of the SMARCB1 locus was excluded. In contrast, Dunham et al. [3] studied one tumor and revealed normal DNA copy-numbers for the MYC and MYCN loci, and for chromosome 17.

In the present study, we performed genome-wide analysis of the ETANTR by using array-CGH with an average resolution of 0.5 Mb. Among the large-scale aberrations, there were trisomies of chromosomes 2 and 19 with a high-level amplicon at 19q13.42 spanning 0.89 Mb and covering numerous genes involved in carcinogenesis. We validated gains of chromosomes 2 and 19 applying FISH with corresponding probes. Besides, we found a few small regions of copy-number aberrations harboring potent oncogenes and tumor suppressor genes, although their biologic relevance for the pathogenesis of ETANTR requires further investigation.

Amplifications on 19q have been described in various human cancers, but in all these cases, more proximal parts of the chromosome arm (e.g., 19q12–13.3) were affected targeting well-established oncogenes (e.g., CCNE and AKT2). To our knowledge, this is the first description of a genomic amplification at 19q13.42. This region harbors 25 genes encoding a cluster of miRNAs. MicroRNAs have been implicated in different biological processes including human cancer. Various miRNAs were found to be involved in the pathogenesis of lung cancer, breast cancer, and lymphomas where they cooperate with oncogenes in cell transformation. Among the miRNAs located within the amplified region at 19q13.42, miRNA-372 and miRNA-373 have been recently identified as potent oncogenes in development of germ cell tumors by numbing the p53 pathway [14]. We performed miRNA expression profiling using a microarray covering most of the known human miRNAs, and found that expression of most of the miRNAs located within the miRNA cluster at 19q13.42 was highly upregulated in comparison with normal cerebellum, in which, we found virtually no expression of these miRNAs. The upregulation of selected miRNAs was confirmed by QRT-PCR. Expression of miRNA-372 and miRNA-373 was upregulated in the tumor by more than 30-fold in comparison with normal cerebellum. These findings suggested a role of the miRNA cluster at 19q13.4 in the molecular pathogenesis of ETANTR.

The distinct nosologic position of ETANTR is still unclear. Clearly, ETANTR is an embryonal neoplasm that shows a wide spectrum of neuronal differentiation and additional features of ependymal differentiation. Eberhart et al. considered ETANTR to be a unique tumor entity that appears to be distinct from other forms of embryonal CNS tumors. In contrast, La Spina et al. suggested that ETANTR overlaps with other PNETs, including medulloblastomas with neuronal differentiation.

We compared the results of the current study with the results of previously performed array-CGH analyses of 100 medulloblastomas, 10 supratentorial PNETs, 12 AT/RTs, and 70 ependymomas applying the same microarray platform. Consequently, we found that the amplification on 19q13.42 has not been identified in any of these tumors indicating a specific aberration in ETANTR. Moreover, CGH analysis of four ependymoblastomas did not reveal aberrations on chromosome 19, although three of these tumors showed complete gains of chromosome 2 [11]. These findings allow us to hypothesize a unique nosologic position of ETANTR. Nevertheless, taking into account the rarity of these tumors, further molecular investigations will be necessary to finally answer this question. Such genome-wide screening analyses will be an important line of future research to identify the cell of origin of this infrequent tumor.