• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Cancer Cell. Author manuscript; available in PMC Aug 31, 2012.
Published in final edited form as:
PMCID: PMC3431561

Frequent amplification of a chr19q13.41 microRNA polycistron (C19MC) in aggressive primitive neuro-ectodermal brain tumors


We discovered a high-level amplicon involving the chr19q13.41 microRNA (miRNA) cluster (C19MC) in 11/45(~25%) primary CNS-PNET which results in striking over-expression of miR-517c and 520g. Constitutive expression of miR-517c or 520g promotes in vitro and in vivo oncogenicity, modulates cell survival and robustly enhances growth of untransformed human neural stem cells (hNSCs) in part by upregulating WNT pathway signaling and restricting differentiation of hNSCs. Remarkably, the C19MC amplicon, which is very rare in other brain tumors (1/263), identify an aggressive sub-group of CNS-PNET with distinct gene expression profiles, characteristic histology and dismal survival. Our data implicate miR-517c and 520g as oncogenes and promising biological markers for CNS-PNET and provide important insights into oncogenic properties of the C19MC locus.


Brain tumours are the leading cause of childhood cancer–related morbidity and mortality. Primitive neuroectodermal tumors (PNET), which are the most frequent malignant pediatric brain tumors include medulloblastoma, a cerebellar PNET, and CNS-PNET (also called supratentorial PNET or sPNET) – an aggressive cerebral tumor (McLendon RE, 2007). Comprehensive genetic studies of a substantial number of CNS-PNET have not yet been undertaken due to their relatively rarity. Lack of insight into the molecular pathogenesis of CNS-PNET is a major obstacle towards development of disease-specific models and treatments for these frequently fatal tumors.

CNS-PNET, which represents 3-7% of all pediatric brain tumors, are a heterogeneous group characterized by primitive neuroepithelial cells with variable neuronal, glial or ependymal differentiation. Unlike rhabdoid tumors, which are defined by INI1gene alterations (Allen et al., 2006), CNS-PNET may pose significant diagnostic challenges due to lack of characteristic genetic or immunohistochemical markers (McLendon RE, 2007). In addition to CNS-PNET, several other PNET histologic variants may arise in the cerebral hemispheres; these include ependymoblastoma and medulloepithelioma, which have characteristic histologic features such as ependymal rosettes and primitive neural tube formation, and ENATR – a newly described entity with ependymoblastic differentiation (Gessi et al., 2008; McLendon RE, 2007). Whether CNS-PNET and PNET variants represent distinct molecular tumor sub-types remains debated(Burger, 2006; Judkins and Ellison, 2008).

CNS-PNET and medulloblastoma share very similar histology and are largely distinguished by tumor location. However, CNS-PNET are more aggressive tumors with significantly inferior outcomes (Fangusaro et al., 2008; Timmermann et al., 2006) than medulloblastoma (Packer et al., 2006). Cumulative data suggest CNS-PNETs and medulloblastoma have distinct genetic features and histogenesis. Notably, CNS-PNET lack expression of cerebellar granule cell specific transcription factors (Pomeroy et al., 2002) and isochromosome 17q (i17q), which characterize medulloblastoma (Inda et al., 2005; McCabe et al., 2006; Pfister et al., 2007) and rarely arise in murine medulloblastoma models (Zindy et al., 2003).

Limited CGH studies indicate CNS-PNETs are genetically heterogeneous with frequent but diverse copy number aberrations (CNAs). Genes targeted by CNAs in CNS-PNET remains largely unknown; to date amplification of MYCN, PDGFB and PDGFRA and CDKN2A/2B deletions have been reported (Inda et al., 2005; McCabe et al., 2006; Pfister et al., 2007). However, the significance of these and other reported genetic alterations to the biology and clinical phenotypes of CNS-PNET is not known.

MicroRNAs (miRNAs) are short ~22 nucleotide non-coding RNA which function as important regulators of gene expression in various cell types. They are implicated in diverse biological processes including self-renewal and pluri-potentcy (Bar et al., 2008) and as etiologic genes in various human malignancies (Esquela-Kerscher and Slack, 2006). Patterns of miRNA expression have been shown to distinguish tumor types and to predict tumor biology (Lu et al., 2005). Direct genetic alterations of a few miRNAs have also been reported in human cancers, these include MiR15/16a in CLL (Calin et al., 2002), MIR-17-92 in B-cell lymphoma (He et al., 2005) and LET-7 in lung cancer (Johnson et al., 2005). Although miRNAs have been implicated in malignant gliomas (Papagiannakopoulos et al., 2008), their role in malignant pediatric brain tumors remains largely unexplored.

In this study we sought to comprehensively define genetic alterations in CNS-PNET by performing high resolution DNA copy number and gene expression analysis on a substantial number of primary tumors.


Identification of a frequent high level amplicon at chr19q13.41 in CNS-PNET

We performed high resolution copy-number and gene expression analyses on 39 and 33 primary tumors respectively (Table S1). Global DNA copy number determined using the 500K Affymetrix SNP arrays and dChip analysis (Lin et al., 2004) revealed frequent and multiple CNAs (10-28/tumor) in most tumors (Fig S1; Table S2, 3) and collectively confirm prior observations (Inda et al., 2005; McCabe et al., 2006; Pfister et al., 2007) that CNS-PNET are a genetically heterogeneous but distinct group of embryonal brain tumours.

Despite substantial genomic heterogeneity, we observed recurrent chr 2 gains and focal chr19q13.41 amplification in a significant proportion (8/39) of CNS-PNET. Two tumors had non-overlapping chr2p24.3 and chr2q33.3 amplicons encompassing MYCN and putative oncogenes, CREB1 and FZD5 (Fig 1A), however, gene expression profiles did not correlate MYCN, CREB1 or FZD5 expression with chr2 gains in tumors (data not shown). In contrast, the chr19q13.41 amplicon in all 8 tumors mapped to a common ~1Mb interval between the DPRX and LILRP2 loci (Fig 1B). Notably, the chr19q13.41 amplicon was absent in matched lymphocyte DNA (Fig 1C), indicating it was tumor-specific and not a normal DNA copy number variant (Iafrate et al., 2004). We used FISH analyses to validate the genotyping data (Fig 1D), and to test for the chr19q13.41 amplicon in 5 additional tumors that were not included in genotyping analysis. Collectively, the amplicon was detected in ~25% (11/45) of tumors analyzed; a remarkably high frequency indicating an important CNS-PNET oncogene(s) maps to chr19q13.41.

Figure 1
Frequent chr2 gains and chr19q13.41 amplication in CNS-PNET

MicroRNAs of the C19MC locus are the targets of the chr19q13.41 amplicon

Genotyping and qPCR analysis indicated the minimal chr19q13.41 amplicon in CNS-PNET spanned 34 protein-coding loci and 2 miRNA clusters (Fig 2A, B). We examined transcriptional profiles of 33 CNS-PNET by t-tests to determine if any coding loci exhibited copy-number driven changes in expression (Fig 2C); qRT-PCR validation of select genes is shown in Fig S2A. Only 1 gene exhibited a modest 2-fold increased expression (CNOT-3; p= 0.02), indicating miRNA clusters as likely targets of the chr19q13.41 amplicon. The chr19q13.41 amplicon spans a large, poorly characterized, primate-specific miRNA cluster, C19MC and the MIR-371-373 locus which map ~100kb apart. Although, miR371-373 has been implicated in germ cell tumors (Voorhoeve et al., 2006), little is known regarding the oncogenic potential of the C19MC locus which encodes 54 predicted miRNA (Bentwich et al., 2005).

Figure 2
MicroRNAs are the oncogenic target of the chr19q13.41 amplicon

To identify potential oncogenic miRNAs, we used a miRNA array to assay expression of the C19MC and MIR371-373 clusters (Table S4) in chr19q13.41 amplified tumors. These analyses revealed robust expression of several C19MC miRNAs in tumors relative to fetal brain, but only modest expression of miR371-373 (Fig 2D). With exception of miR-512-3p and 520c-3p, miRNA levels was confirmed by gene-specific qRT-PCR (Fig S2B-D), and significantly correlated (p<0.0001) with chr19q13.41 amplification (Fig 2E). We observed highest tumor-specific expression (up to 40X greater than other miRNAs) of miR-520g and miR-517c in chr19q13.41 amplified tumors (Fig 2F) and therefore focused on miR-520g and 517c to test for oncogenic potential of the C19MC locus.

MiR-520g and 517c have oncogenic effects in vitro and in vivo

To test if miR-520g and 517c have oncogenic properties we characterized the effects of stable miRNA expression (Fig 3A) in hemispheric PNET (PFSK) and medulloblastoma, (Daoy) cell lines, which have low endogenous miR-520g and 517c expression (Fig S3), as well as murine NIH3T3 cells using growth, soft agar transformation and nude mouse xenograft assays. Although, the C19MC locus is primate-specific, we hypothesized that genes with important oncogenic functions will target conserved, critical downstream pathways across species and cell-types. Indeed miR-520g or 517c expression enhanced cell growth and/or transformation in all 3 cell types (Fig 3B-D). Interestingly we observed cell-type and miRNA dependent phenotypic changes that were not associated with relative miRNA levels. While miR-517c expression enhanced growth of all 3 lines (p<0.001) and promoted soft agar growth and xenograft formation in Daoy cells, NIH3T3-miR-517c exhibited limited in vitro transformation (Fig 3C), but formed robust xenografts (Fig 3D). Similarly, miR-520g expression had limited effects on cell growth (Fig 3B) but promoted soft agar growth and xenograft formation in Daoy and NIH-3T3 cells (Fig 3C, D). These observations collectively support oncogenic functions for miR-520g and 517c and suggest that, like other oncogenic miRNAs (Bonci et al., 2008), miR-520g and 517c may effect cell transformation via multiple mechanisms.

Figure 3
MiR-520g and 517c have oncogenic effects in vitro and in vivo

miR-520g and 517c expression promotes cell survival and inhibits neural stem cell differentiation

To investigate how miR-520g or 517c might drive oncogenesis, we examined tumor transcriptional profiles to identify gene expression changes most significantly correlated with C19MC amplification using t-test statistics (Fig 4A). A resulting set of 381 and 501 genes respectively with up and down-regulation (Table S5), in C19MC amplified tumors were further investigated using the gene function annotation tool, GOSTAT (Beissbarth and Speed, 2004) (Table S6). Interestingly, apoptosis regulators were significantly down-regulated in C19MC amplified tumors (22 of 501 genes, p=5.57e-8); 11/22 genes, including API5, BAD and TRADD, had pro-apoptotic functions. Consistent with these observations, miR-520g or 517c expressing PFSK, Daoy and NIH3T3 cells exhibited increased cell viability and decreased apoptosis in serum deprived media (Fig 4B, C). Collectively these data indicate that miR-520g and 517c promote oncogenesis in part by modulating cell survival pathways.

Figure 4
C19MC amplification and constitutive miR-517c and 520g expression is associated with a survival phenotype

Interestingly, GOSTAT analyses also revealed enrichment of developmental signaling genes in C19MC amplified tumors (83 of 381 genes, p= 4.79 e-26; Table S6). In particular, self-renewal and survival genes including SOX11, NR2F1, NKX2-2, FGF13, FGFr3, SALL4 and WNT pathway ligands (WNT5A, 7A) and receptors (FZD2, 3, 7, 10) were over-represented. We investigated if ectopic miR-520g or 517c expression could alter the phenotype of primary human neural stem cell line (hNSCs) which normally continually replicate in growth-promoting/proliferative media containing high FGF and EGF but undergo growth arrest in factor-deprived/differentiation conditions such as 10% serum media (Sun et al., 2008). Stable hNSC-miR-520g and 517c cells did not exhibit significant changes in growth under proliferative conditions (Fig S4), however miR-520g and to a lesser extent, miR-517c expression, strikingly enhanced hNSCs growth in differentiation media (Fig 5A, B; S4) indicating that miR-520g or 517c may alter the differentiation potential of hNSCs.

Figure 5
MiR-520g and 517c expression drives growth and inhibits differentiation of human neural stem cells

We used immuno-cytochemical and qRT-PCR analyses to compare expression of Tubulin β-III and GFAP, which respectively mark neuronal and astrocytic differentiation, in hNSC-miR-520g and 517c cells grown in proliferative and differentiation conditions. GFAP expression was not significantly altered in hNSCs-miR-520g and 517c cells as compared to controls (data not shown). Notably, although hNSCs-miR-520g or 517c cells did not exhibit significant changes in Tubulin β-III expression under proliferative conditions, Tubulin β-III expression in hNSC-miR-520g cells grown in differentiation media was strikingly diminished as compared to controls (15% vs 65% Tubulin β-III positivity; Fig 5C, D). Interestingly qRT-PCR analysis revealed expression of NKX2-2 and OLIG1 two genes with essential roles in normal neuronal cell fate specification and differentiation (Briscoe et al., 2000; Zhou and Anderson, 2002), was also diminished in miR-520g as well as 517c expressing hNSCs (Fig 5D). We also observed that robust effects of miR-520g on hNSC differentiation was accompanied by striking changes in expression of neural stem/progenitor cell markers in hNSCs grown under differentiation conditions. Specifically qRT-PCR analysis revealed up-regulation of early neural stem/progenitor cell markers CD44, CD133 and POU3F2, in stable hNSC-miR-520g as well as 517c cells grown in proliferative media, but only hNSC-miR-520g cells maintained or exhibited robust expression (~3-7 fold greater than controls) of neural stem/progenitor cell markers in differentiation media (Fig 5E). These data suggest that miR-520g, and to a more limited extent miR-517c, promotes hNSC growth by restricting differentiation and enhancing neural stem/progenitor cell phenotypes. Interestingly, expression of several C19MC miRNAs, including miR-520g, has also been correlated with human ES cell (hESC) differentiation (Bar et al., 2008). Together with our observations, these data highlight an important role for miR-520g and possibly other oncogenic C19MC miRNAs in maintenance of primitive/progenitor cell phenotypes.

miR-520g expression correlates with altered WNT signaling

Substantial data implicate aberrant developmental signaling in the pathogenesis of embryonal brain tumors including CNS-PNET (Fogarty et al., 2005). Indeed, expression analysis revealed significant enrichment of developmental signaling genes, particularly those of the WNT pathway in C19MC amplified CNS-PNET (Fig 4A, Table S6) indicating that one or more C19MC miRNA genes may function in WNT signaling. Notably, the WNT pathway, which plays a critical role in neural stem cell survival and lineage determination (Kalani et al., 2008; Wexler et al., 2009), has been linked to miR-520g by recent computational studies (Gaidatzis et al., 2007). We focused on miR-520g to investigate signaling pathways associated with the C19MC oncogenic locus, as miR-520g expression correlated with the most robust phenotypic changes in hNSCs.

To uncover signaling pathways associated with miR-520g cellular phenotypes, we performed gene expression analysis of hNSC-miR-520g and PFSK-miR-520g and control cell line using the Illumina BeadChip arrays and used t-tests to identify gene sets most significantly correlated with miR-520g expression. Not surprisingly, consistent with broad regulatory functions of miRNAs, we observed substantial gene expression changes in both hNSCs and PFSK cells with stable miR-520g expression; ~1000 to 5000 differentially expressed genes were respectively identified in PFSK and hNSC cell lines with miR-520g expression (p<0.05). Functional analyses of miR-520g targets predicted by several different algorithms and of the miR-520g associated gene sets determined from both cell line gene expression data, consistently revealed most significant enrichment of genes with functions in neural development (data not shown). We therefore performed gene enrichment analysis on miR-520g associated gene sets to investigate if developmental signaling pathways which have been implicated in embryonal brain tumor development, including the WNT, SHH, NOTCH and TGF-β pathways, were altered by miR-520g expression. We used the KEGG pathway database to identify and determine the number of specific pathway genes assayed by the Illumina BeadChip and applied multiple hypotheses testing to calculate relative representation of each pathway in the miR-520g associated gene sets. Strikingly, consistent with observations in C19MC amplified primary tumors (p=1.3 e-5), these analyses revealed most significant enrichment of WNT pathway loci in hNSC-miR-520g (p=1.3 e-8) and a similar more modest trend in PFSK-miR-520g cells (p<0.016) (Table S7; Fig 6A).

Figure 6
MiR-520g expression correlates with altered WNT signaling

The WNT pathway encodes a complex network of ligand and receptor combinations which signal via β-catenin-dependent (canonical) or independent (non-canonical) pathways in various cell contexts. While expression of WNT2/3 class ligands typically increase canonical WNT activity and upregulate downstream proliferative genes, like CCND1 and MYCC, the WNT5A ligand family classically activate non-canonical signaling, but may also regulate the canonical pathway (Miller, 2002). Interestingly, in both hNSCs and PFSK cells, miR-520g expression correlated with up-regulation of WNT5A as well as SFRP2/3 and DKK1 genes which encode antagonists of the canonical WNT pathway (Table S8). We confirmed up-regulation of WNT5A, SFRP2/3 and DKK1 in hNSCs and PFSK-miR-520g cells by qRT-PCR analysis (Fig 6B, C). Of note, although expression of CCND1 and MYCC was diminished in miR-520g expressing cells (Fig 6B & C) we did not observe significant changes in transcriptional activity of a TCF/LEF-driven reporter gene (pTOP-FLASH) in PFSK-miR-520g cells (Fig S5). These collective observations suggests that dysregulation of WNT pathway genes observed in miR-520g expressing stable cell lines and C19MC amplified primary tumors may reflect altered non-canonical WNT signaling.

Downstream effectors and targets of non-canonical WNT signaling are poorly defined. It is known however that WNT5A can associate with specific Frizzled receptors to induce non-canonical WNT signaling and activate several downstream effector kinases including those of the Ca2+ dependent PKC and CAMKII, or the JNK pathways in various cell systems (Mikels and Nusse, 2006; Miller, 2002). To investigate if up-regulation of WNT5A in CNS-PNET cells had any effect on non-canonical WNT signaling we performed western blot analysis of PFSK-miR-520g cells using pan-PKC antibodies which identifies most Ca2+ dependent phospho-PKC isoforms (PKC α, βI, βII), anti-phospho-CAMKII as well total and phospho-specific JNK1/2 antibodies. Interestingly, whereas phospho-PKC or CAMKII levels were not altered with miR-520g expression, phospho-JNK2, but not JNK1 levels was consistently upregulated in PFSK cells with stable miR-520g expression as compared to controls (Fig 6D shows representative of 4 independent experiments) indicating that miR-520g maps upstream of JNK2. In aggregate our data correlates increased expression of WNT5A and canonical WNT inhibitors, SFRP2 and DKK1, with specific up-regulation of phospho-JNK2 in PFSK-miR-520g cells, and indicate a role for miR-520g in non-canonical WNT/JNK2 signaling. These findings are particularly intriguing as recent reports indicate canonical WNT activity may be linked to better survival in CNS-PNET (Rogers et al., 2009), and suggest C19MC amplification may portend poorer CNS-PNET phenotypes.

Amplification of the C19MC locus defines an aggressive group of CNS-PNET

To assess the significance of the C19MC locus in CNS-PNET pathogenesis, we investigated if C19MC amplification correlated with specific clinical features. Indeed, Kaplan Meier analysis of 26 evaluable patients revealed substantially poorer overall survival of patients with C19MC amplified tumors (mean survival 4+/−1.3 months in amplified vs 44+/−12.8 months in non-amplified cases; p<0.0001) (Fig 7A). Interestingly, the C19MC amplicon was also present at higher frequency in tumors with variant histological features of ependymal or ependymoblastic differentiation, (8/10 amplified versus 3/35 non-amplified tumors; p <0.0005) and younger age at diagnosis (10/11 amplified vs 14/26 non-amplified CNS-PNETs were <4yrs, p<0.026) (Fig 7B, C). These collective data suggested that the C19MC amplicon may identify a CNS-PNET sub-group with distinct clinico-pathological and genetic features. Consistent with this hypothesis, we observed co-segregation of C19MC amplified tumors in unsupervised cluster analysis of transcriptional profiles from 33 tumors (Fig 7C). Remarkably, in a survey of 263 other pediatric brain tumors by FISH analysis of tissue microarrays, we detected C19MC amplicon in only one non-CNS-PNET tumor (an ependymoma) (p<0.00001) (Fig 7D). These data not only highlight the importance of the oncogenic C19MC locus in CNS-PNET pathogenesis but suggest that the C19MC locus may represent a specific marker for this tumor group.

Figure 7
Amplification of the C19MC locus identifies CNS-PNET with poor survival and distinct tumor pathology


In this study we report the discovery of a large miRNA cluster on chr19q13.41, C19MC, that is frequently amplified in aggressive CNS-PNET and harbors miRNA oncogenes capable of driving growth of normal hNSCs. We provide evidence that miR-520g and 517c of the C19MC locus have oncogenic activity in vitro and in vivo, confers cell survival and promotes primitive/progenitor cell phenotypes by attenuating differentiation of hNSCs. We correlate C19MC amplification in primary tumors and stable miR-520g expression in hNSC and CNS-PNET cells with robust dysregulation of WNT pathway loci and propose that C19MC miRNA oncogenes confers aggressive tumor phenotype in part by modulating WNT pathway signaling. Our findings have significant implications for clinical management of embryonal brain tumors and provide exciting insights into the molecular pathogenesis of CNS-PNET and oncogenic properties of the C19MC locus.

Oncogenic functions of the C19MC locus

The functions of C19MC miRNAs are largely unknown, however many exhibit restricted expression in undifferentiated/germinal tissues and cell types, such as placenta, testes (Bentwich et al., 2005) and hESCs (Bar et al., 2008; Laurent et al., 2008) indicating functions in cellular differentiation. Indeed, Bar et al (2008) observed that expression of miR-520g, as well as 4 other C19MC miRNAs diminishes with hESCs differentiation. Intriguingly, they also demonstrate that C19MC may be regulated by NANOG and OCT4, known critical transcriptional regulators of ESC pluripotentcy, and propose that select C19MC miRNAs, including miR-520g, may mediate self-renewal and represent “signatures” of pluri-potentcy in hESCs. In agreement with these studies, we observed that miR-520g and 517c increases growth of hNSCs under differentiation conditions. Importantly, our studies indicate that miR-520g and 517c expression not only confers increased cell survival but also promotes stem/progenitor cell phenotypes. Therefore, we propose that aberrant expression of miR-520g and possibly other C19MC miRNAs, alters normal neuro-epithelial stem/precursor response to cell differentiation and survival stimuli, and promotes oncogenesis in part by enhancing target stem/progenitor cell pools available for additional transforming events. Notably although both miR-520g and 517c cells confers cell survival in Daoy and PFSK tumor cells, we observed more limited effects of miR-520g and 517c on growth and differentiation phenotypes of these tumor cell lines (Fig S6), indicating that miR-520g and 517c, may also manifest cell stage and type restricted cellular effects/functions. Future in vivo experiments will be necessary to directly assess the individual and collective effects of miR-520g, 517c as well other candidate C19MC miRNAs oncogenes, on survival and self-renewal properties of neural stem/progenitors cells.

A role for miR-520g in WNT signaling

We show that C19MC amplification and miR-520g expression strongly correlates with altered WNT signaling in primary tumors and stable tumor/hNSC cell lines respectively. These findings, which concur with computational analyses of miR-520g (Gaidatzis et al., 2007), are particularly interesting as substantial evidence indicate aberrant WNT signaling may mediate malignant transformation by promoting stem/progenitor cell expansion in various developmental context including the nervous system (Kalani et al., 2008; Reya and Clevers, 2005). Interestingly, WNT5A mediated inhibition of canonical WNT signaling has been shown to enhance hematopoietic stem/progenitor cell repopulation (Nemeth et al., 2007). We observed decreased CCND1 and MYCC expression in hNSCs and CNS-PNET cell line PFSK with stable miR-520g expression, however, reporter gene assays suggest that miR-520g does not significantly alter canonical WNT signaling in PFSK cells. Although the canonical WNT pathway has been most often highlighted in brain tumor development, our collective studies suggest that WNT5A associated non-canonical WNT/JNK2 signaling, which is known to regulate cell polarity and neuronal maturation (Ciani and Salinas, 2005), is upregulated by miR-520g expression in hNSC and PFSK cells. The role of non-canonical WNT signaling in neural stem cells remains to be completely elucidated however; our observations lead us to speculate that oncogenic miR-520g, and possibly other related C19MC miRNA, promotes transformation of neural stem/progenitor cell pools by subverting WNT signaling.

While our aggregate data links miR-520g with altered WNT signaling, the mechanism(s) by which miR-520g mediates this effect remains to be determined. An analysis of miR-520g associated gene sets in PFSK and hNSCs revealed few genes that represented strong direct miR-520g targets – based on criteria of target site prediction by at least 2 of 3 target prediction programs and modest to strong conservation of target sites (Watanabe et al., 2007). However, a comparative analysis of 2460 miR-520g target genes predicted by 3 algorithms (Table S9) reveal a common set of 63 candidate genes, a subset of which (TLE4, PRICKLE 2, BMP6, CAMK2N1, CUL3, LHX8, MAPRE1/EB1), represent promising candidates as they are implicated in WNT signaling and also exhibit one or more conserved miR-520g binding sites. Our studies to date reveal only limited gene expression changes of most candidate targets in miR-520g expressing stable hNSC and PFSK cell lines, therefore high-through put reporter gene assays and proteomic analysis are underway to validate these predicted miR-520g targets.

It is important to point out that the three algorithms used for identification of candidate miR-520g targets in our study rely largely on sequence conservation. Although use of evolutionary conservation has represented a key methodological advance for target gene identification of highly conserved miRNAs, the importance of such parameters for defining targets of poorly conserved, primate specific miRNAs, such as those of the C19MC locus remain unknown (Bartel, 2009). Significantly, functionally validated targets without matched seed pairing has been reported for a number of miRNAs including LIN28 (Wu and Belasco, 2005) and miR-24 (Lal et al., 2009). Targeting of “seedless” 3′UTR may represent a particularly important mechanism for C19MC miRNAs as they are characterized by poorly defined seed sequence, thus conventional seed-matched algorithms as used in this study may only identify a small subset of biologically relevant miR-520g targets.

Notably, many predicted miR-520g target genes also contain binding sites for other candidate C19MC miRNA oncogenes including miR-517c and miR-519a. As recent studies indicate that most miRNAs individually only have modest, modulatory effects on gene transcription or translation (Baek et al., 2008; Selbach et al., 2008), it is likely that miR-520g will require co-operativity with other non-coding or coding loci to effect robust changes in gene regulation and cell signaling. In this regard, it is interesting to note that in our study we observed frequent gains of chr2 in tumors with C19MC amplification, thus in addition to other C19MC miRNAs which are over expressed in CNS-PNET, we speculate that one or more loci on chr2 may also encode oncogenic partners of miR-520g.

Clinical implications

Current treatment of malignant pediatric brain tumors with chemo-radiotherapy engenders significant long-term neuro-cognitive sequelae, therefore there is substantial interest in applying molecular markers to optimize treatment. In addition to potential applications in tumor diagnosis, our data suggest the C19MC locus may also have prognostic role. A recent report described the C19MC amplicon in a case of an embryonal brain tumor characterized by ependymoblastic differentiation, and proposed association of C19MC with a specific nosologic entity (Pfister et al., 2008). However, our studies indicate a significant but not exclusive correlation of C19MC amplification with histology, pointing to a broader role for the C19MC locus in CNS-PNET pathogenesis. Specifically, a proportion of tumors that segregate with the C19MC amplified tumors lack characteristic histology or C19MC amplification, but exhibit increased expression of select C19MC miRNA oncogenes (Fig S7) and/or chr2 gains. Future investigations of alternative mechanisms of C19MC activation and/or other cooperating oncogenic loci suggested by our studies will be an important step towards delineating the scope of C19MC functions and how they contribute to the molecular pathogenesis of CNS-PNET.

Although the effects of treatment-related parameters could not be fully evaluated in our patient cohort, the collective data from this study provides a compelling basis for investigating the C19MC locus as a bio-marker in larger, homogenously treated cohorts of CNS-PNET. As our experimental data link WNT pathway dysregulation with oncogenic functions of the C19MC miRNAs, it will also be of interest to determine whether tumor phenotypes correlate with specific patterns of WNT pathway activity.

We have identified recurrent amplification leading to over-expression of the C19MC locus in a human cancer and demonstrate that at least 2 members of the cluster mediate oncogenic effects. Our data links the C19MC amplicon to a distinctly aggressive clinical phenotype and point to promising diagnostic and therapeutic applications of this locus for pediatric embryonal brain tumors. Together with recent reports which implicate C19MC miRNAs in aggressive primary breast cancer (Foekens et al., 2008), our findings suggest activation of the C19MC locus may generally define aggressive human tumors. Combined with recent success in blood-based detection (Mitchell et al., 2008) and in vivo targeting of oncogenic miRNAs (Krutzfeldt et al., 2005), our findings raise exciting possibilities to exploit C19MC as a diagnostic tool and therapeutic target for a broader spectrum of human cancers.


Tumor and nucleic acid samples

Tumor samples and clinical information were collected with consent as per protocols approved by the Hospital Research Ethics Board at participating institutions (listed in supplemental information). Clinicopathological data is listed in Table S1. Only tumors classified as CNS-PNET according to 2007 WHO CNS tumor classification criteria (McLendon RE, 2007) were included. DNA and RNA were extracted by standard methods; tumor histology was reviewed by CH or CE. Pediatric brain tumor tissue microarrays were previously described (Liang et al., 2008; Ray et al., 2004; Tabori et al., 2006)

Microarray and Fluorescence-in-situ hybridization (FISH) analysis

DNA and RNA array hybridizations were carried out by The Centre of Applied Genomics facility (TCAG), Hospital for Sick Children, Toronto (http://www.tcag.ca/) using the Affymetrix 500K/250K NspI and 250 StyI GeneChip Mapping arrays and the Illumina Human-6 V2 BeadChip array (http://www.illumina.com) respectively. Hybridization methods, data collection and analysis are detailed in Supplemental Information. Dual-colored FISH analyses were performed using pre-labeled BAC probes as detailed in Supplemental Information.

Accession numbers

Microarray data from this study are available at NCBI Gene Expression Omnibus (htt://www.ncbi.nlm.nih.gov/geo/) under accession numbers GSE14296 and GSE16988.

Immunoblot assays and quantitative real-time PCR (qPCR) analysis for DNA and mRNA

In Supplemental Information.

MicroRNA expression analysis with TaqMan® MicroRNA arrays and qRT-PCR

In Supplemental Information.

Stable microRNA expression

To generate miR 517c and 520g expression plasmids, synthetic oligonucleotides (Integrated DNA Technologies) containing mature miR-517c or miR-520g and 25-40bp of flanking precursor miRNAs sequences (Accession MI0003166 and M10003174 respectively from http://microrna.sanger.ac.uk/sequences) were cloned into a pcDNA3 vector (Invitrogen). Plasmid sequences are available on request. Stable cell lines were generated by transfecting Daoy, NIH3T3 and PFSK cells with pcDNA-miR-520g, miR-517c or empty vector, cells selected in G418 (2.5 μg/ml) were tested for miRNA expression by qRT-PCR.

Cell line propagation, growth, survival and transformation assays

In Supplemental Information.

Nude Mice Xenograft Assays

Protocols for animal experimentation were approved by the Hospital for Sick Children Animal Care Committee and are described in Supplemental Information.

Human neural stem cell line propagation, growth and differentiation assays

An adherent human neural stem (hNSC) cell line was generated from a 50 day old fetal cortex with consent obtained under a Hospital for Sick Children Research Ethics Board approved protocol and characterized according to published methods (Sun et al., 2008). The hNSC line was propagated in RHB-A media (Stem Cell Sciences Ltd) with recombinant murine EGF and human FGF-2 (10ng/ml final concentration; Peprotech). Stable hNSC lines were generated by nucleofection (Amaxa Biosystems) with pcDNA-miR-520 and 517c plasmids and selection in G418 containing media for 14 days. For growth assays, stable hNSCs-miR-520g and 517c cell lines were plated in 96-well dishes at 1000 cells/well, allowed to attach overnight under proliferative conditions and then switched to DMEM +10% FBS and monitored for growth using an InCyte imaging system (Essen Instruments) or MTT assay as described in Supplemental Information. Expression of differentiation markers in stable miR-520g and 517c expressing hNSCs was determined using FACS analysis with anti-Tubulin β-III (TuJ1 antibodies, 1:500, Chemicon), anti-GFAP (1:300, monoclonal GA-5, Sigma monoclonal) and Alexa 405 or 488 anti-species antibodies (Invitrogen) as described previously (Pollard et al., 2009).

Statistical analyses

ANOVA method was used for time series cell growth and survival experiments. Significance of soft agar assays and correlation of gene expression with chr19q13.41 amplification was determined using unpaired, 2-tailed Student’s t-tests; p-values were adjusted for multiple testing by FDR method for comparison of tumor or cell line gene expression profiles. A Wilcoxon rank sum test was used to determine significance of xenograft studies. Correlation of amplicon to survival was assessed using the Kaplan Meier method and log-rank test (α=0.05) and to age and histology by a two-sided Fisher’s exact test.


Lack of insight into disease mechanisms impedes development of effective therapies for many human neoplasm including childhood brain tumors such as CNS-PNET. Through global genomic studies, we discovered a frequent high-level amplicon of a polycistronic miRNA locus (C19MC) which defines aggressive CNS-PNET, and provide functional evidence that C19MC miRNAs act as oncogenes. Interestingly, although C19MC amplification remains to be reported in other cancers, expression of C19MC miRNAs has also been linked to unfavorable breast cancer phenotypes. Together with our discovery of recurrent C19MC gene amplification in a highly malignant brain tumor, these data highlight C19MC as an attractive candidate biomarker and therapeutic target for aggressive human cancers and support an emerging theme that miRNAs are important determinants of tumor biology.

Supplementary Material

Li Et Supplementry


P.Burger, J. Squire, L. Penn for mentorship; M. Zielenska, S. Chilton, P. Northcott, P. Kongkam, T. Paton and G. Casallo for generous help; J. Dick, S. Egan, CC Hui, E Lechman, R. Gilbertson for helpful discussions; P Hu for statistical analysis. Funding from CBTF, SickKids (Brainchild) and JM Foundations to AH, and AIRC to PM and tissue contributions from CHTN, NINDs and the London Regional Brain Tumor Bank are gratefully acknowledged. ML was an ABTA Fellow; KFL and DS received SickKids foundation RESTRACOMP pre-doctoral awards.


AUTHOR CONTRIBUTIONS ML, KFL, YL and D.S performed all experiments and analyses with support from DP, LZ and PB. NSC studies were performed by IC; EB, PM, ML, PC, MC, TV, SP, CE and CL provided tumor tissue and clinical data. A.H supervised the project. A.H, MDT, SS, JTR, CEH and PBD contributed to manuscript writing and review.


  • Allen JC, Judkins AR, Rosenblum MK, Biegel JA. Atypical teratoid/rhabdoid tumor evolving from an optic pathway ganglioglioma: case study. Neuro Oncol. 2006;8:79–82. [PMC free article] [PubMed]
  • Baek D, Villen J, Shin C, Camargo FD, Gygi SP, Bartel DP. The impact of microRNAs on protein output. Nature. 2008;455:64–71. [PMC free article] [PubMed]
  • Bar M, Wyman SK, Fritz BR, Qi J, Garg KS, Parkin RK, Kroh EM, Bendoraite A, Mitchell PS, Nelson AM, et al. MicroRNA discovery and profiling in human embryonic stem cells by deep sequencing of small RNA libraries. Stem Cells. 2008;26:2496–2505. [PMC free article] [PubMed]
  • Bartel DP. MicroRNAs: target recognition and regulatory functions. Cell. 2009;136:215–233. [PMC free article] [PubMed]
  • Beissbarth T, Speed TP. GOstat: find statistically overrepresented Gene Ontologies within a group of genes. Bioinformatics. 2004;20:1464–1465. [PubMed]
  • Bentwich I, Avniel A, Karov Y, Aharonov R, Gilad S, Barad O, Barzilai A, Einat P, Einav U, Meiri E, et al. Identification of hundreds of conserved and nonconserved human microRNAs. Nat Genet. 2005;37:766–770. [PubMed]
  • Bonci D, Coppola V, Musumeci M, Addario A, Giuffrida R, Memeo L, D’Urso L, Pagliuca A, Biffoni M, Labbaye C, et al. The miR-15a-miR-16-1 cluster controls prostate cancer by targeting multiple oncogenic activities. Nat Med. 2008;14:1271–1277. [PubMed]
  • Briscoe J, Pierani A, Jessell TM, Ericson J. A homeodomain protein code specifies progenitor cell identity and neuronal fate in the ventral neural tube. Cell. 2000;101:435–445. [PubMed]
  • Burger PC. Supratentorial primitive neuroectodermal tumor (sPNET) Brain Pathol. 2006;16:86. [PubMed]
  • Calin GA, Dumitru CD, Shimizu M, Bichi R, Zupo S, Noch E, Aldler H, Rattan S, Keating M, Rai K, et al. Frequent deletions and down-regulation of micro-RNA genes miR15 and miR16 at 13q14 in chronic lymphocytic leukemia. Proc Natl Acad Sci U S A. 2002;99:15524–15529. [PMC free article] [PubMed]
  • Ciani L, Salinas PC. WNTs in the vertebrate nervous system: from patterning to neuronal connectivity. Nat Rev Neurosci. 2005;6:351–362. [PubMed]
  • Esquela-Kerscher A, Slack FJ. Oncomirs - microRNAs with a role in cancer. Nat Rev Cancer. 2006;6:259–269. [PubMed]
  • Fangusaro J, Finlay J, Sposto R, Ji L, Saly M, Zacharoulis S, Asgharzadeh S, Abromowitch M, Olshefski R, Halpern S, et al. Intensive chemotherapy followed by consolidative myeloablative chemotherapy with autologous hematopoietic cell rescue (AuHCR) in young children with newly diagnosed supratentorial primitive neuroectodermal tumors (sPNETs): report of the Head Start I and II experience. Pediatr Blood Cancer. 2008;50:312–318. [PubMed]
  • Foekens JA, Sieuwerts AM, Smid M, Look MP, de Weerd V, Boersma AW, Klijn JG, Wiemer EA, Martens JW. Four miRNAs associated with aggressiveness of lymph node-negative, estrogen receptor-positive human breast cancer. Proc Natl Acad Sci U S A. 2008;105:13021–13026. [PMC free article] [PubMed]
  • Fogarty MP, Kessler JD, Wechsler-Reya RJ. Morphing into cancer: the role of developmental signaling pathways in brain tumor formation. J Neurobiol. 2005;64:458–475. [PubMed]
  • Gaidatzis D, van Nimwegen E, Hausser J, Zavolan M. Inference of miRNA targets using evolutionary conservation and pathway analysis. BMC Bioinformatics. 2007;8:69. [PMC free article] [PubMed]
  • Gessi M, Giangaspero F, Lauriola L, Gardiman M, Scheithauer BW, Halliday W, Hawkins C, Rosenblum MK, Burger PC, Eberhart CG. Embryonal Tumors With Abundant Neuropil and True Rosettes: A Distinctive CNS Primitive Neuroectodermal Tumor. Am J Surg Pathol. 2008 [PubMed]
  • He L, Thomson JM, Hemann MT, Hernando-Monge E, Mu D, Goodson S, Powers S, Cordon-Cardo C, Lowe SW, Hannon GJ, Hammond SM. A microRNA polycistron as a potential human oncogene. Nature. 2005;435:828–833. [PubMed]
  • Iafrate AJ, Feuk L, Rivera MN, Listewnik ML, Donahoe PK, Qi Y, Scherer SW, Lee C. Detection of large-scale variation in the human genome. Nat Genet. 2004;36:949–951. [PubMed]
  • Inda MM, Perot C, Guillaud-Bataille M, Danglot G, Rey JA, Bello MJ, Fan X, Eberhart C, Zazpe I, Portillo E, et al. Genetic heterogeneity in supratentorial and infratentorial primitive neuroectodermal tumours of the central nervous system. Histopathology. 2005;47:631–637. [PubMed]
  • Johnson SM, Grosshans H, Shingara J, Byrom M, Jarvis R, Cheng A, Labourier E, Reinert KL, Brown D, Slack FJ. RAS is regulated by the let-7 microRNA family. Cell. 2005;120:635–647. [PubMed]
  • Judkins AR, Ellison DW. Ependymoblastoma: Dear, Damned, Distracting Diagnosis, Farewell!*. Brain Pathol. 2008 [PubMed]
  • Kalani MY, Cheshier SH, Cord BJ, Bababeygy SR, Vogel H, Weissman IL, Palmer TD, Nusse R. Wnt-mediated self-renewal of neural stem/progenitor cells. Proc Natl Acad Sci U S A. 2008;105:16970–16975. [PMC free article] [PubMed]
  • Krutzfeldt J, Rajewsky N, Braich R, Rajeev KG, Tuschl T, Manoharan M, Stoffel M. Silencing of microRNAs in vivo with ’antagomirs’ Nature. 2005;438:685–689. [PubMed]
  • Lal A, Navarro F, Maher CA, Maliszewski LE, Yan N, O’Day E, Chowdhury D, Dykxhoorn DM, Tsai P, Hofmann O, et al. miR-24 Inhibits cell proliferation by targeting E2F2, MYC, and other cell-cycle genes via binding to “seedless” 3′UTR microRNA recognition elements. Mol Cell. 2009;35:610–625. [PMC free article] [PubMed]
  • Laurent LC, Chen J, Ulitsky I, Mueller FJ, Lu C, Shamir R, Fan JB, Loring JF. Comprehensive microRNA profiling reveals a unique human embryonic stem cell signature dominated by a single seed sequence. Stem Cells. 2008;26:1506–1516. [PubMed]
  • Liang ML, Ma J, Ho M, Solomon L, Bouffet E, Rutka JT, Hawkins C. Tyrosine kinase expression in pediatric high grade astrocytoma. J Neurooncol. 2008;87:247–253. [PubMed]
  • Lin M, Wei LJ, Sellers WR, Lieberfarb M, Wong WH, Li C. dChipSNP: significance curve and clustering of SNP-array-based loss-of-heterozygosity data. Bioinformatics. 2004;20:1233–1240. [PubMed]
  • Lu J, Getz G, Miska EA, Alvarez-Saavedra E, Lamb J, Peck D, Sweet-Cordero A, Ebert BL, Mak RH, Ferrando AA, et al. MicroRNA expression profiles classify human cancers. Nature. 2005;435:834–838. [PubMed]
  • McCabe MG, Ichimura K, Liu L, Plant K, Backlund LM, Pearson DM, Collins VP. High-resolution array-based comparative genomic hybridization of medulloblastomas and supratentorial primitive neuroectodermal tumors. J Neuropathol Exp Neurol. 2006;65:549–561. [PMC free article] [PubMed]
  • McLendon RE JA, Eberhart CG, et al. WHO Classification of Tumors of the Central Nervous System. IARC Press (IARC Press); Lyon: 2007. Central nervous system Primitive Neuroectodermal Tumors (PNET) pp. 141–146.
  • Mikels AJ, Nusse R. Purified Wnt5a protein activates or inhibits beta-catenin-TCF signaling depending on receptor context. PLoS Biol. 2006;4:e115. [PMC free article] [PubMed]
  • Miller JR. The Wnts. Genome Biol. 2002;3 REVIEWS3001. [PMC free article] [PubMed]
  • Mitchell PS, Parkin RK, Kroh EM, Fritz BR, Wyman SK, Pogosova-Agadjanyan EL, Peterson A, Noteboom J, O’Briant KC, Allen A, et al. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci U S A. 2008;105:10513–10518. [PMC free article] [PubMed]
  • Nemeth MJ, Topol L, Anderson SM, Yang Y, Bodine DM. Wnt5a inhibits canonical Wnt signaling in hematopoietic stem cells and enhances repopulation. Proc Natl Acad Sci U S A. 2007;104:15436–15441. [PMC free article] [PubMed]
  • Packer RJ, Gajjar A, Vezina G, Rorke-Adams L, Burger PC, Robertson PL, Bayer L, LaFond D, Donahue BR, Marymont MH, et al. Phase III study of craniospinal radiation therapy followed by adjuvant chemotherapy for newly diagnosed average-risk medulloblastoma. J Clin Oncol. 2006;24:4202–4208. [PubMed]
  • Papagiannakopoulos T, Shapiro A, Kosik KS. MicroRNA-21 targets a network of key tumor-suppressive pathways in glioblastoma cells. Cancer Res. 2008;68:8164–8172. [PubMed]
  • Pfister S, Remke M, Castoldi M, Bai AH, Muckenthaler MU, Kulozik A, von Deimling A, Pscherer A, Lichter P, Korshunov A. Novel genomic amplification targeting the microRNA cluster at 19q13.42 in a pediatric embryonal tumor with abundant neuropil and true rosettes. Acta Neuropathol. 2008 [PubMed]
  • Pfister S, Remke M, Toedt G, Werft W, Benner A, Mendrzyk F, Wittmann A, Devens F, von Hoff K, Rutkowski S, et al. Supratentorial primitive neuroectodermal tumors of the central nervous system frequently harbor deletions of the CDKN2A locus and other genomic aberrations distinct from medulloblastomas. Genes Chromosomes Cancer. 2007;46:839–851. [PubMed]
  • Pollard SM, Yoshikawa K, Clarke ID, Danovi D, Stricker S, Russell R, Bayani J, Head R, Lee M, Bernstein M, et al. Glioma stem cell lines expanded in adherent culture have tumor-specific phenotypes and are suitable for chemical and genetic screens. Cell Stem Cell. 2009;4:568–580. [PubMed]
  • Pomeroy SL, Tamayo P, Gaasenbeek M, Sturla LM, Angelo M, McLaughlin ME, Kim JY, Goumnerova LC, Black PM, Lau C, et al. Prediction of central nervous system embryonal tumour outcome based on gene expression. Nature. 2002;415:436–442. [PubMed]
  • Ray A, Ho M, Ma J, Parkes RK, Mainprize TG, Ueda S, McLaughlin J, Bouffet E, Rutka JT, Hawkins CE. A clinicobiological model predicting survival in medulloblastoma. Clin Cancer Res. 2004;10:7613–7620. [PubMed]
  • Reya T, Clevers H. Wnt signalling in stem cells and cancer. Nature. 2005;434:843–850. [PubMed]
  • Rogers HA, Miller S, Lowe J, Brundler MA, Coyle B, Grundy RG. An investigation of WNT pathway activation and association with survival in central nervous system primitive neuroectodermal tumours (CNS PNET) Br J Cancer. 2009;100:1292–1302. [PMC free article] [PubMed]
  • Selbach M, Schwanhausser B, Thierfelder N, Fang Z, Khanin R, Rajewsky N. Widespread changes in protein synthesis induced by microRNAs. Nature. 2008;455:58–63. [PubMed]
  • Sun Y, Pollard S, Conti L, Toselli M, Biella G, Parkin G, Willatt L, Falk A, Cattaneo E, Smith A. Long-term tripotent differentiation capacity of human neural stem (NS) cells in adherent culture. Mol Cell Neurosci. 2008;38:245–258. [PubMed]
  • Tabori U, Ma J, Carter M, Zielenska M, Rutka J, Bouffet E, Bartels U, Malkin D, Hawkins C. Human telomere reverse transcriptase expression predicts progression and survival in pediatric intracranial ependymoma. J Clin Oncol. 2006;24:1522–1528. [PubMed]
  • Timmermann B, Kortmann RD, Kuhl J, Rutkowski S, Meisner C, Pietsch T, Deinlein F, Urban C, Warmuth-Metz M, Bamberg M. Role of radiotherapy in supratentorial primitive neuroectodermal tumor in young children: results of the German HIT-SKK87 and HIT-SKK92 trials. J Clin Oncol. 2006;24:1554–1560. [PubMed]
  • Voorhoeve PM, le Sage C, Schrier M, Gillis AJ, Stoop H, Nagel R, Liu YP, van Duijse J, Drost J, Griekspoor A, et al. A genetic screen implicates miRNA-372 and miRNA-373 as oncogenes in testicular germ cell tumors. Cell. 2006;124:1169–1181. [PubMed]
  • Watanabe Y, Tomita M, Kanai A. Computational methods for microRNA target prediction. Methods Enzymol. 2007;427:65–86. [PubMed]
  • Wexler EM, Paucer A, Kornblum HI, Plamer TD, Geschwind DH. Endogenous Wnt signaling maintains neural progenitor cell potency. Stem Cells. 2009;27:1130–1141. [PMC free article] [PubMed]
  • Wu L, Belasco JG. Micro-RNA regulation of the mammalian lin-28 gene during neuronal differentiation of embryonal carcinoma cells. Mol Cell Biol. 2005;25:9198–9208. [PMC free article] [PubMed]
  • Zhou Q, Anderson DJ. The bHLH transcription factors OLIG2 and OLIG1 couple neuronal and glial subtype specification. Cell. 2002;109:61–73. [PubMed]
  • Zindy F, Nilsson LM, Nguyen L, Meunier C, Smeyne RJ, Rehg JE, Eberhart C, Sherr CJ, Roussel MF. Hemangiosarcomas, medulloblastomas, and other tumors in Ink4c/p53-null mice. Cancer Res. 2003;63:5420–5427. [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...