We have extended our understanding of the molecular biology underlying adult glioblastoma over many years. In contrast, high-grade gliomas in children and adolescents have remained a relatively under-investigated disease. The latest large-scale genomic and epigenomic profiling studies have yielded an unprecedented abundance of novel data and revealed deeper insights into gliomagenesis across all age groups, highlighting key distinctions, but also some commonalities. As we are on the verge of dissecting glioblastomas into meaningful biological subgroups, this Review summarizes the hallmark genetic alterations associated with distinct epigenetic features and patient characteristics in both paediatric and adult disease, and examines the complex interplay between the glioblastoma genome and epigenome.
Pilocytic astrocytoma, the most common childhood brain tumor1, is typically associated with mitogen-activated protein kinase (MAPK) pathway alterations2. Surgically inaccessible midline tumors are therapeutically challenging, showing sustained tendency for progression3 and often becoming a chronic disease with substantial morbidities4.
Here we describe whole-genome sequencing of 96 pilocytic astrocytomas, with matched RNA sequencing (n=73), conducted by the International Cancer Genome Consortium (ICGC) PedBrain Tumor Project. We identified recurrent activating mutations in FGFR1 and PTPN11 and novel NTRK2 fusion genes in non-cerebellar tumors. New BRAF activating changes were also observed. MAPK pathway alterations affected 100% of tumors analyzed, with no other significant mutations, indicating pilocytic astrocytoma as predominantly a single-pathway disease.
Notably, we identified the same FGFR1 mutations in a subset of H3F3A-mutated pediatric glioblastoma with additional alterations in NF15. Our findings thus identify new potential therapeutic targets in distinct subsets of pilocytic astrocytoma and childhood glioblastoma.
All cancers are caused by somatic mutations. However, understanding of the biological processes generating these mutations is limited. The catalogue of somatic mutations from a cancer genome bears the signatures of the mutational processes that have been operative. Here, we analysed 4,938,362 mutations from 7,042 cancers and extracted more than 20 distinct mutational signatures. Some are present in many cancer types, notably a signature attributed to the APOBEC family of cytidine deaminases, whereas others are confined to a single class. Certain signatures are associated with age of the patient at cancer diagnosis, known mutagenic exposures or defects in DNA maintenance, but many are of cryptic origin. In addition to these genome-wide mutational signatures, hypermutation localized to small genomic regions, kataegis, is found in many cancer types. The results reveal the diversity of mutational processes underlying the development of cancer with potential implications for understanding of cancer etiology, prevention and therapy.
Subgrouping of medulloblastoma by microarray expression profiling has dramatically changed our perspective of this malignant childhood brain tumour. Now, the availability of next-generation sequencing and complementary high-density genomic technologies has unmasked novel driver mutations in each medulloblastoma subgroup. The implications of these findings for the management of patients are readily apparent, pinpointing previously unappreciated diagnostic and therapeutic targets. Here, we summarize the ’explosion’ of data emerging from the application of modern genomics to medulloblastoma, and in particular the recurrent targets of mutation in medulloblastoma subgroups. These data are making their way into contemporary clinical trials as we seek to integrate conventional and molecularly targeted therapies.
High-grade soft tissue sarcomas are a heterogeneous, complex group of aggressive malignant tumors showing mesenchymal differentiation. Recently, soft tissue sarcomas have increasingly been classified on the basis of underlying genetic alterations; however, the role of aberrant DNA methylation in these tumors is not well understood and, consequently, the usefulness of methylation-based classification is unclear.
We used the Infinium HumanMethylation27 platform to profile DNA methylation in 80 primary, untreated high-grade soft tissue sarcomas, representing eight relevant subtypes, two non-neoplastic fat samples and 14 representative sarcoma cell lines. The primary samples were partitioned into seven stable clusters. A classification algorithm identified 216 CpG sites, mapping to 246 genes, showing different degrees of DNA methylation between these seven groups. The differences between the clusters were best represented by a set of eight CpG sites located in the genes SPEG, NNAT, FBLN2, PYROXD2, ZNF217, COL14A1, DMRT2 and CDKN2A. By integrating DNA methylation and mRNA expression data, we identified 27 genes showing negative and three genes showing positive correlation. Compared with non-neoplastic fat, NNAT showed DNA hypomethylation and inverse gene expression in myxoid liposarcomas, and DNA hypermethylation and inverse gene expression in dedifferentiated and pleomorphic liposarcomas. Recovery of NNAT in a hypermethylated myxoid liposarcoma cell line decreased cell migration and viability.
Our analysis represents the first comprehensive integration of DNA methylation and transcriptional data in primary high-grade soft tissue sarcomas. We propose novel biomarkers and genes relevant for pathogenesis, including NNAT as a potential tumor suppressor in myxoid liposarcomas.
Medulloblastoma is curable in approximately 70 % of patients. Over the past decade, progress in improving survival using conventional therapies has stalled, resulting in reduced quality of life due to treatment-related side effects, which are a major concern in survivors. The vast amount of genomic and molecular data generated over the last 5–10 years encourages optimism that improved risk stratification and new molecular targets will improve outcomes. It is now clear that medulloblastoma is not a single-disease entity, but instead consists of at least four distinct molecular subgroups: WNT/Wingless, Sonic Hedgehog, Group 3, and Group 4. The Medulloblastoma Down Under 2013 meeting, which convened at Bunker Bay, Australia, brought together 50 leading clinicians and scientists. The 2-day agenda included focused sessions on pathology and molecular stratification, genomics and mouse models, high-throughput drug screening, and clinical trial design. The meeting established a global action plan to translate novel biologic insights and drug targeting into treatment regimens to improve outcomes. A consensus was reached in several key areas, with the most important being that a novel classification scheme for medulloblastoma based on the four molecular subgroups, as well as histopathologic features, should be presented for consideration in the upcoming fifth edition of the World Health Organization’s classification of tumours of the central nervous system. Three other notable areas of agreement were as follows: (1) to establish a central repository of annotated mouse models that are readily accessible and freely available to the international research community; (2) to institute common eligibility criteria between the Children’s Oncology Group and the International Society of Paediatric Oncology Europe and initiate joint or parallel clinical trials; (3) to share preliminary high-throughput screening data across discovery labs to hasten the development of novel therapeutics. Medulloblastoma Down Under 2013 was an effective forum for meaningful discussion, which resulted in enhancing international collaborative clinical and translational research of this rare disease. This template could be applied to other fields to devise global action plans addressing all aspects of a disease, from improved disease classification, treatment stratification, and drug targeting to superior treatment regimens to be assessed in cooperative international clinical trials.
Mutation is a fundamental process in tumorigenesis. However, the degree to which the rate of somatic mutation varies across the human genome and the mechanistic basis underlying this variation remain to be fully elucidated. Here, we performed a cross-cancer comparison of 402 whole genomes comprising a diverse set of childhood and adult tumors, including both solid and hematopoietic malignancies. Surprisingly, we found that the inactive X chromosome of many female cancer genomes accumulates on average twice and up to four times as many somatic mutations per megabase, as compared to the individual autosomes. Whole-genome sequencing of clonally expanded hematopoietic stem/progenitor cells (HSPCs) from healthy individuals and a premalignant myelodysplastic syndrome (MDS) sample revealed no X chromosome hypermutation. Our data suggest that hypermutation of the inactive X chromosome is an early and frequent feature of tumorigenesis resulting from DNA replication stress in aberrantly proliferating cells.
•X chromosome has up to 4× more mutations than the autosomes in female cancer genomes•Hypermutations only affect the inactive X chromosome•X hypermutation involves somatic point mutations and indels, but not germline mutations•No X hypermutation is found in clonal expansions of normal or premalignant cells
A comparison of 402 cancer genomes identifies a surprisingly high level of somatic mutations in the inactive X chromosome of female cancer genomes. As hypermutability of the inactive X was not observed in clonal hematopoietic progenitor or preleukemic samples, it is likely that it may be a contributing factor to tumorigenesis.
The emergence of high-throughput, next-generation sequencing technologies has dramatically altered the way we assess genomes in population genetics and in cancer genomics. Currently, there are four commonly used whole-genome sequencing platforms on the market: Illumina’s HiSeq2000, Life Technologies’ SOLiD 4 and its completely redesigned 5500xl SOLiD, and Complete Genomics’ technology. A number of earlier studies have compared a subset of those sequencing platforms or compared those platforms with Sanger sequencing, which is prohibitively expensive for whole genome studies. Here we present a detailed comparison of the performance of all currently available whole genome sequencing platforms, especially regarding their ability to call SNVs and to evenly cover the genome and specific genomic regions. Unlike earlier studies, we base our comparison on four different samples, allowing us to assess the between-sample variation of the platforms. We find a pronounced GC bias in GC-rich regions for Life Technologies’ platforms, with Complete Genomics performing best here, while we see the least bias in GC-poor regions for HiSeq2000 and 5500xl. HiSeq2000 gives the most uniform coverage and displays the least sample-to-sample variation. In contrast, Complete Genomics exhibits by far the smallest fraction of bases not covered, while the SOLiD platforms reveal remarkable shortcomings, especially in covering CpG islands. When comparing the performance of the four platforms for calling SNPs, HiSeq2000 and Complete Genomics achieve the highest sensitivity, while the SOLiD platforms show the lowest false positive rate. Finally, we find that integrating sequencing data from different platforms offers the potential to combine the strengths of different technologies. In summary, our results detail the strengths and weaknesses of all four whole-genome sequencing platforms. It indicates application areas that call for a specific sequencing platform and disallow other platforms. This helps to identify the proper sequencing platform for whole genome studies with different application scopes.
Medulloblastoma is an aggressively-growing tumour, arising in the cerebellum or medulla/brain stem. It is the most common malignant brain tumour in children, and displays tremendous biological and clinical heterogeneity1. Despite recent treatment advances, approximately 40% of children experience tumour recurrence, and 30% will die from their disease. Those who survive often have a significantly reduced quality of life.
Four tumour subgroups with distinct clinical, biological and genetic profiles are currently discriminated2,3. WNT tumours, displaying activated wingless pathway signalling, carry a favourable prognosis under current treatment regimens4. SHH tumours show hedgehog pathway activation, and have an intermediate prognosis2. Group 3 & 4 tumours are molecularly less well-characterised, and also present the greatest clinical challenges2,3,5. The full repertoire of genetic events driving this distinction, however, remains unclear.
Here we describe an integrative deep-sequencing analysis of 125 tumour-normal pairs. Tetraploidy was identified as a frequent early event in Group 3 & 4 tumours, and a positive correlation between patient age and mutation rate was observed. Several recurrent mutations were identified, both in known medulloblastoma-related genes (CTNNB1, PTCH1, MLL2, SMARCA4) and in genes not previously linked to this tumour (DDX3X, CTDNEP1, KDM6A, TBR1), often in subgroup-specific patterns. RNA-sequencing confirmed these alterations, and revealed the expression of the first medulloblastoma fusion genes. Chromatin modifiers were frequently altered across all subgroups.
These findings enhance our understanding of the genomic complexity and heterogeneity underlying medulloblastoma, and provide several potential targets for new therapeutics, especially for Group 3 & 4 patients.
Non-coding RNAs are much more common than previously thought. However, for the vast majority of non-coding RNAs, the cellular function remains enigmatic. The two long non-coding RNA (lncRNA) genes DLEU1 and DLEU2 map to a critical region at chromosomal band 13q14.3 that is recurrently deleted in solid tumors and hematopoietic malignancies like chronic lymphocytic leukemia (CLL). While no point mutations have been found in the protein coding candidate genes at 13q14.3, they are deregulated in malignant cells, suggesting an epigenetic tumor suppressor mechanism. We therefore characterized the epigenetic makeup of 13q14.3 in CLL cells and found histone modifications by chromatin-immunoprecipitation (ChIP) that are associated with activated transcription and significant DNA-demethylation at the transcriptional start sites of DLEU1 and DLEU2 using 5 different semi-quantitative and quantitative methods (aPRIMES, BioCOBRA, MCIp, MassARRAY, and bisulfite sequencing). These epigenetic aberrations were correlated with transcriptional deregulation of the neighboring candidate tumor suppressor genes, suggesting a coregulation in cis of this gene cluster. We found that the 13q14.3 genes in addition to their previously known functions regulate NF-kB activity, which we could show after overexpression, siRNA–mediated knockdown, and dominant-negative mutant genes by using Western blots with previously undescribed antibodies, by a customized ELISA as well as by reporter assays. In addition, we performed an unbiased screen of 810 human miRNAs and identified the miR-15/16 family of genes at 13q14.3 as the strongest inducers of NF-kB activity. In summary, the tumor suppressor mechanism at 13q14.3 is a cluster of genes controlled by two lncRNA genes that are regulated by DNA-methylation and histone modifications and whose members all regulate NF-kB. Therefore, the tumor suppressor mechanism in 13q14.3 underlines the role both of epigenetic aberrations and of lncRNA genes in human tumorigenesis and is an example of colocalization of a functionally related gene cluster.
Recent results suggest that genome regions not coding for proteins are read and transcribed into RNA. While the function for the majority of the resulting non-coding RNA molecules remains unclear, some of them are termed according to their length (typically 200–2,000 nucleotides) as long non-coding RNA (lncRNA) genes that play a role in regulating the activity of target genes. In most instances, this deregulation involves changes of so-called “epigenetic” marks associated with the DNA that are inherited to the cellular progeny without changes in the DNA sequence. Here we describe an example where two lncRNA genes (DLEU1 and DLEU2) are epigenetically deregulated together with a cluster of neighboring protein-coding tumor suppressor genes in almost all patients suffering from chronic lymphocytic leukemia. Such a common regulation suggests that the affected genes are involved in the same cellular pathway. In line with this notion, the 13q14.3 genes modulate the NF-kB signalling pathway, either inducing or repressing its activity. An activation of NF-kB has previously been shown to promote survival of the leukemic cells, underlining the importance of the 13q14.3 tumor suppressor locus for the pathomechanism of the disease.
Recently, we found strong overexpression of the mucin-type glycoprotein podoplanin (PDPN) in human astrocytic brain tumors, specifically in primary glioblastoma multiforme (GB). In the current study, we show an inverse correlation between PDPN expression and PTEN levels in primary human GB and glioma cell lines, and we report elevated PDPN protein levels in the subventricular zone of brain tissue sections of PTEN-deficient mice. In human glioma cells lacking functional PTEN, reintroduction of wild-type PTEN, inhibition of the PTEN downstream target protein kinase B/AKT, or interference with transcription factor AP-1 function resulted in efficient downregulation of PDPN expression. In addition, we observed hypoxia-dependent PDPN transcriptional control and demonstrated that PDPN expression is subject to negative transcriptional regulation by promoter methylation in human GB and in glioma cell lines. Treatment of PTEN-negative glioma cells with demethylating agents induced expression of PDPN. Together, our findings show that increased PDPN expression in human GB is caused by loss of PTEN function and activation of the PI3K-AKT-AP-1 signaling pathway, accompanied by epigenetic regulation of PDPN promoter activity. Silencing of PDPN expression leads to reduced proliferation and migration of glioma cells, suggesting a functional role of PDPN in glioma progression and malignancy. Thus, specific targeting of PDPN expression and/or function could be a promising strategy for the treatment of patients with primary GB.
glioma; Jun; podoplanin; promoter methylation; PTEN
Medulloblastomas are the most common malignant brain tumors in children1. Identifying and understanding the genetic events that drive these tumors is critical for the development of more effective diagnostic, prognostic and therapeutic strategies. Recently, our group and others described distinct molecular subtypes of medulloblastoma based on transcriptional and copy number profiles2–5. Here, we utilized whole exome hybrid capture and deep sequencing to identify somatic mutations across the coding regions of 92 primary medulloblastoma/normal pairs. Overall, medulloblastomas exhibit low mutation rates consistent with other pediatric tumors, with a median of 0.35 non-silent mutations per megabase. We identified twelve genes mutated at statistically significant frequencies, including previously known mutated genes in medulloblastoma such as CTNNB1, PTCH1, MLL2, SMARCA4 and TP53. Recurrent somatic mutations were identified in an RNA helicase gene, DDX3X, often concurrent with CTNNB1 mutations, and in the nuclear co-repressor (N-CoR) complex genes GPS2, BCOR, and LDB1, novel findings in medulloblastoma. We show that mutant DDX3X potentiates transactivation of a TCF promoter and enhances cell viability in combination with mutant but not wild type beta-catenin. Together, our study reveals the alteration of Wnt, Hedgehog, histone methyltransferase and now N-CoR pathways across medulloblastomas and within specific subtypes of this disease, and nominates the RNA helicase DDX3X as a component of pathogenic beta-catenin signaling in medulloblastoma.
Genomic rearrangements are thought to occur progressively during tumor development. Recent findings, however, suggest an alternative mechanism, involving massive chromosome rearrangements in a one-step catastrophic event termed chromothripsis. We report the whole-genome sequencing-based analysis of a Sonic-Hedgehog medulloblastoma (SHH-MB) brain tumor from a patient with a germline TP53 mutation (Li-Fraumeni syndrome), uncovering massive, complex chromosome rearrangements. Integrating TP53 status with microarray and deep sequencing-based DNA rearrangement data in additional patients reveals a striking association between TP53 mutation and chromothripsis in SHH-MBs. Analysis of additional tumor entities substantiates a link between TP53 mutation and chromothripsis, and indicates a context-specific role for p53 in catastrophic DNA rearrangements. Among these, we observed a strong association between somatic TP53 mutations and chromothripsis in acute myeloid leukemia. These findings connect p53 status and chromothripsis in specific tumor types, providing a genetic basis for understanding particularly aggressive subtypes of cancer.
Embryonal tumor with multilayered rosettes (ETMR, previously known as ETANTR) is a highly aggressive embryonal CNS tumor, which almost exclusively affects infants and is associated with a dismal prognosis. Accurate diagnosis is of critical clinical importance because of its poor response to current treatment protocols and its distinct biology. Amplification of the miRNA cluster at 19q13.42 has been identified previously as a genetic hallmark for ETMR, but an immunohistochemistry-based assay for clinical routine diagnostics [such as INI-1 for atypical teratoid rhabdoid tumor (AT/RT)] is still lacking. In this study, we screened for an ETMR-specific marker using a gene-expression profiling dataset of more than 1,400 brain tumors and identified LIN28A as a highly specific marker for ETMR. The encoded protein binds small RNA and has been implicated in stem cell pluripotency, metabolism and tumorigenesis. Using an LIN28A specific antibody, we carried out immunohistochemical analysis of LIN28A in more than 800 childhood brain-tumor samples and confirmed its high specificity for ETMR. Strong LIN28A immunoexpression was found in all 37 ETMR samples tested, whereas focal reactivity was only present in a small (6/50) proportion of AT/RT samples. All other pediatric brain tumors were completely LIN28A-negative. In summary, we established LIN28A immunohistochemistry as a highly sensitive and specific, rapid, inexpensive diagnostic tool for routine pathological verification of ETMR.
ETMR; Pediatric brain tumor; LIN28A; Diagnostic marker
We identified the far upstream element binding protein 1 (FBP1), an activator of transcription of the proto-oncogene c-myc, in a functional yeast survival screen for tumor-related antiapoptotic proteins and demonstrated strong overexpression of FBP1 in human hepato-cellular carcinoma (HCC). Knockdown of the protein in HCC cells resulted in increased sensitivity to apoptotic stimuli, reduced cell proliferation, and impaired tumor formation in a mouse xenograft transplantation model. Interestingly, analysis of gene regulation in these cells revealed that c-myc levels were not influenced by FBP1 in HCC cells. Instead, we identified the cell cycle inhibitor p21 as a direct target gene repressed by FBP1, and in addition, expression levels of the proapoptotic genes tumor necrosis factor α, tumor necrosis factor–related apoptosis-inducing ligand, Noxa, and Bik were elevated in the absence of FBP1.
Our data establish FBP1 as an important oncoprotein overexpressed in HCC that induces tumor propagation through direct or indirect repression of cell cycle inhibitors and proapoptotic target genes.
Pediatric glioblastomas (GBM) including diffuse intrinsic pontine gliomas (DIPG) are devastating brain tumors with no effective therapy. Here, we investigated clinical and biological impacts of histone H3.3 mutations. Forty-two DIPGs were tested for H3.3 mutations. Wild-type versus mutated (K27M-H3.3) subgroups were compared for HIST1H3B, IDH, ATRX and TP53 mutations, copy number alterations and clinical outcome. K27M-H3.3 occurred in 71 %, TP53 mutations in 77 % and ATRX mutations in 9 % of DIPGs. ATRX mutations were more frequent in older children (p < 0.0001). No G34V/R-H3.3, IDH1/2 or H3.1 mutations were identified. K27M-H3.3 DIPGs showed specific copy number changes, including all gains/amplifications of PDGFRA and MYC/PVT1 loci. Notably, all long-term survivors were H3.3 wild type and this group of patients had better overall survival. K27M-H3.3 mutation defines clinically and biologically distinct subgroups and is prevalent in DIPG, which will impact future therapeutic trial design. K27M- and G34V-H3.3 have location-based incidence (brainstem/cortex) and potentially play distinct roles in pediatric GBM pathogenesis. K27M-H3.3 is universally associated with short survival in DIPG, while patients wild-type for H3.3 show improved survival. Based on prognostic and therapeutic implications, our findings argue for H3.3-mutation testing at diagnosis, which should be rapidly integrated into the clinical decision-making algorithm, particularly in atypical DIPG.
Electronic supplementary material
The online version of this article (doi:10.1007/s00401-012-0998-0) contains supplementary material, which is available to authorized users.
DIPG; H3.3; ATRX; TP53; Survival; Targeted therapy
Medulloblastoma is the most common malignant brain tumor in children. A subset of medulloblastoma originates from granule cell precursors (GCPs) of the developing cerebellum and demonstrates aberrant hedgehog signaling, typically due to inactivating mutations in the receptor PTCH1, a pathomechanism recapitulated in Ptch1+/− mice. As nitric oxide may regulate GCP proliferation and differentiation, we crossed Ptch1+/− mice with mice lacking inducible nitric oxide synthase (Nos2) to investigate a possible influence on tumorigenesis. We observed a two-fold higher medulloblastoma rate in Ptch1+/− Nos2−/− mice compared to Ptch1+/− Nos2+/+ mice. To identify the molecular mechanisms underlying this finding, we performed gene expression profiling of medulloblastomas from both genotypes, as well as normal cerebellar tissue samples of different developmental stages and genotypes. Downregulation of hedgehog target genes was observed in postnatal cerebellum from Ptch1+/+ Nos2−/− mice but not from Ptch1+/− Nos2−/− mice. The most consistent effect of Nos2 deficiency was downregulation of growth-associated protein 43 (Gap43). Functional studies in neuronal progenitor cells demonstrated nitric oxide dependence of Gap43 expression and impaired migration upon Gap43 knock-down. Both effects were confirmed in situ by immunofluorescence analyses on tissue sections of the developing cerebellum. Finally, the number of proliferating GCPs at the cerebellar periphery was decreased in Ptch1+/+ Nos2−/− mice but increased in Ptch1+/− Nos2−/− mice relative to Ptch1+/− Nos2+/+ mice. Taken together, these results indicate that Nos2 deficiency promotes medulloblastoma development in Ptch1+/− mice through retention of proliferating GCPs in the external granular layer due to reduced Gap43 expression. This study illustrates a new role of nitric oxide signaling in cerebellar development and demonstrates that the localization of pre-neoplastic cells during morphogenesis is crucial for their malignant progression.
Medulloblastoma is a common pediatric brain tumor, a subtype of which is driven by aberrant hedgehog pathway activation in cerebellar granule cell precursors. Although this tumor etiology has been intensively investigated in the well-established Ptch1+/− mouse model, knowledge is still lacking about the molecular interactions between neoplastic transformation and other developmental processes. Nitric oxide (NO) has been reported to be involved in controlling proliferation and differentiation of these cells. Therefore, inactivation of the NO–producing enzyme Nos2 in combination with the mutated Ptch1 gene should provide insight into how developmental regulation influences pathogenesis. Here, we describe a new role for NO in developing neuronal precursors of the cerebellum facilitating physiologically accurate migration via regulation of Gap43. We further demonstrate that disturbance of these processes leads to retention of granule precursor cells to the cerebellar periphery. Together with the sustained proliferation of these cells in combined Ptch1+/− Nos2−/− mice, this effect results in an increased medulloblastoma incidence relative to Ptch1+/− mice and demonstrates a new disease-promoting mechanism in this tumor entity.
Medulloblastoma is the most common malignant brain tumor in childhood. Molecular studies from several groups around the world demonstrated that medulloblastoma is not one disease but comprises a collection of distinct molecular subgroups. However, all these studies reported on different numbers of subgroups. The current consensus is that there are only four core subgroups, which should be termed WNT, SHH, Group 3 and Group 4. Based on this, we performed a meta-analysis of all molecular and clinical data of 550 medulloblastomas brought together from seven independent studies. All cases were analyzed by gene expression profiling and for most cases SNP or array-CGH data were available. Data are presented for all medulloblastomas together and for each subgroup separately. For validation purposes, we compared the results of this meta-analysis with another large medulloblastoma cohort (n = 402) for which subgroup information was obtained by immunohistochemistry. Results from both cohorts are highly similar and show how distinct the molecular subtypes are with respect to their transcriptome, DNA copy-number aberrations, demographics, and survival. Results from these analyses will form the basis for prospective multi-center studies and will have an impact on how the different subgroups of medulloblastoma will be treated in the future.
Electronic supplementary material
The online version of this article (doi:10.1007/s00401-012-0958-8) contains supplementary material, which is available to authorized users.
Medulloblastoma; Pediatric brain tumor; Subgroups; Meta-analysis
Pilocytic astrocytoma (PA) is the most common tumor of the pediatric central nervous system (CNS). A body of research over recent years has demonstrated a key role for mitogen-activated protein kinase (MAPK) pathway signaling in the development and behavior of PAs. Several mechanisms lead to activation of this pathway in PA, mostly in a mutually exclusive manner, with constitutive BRAF kinase activation subsequent to gene fusion being the most frequent. The high specificity of this fusion to PA when compared with other CNS tumors has diagnostic utility. In addition, the frequency of alteration of this key pathway provides an opportunity for molecularly targeted therapy in this tumor. Here, we review the current knowledge on mechanisms of MAPK activation in PA and some of the downstream consequences of this activation, which are now starting to be elucidated both in vitro and in vivo, as well as clinical considerations and possible future directions.
Pilocytic; Astrocytoma; Low grade glioma; LGG; BRAF; Fusion; MAPK; Senescence
Medulloblastoma, a small blue cell malignancy of the cerebellum, is a major cause of morbidity and mortality in pediatric oncology. Current mechanisms for clinical prognostication and stratification include clinical factors (age, presence of metastases, and extent of resection) as well as histological subgrouping (classic, desmoplastic, and large cell/anaplastic histology). Transcriptional profiling studies of medulloblastoma cohorts from several research groups around the globe have suggested the existence of multiple distinct molecular subgroups that differ in their demographics, transcriptomes, somatic genetic events, and clinical outcomes. Variations in the number, composition, and nature of the subgroups between studies brought about a consensus conference in Boston in the fall of 2010. Discussants at the conference came to a consensus that the evidence supported the existence of four main subgroups of medulloblastoma (Wnt, Shh, Group 3, and Group 4). Participants outlined the demographic, transcriptional, genetic, and clinical differences between the four subgroups. While it is anticipated that the molecular classification of medulloblastoma will continue to evolve and diversify in the future as larger cohorts are studied at greater depth, herein we outline the current consensus nomenclature, and the differences between the medulloblastoma subgroups.
Medulloblastoma; Consensus; Subgroups; SHH; WNT; Group 3; Group 4
MicroRNAs are 22 nucleotides long non-coding RNAs and exert their function either by transcriptional or translational inhibition. Although many microRNA profiles in different tissues and disease states have already been discovered, only little is known about their target proteins. The microRNA miR-155 is deregulated in many diseases, including cancer, where it might function as an oncoMir.
We employed a proteomics technique called “stable isotope labelling by amino acids in cell culture” (SILAC) allowing relative quantification to reliably identify target proteins of miR-155. Using SILAC, we identified 46 putative miR-155 target proteins, some of which were previously reported. With luciferase reporter assays, CKAP5 was confirmed as a new target of miR-155. Functional annotation of miR-155 target proteins pointed to a role in cell cycle regulation.
To the best of our knowledge we have investigated for the first time miR-155 target proteins in the HEK293T cell line in large scale. In addition, by comparing our results to previously identified miR-155 target proteins in other cell lines, we provided further evidence for the cell line specificity of microRNAs.
Classification and variable selection play an important role in knowledge discovery in high-dimensional data. Although Support Vector Machine (SVM) algorithms are among the most powerful classification and prediction methods with a wide range of scientific applications, the SVM does not include automatic feature selection and therefore a number of feature selection procedures have been developed. Regularisation approaches extend SVM to a feature selection method in a flexible way using penalty functions like LASSO, SCAD and Elastic Net.
We propose a novel penalty function for SVM classification tasks, Elastic SCAD, a combination of SCAD and ridge penalties which overcomes the limitations of each penalty alone.
Since SVM models are extremely sensitive to the choice of tuning parameters, we adopted an interval search algorithm, which in comparison to a fixed grid search finds rapidly and more precisely a global optimal solution.
Feature selection methods with combined penalties (Elastic Net and Elastic SCAD SVMs) are more robust to a change of the model complexity than methods using single penalties. Our simulation study showed that Elastic SCAD SVM outperformed LASSO (L1) and SCAD SVMs. Moreover, Elastic SCAD SVM provided sparser classifiers in terms of median number of features selected than Elastic Net SVM and often better predicted than Elastic Net in terms of misclassification error.
Finally, we applied the penalization methods described above on four publicly available breast cancer data sets. Elastic SCAD SVM was the only method providing robust classifiers in sparse and non-sparse situations.
The proposed Elastic SCAD SVM algorithm provides the advantages of the SCAD penalty and at the same time avoids sparsity limitations for non-sparse data. We were first to demonstrate that the integration of the interval search algorithm and penalized SVM classification techniques provides fast solutions on the optimization of tuning parameters.
The penalized SVM classification algorithms as well as fixed grid and interval search for finding appropriate tuning parameters were implemented in our freely available R package 'penalizedSVM'.
We conclude that the Elastic SCAD SVM is a flexible and robust tool for classification and feature selection tasks for high-dimensional data such as microarray data sets.
Pilocytic astrocytoma (PA) is the most common type of primary brain tumor in children and the second most frequent cancer in childhood. Children with incompletely resected PA represent a clinically challenging patient cohort for whom conventional adjuvant therapies are only moderately effective. This has produced high clinical demand for testing of new molecularly targeted treatments. However, the development of new therapeutics for PA has been hampered by the lack of an adequate in vivo tumor model. Recent studies have identified activation of MAPK signaling, mainly by oncogenic BRAF activation, as a hallmark genetic event in the pathogenesis of human PA. Using in vivo retroviral somatic gene transfer into mouse neural progenitor cells, we have shown here that ectopic expression of the activated BRAF kinase domain is sufficient to induce PA in mice. Further in vitro analyses demonstrated that overexpression of activated BRAF led to increased proliferation of primary mouse astrocytes that could be inhibited by treatment with the kinase inhibitor sorafenib. Our in vivo model for PA shows that the activated BRAF kinase domain is sufficient to induce PA and highlights its role as a potential therapeutic target.