Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Jan 5.
Published in final edited form as: Nat Genet. 2014 Apr 6;46(5):462–466. doi: 10.1038/ng.2950

Recurrent somatic mutations in ACVR1 in pediatric midline high-grade astrocytoma

Adam M Fontebasso 1,33, Simon Papillon-Cavanagh 2,33, Jeremy Schwartzentruber 3,33, Hamid Nikbakht 2, Noha Gerges 2, Pierre-Olivier Fiset 4, Denise Bechet 2, Damien Faury 2,5, Nicolas De Jay 2, Lori Ramkissoon 6, Aoife Corcoran 6, David T W Jones 7, Dominik Sturm 7, Pascal Johann 7, Tadanori Tomita 8, Stewart Goldman 9, Mahmoud Nagib 10, Anne Bendel 11, Liliana Goumnerova 12,13, Daniel C Bowers 14, Jeffrey R Leonard 15, Joshua B Rubin 16, Tord Alden 8, Samuel Browd 17, J Russell Geyer 18, Sarah Leary 18, George Jallo 19, Kenneth Cohen 20, Nalin Gupta 21, Michael D Prados 21, Anne-Sophie Carret 22, Benjamin Ellezam 23, Louis Crevier 24, Almos Klekner 25, Laszlo Bognar 25, Peter Hauser 26, Miklos Garami 26, John Myseros 27, Zhifeng Dong 28, Peter M Siegel 28, Hayley Malkin 29, Azra Ligon 13,30, Steffen Albrecht 4, Stefan M Pfister 7, Keith L Ligon 6,13,30,31,34, Jacek Majewski 2,34, Nada Jabado 1,2,5,34, Mark W Kieran 13,27,32,34
PMCID: PMC4282994  NIHMSID: NIHMS650089  PMID: 24705250

Abstract

Midline pediatric high-grade astrocytomas (pHGAs) are incurable with few treatment targets identified. Most tumors harbor K27M mutations on histone 3 variants. In 40 treatment-naïve midline pHGAs, 39 analyzed by whole-exome sequencing, we find additional somatic mutations specific to tumor location. Gain-of-function mutations in ACVR1 occur in tumors of the pons in conjunction with H3.1 K27M, while FGFR1 mutations/fusions occur in thalamic tumors associated with H3.3 K27M. Hyper-activation of the bone morphogenetic protein (BMP)/ACVR1 developmental pathway in pHGAs harbouring ACVR1 mutations led to increased phospho-SMAD1/5/8 expression and up-regulation of BMP downstream early response genes in tumour cells. Global DNA methylation profiles were significantly associated with the K27M mutation regardless of the mutant H3 variant and irrespective of tumor location, supporting its role in driving the epigenetic phenotype. This significantly expands the potential treatment targets and further justifies pre-treatment biopsy in pHGA as a means to orient therapeutic efforts in this disease.

We and others recently identified recurrent mutations in H3F3A, encoding histone 3 variant 3 (H3.3), in 38% of pediatric supratentorial high-grade astrocytomas (HGAs), and mutations in H3.3 or H3.1 in ~80% of brainstem HGA (diffuse intrinsic pontine gliomas, DIPG)1–3. Mutations of the histone H3 tail altering glycine 34 to either arginine or valine (G34R, G34V), as well as mutations in genes affecting H3 post-translational modifications (PTM) at lysine 36 predominate in cortical tumors, while lysine 27 to methionine (K27M) mutations occur in midline tumors1,2,4,5.

Herein, we focus on the genomic (mutational spectrum, copy number alterations) and epigenetic (DNA methylation) landscape of treatment-naïve midline (thalamus, cerebellum, spine and pons; non-cortical regions) pediatric HGA. These tumors are often surgically challenging or inoperable, and published studies have mainly used post-therapy material and provided limited genomic data other than structural alterations and mutational analysis of H3 variants and TP531,3,6,7. We analyzed 40 midline HGAs, with whole-exome sequencing (WES) data for 39 tumors, including 25 DIPG patient biopsies (11 previously reported2,8, Table S1 and Online Methods). The recurrent H3 mutation K27M was found in 37/40 (93%) cases and was distributed among three histone variants including H3F3A (32/40 samples), HIST1H3B (4/39) and HIST1H3C (1/39). HIST1H3C also encodes the canonical H3.1 (as HIST1H3B) and has not been previously reported to be mutated (Figure 1). All five H3.1 mutations occurred in the pons of younger patients (Figure S1). H3.3 K27M mutations occurred in multiple midline locations, including the brainstem (19/25), thalamus (7/9) and rare locations for HGA – the cerebellum, 4th ventricle and spinal cord (4/4). In contrast, only one H3.3 K27M mutant was identified in 42 pHGAs located in the cerebral hemispheres in a previously reported dataset3,5 (Figure 1, Table S1)

Figure 1. Genomic landscape of pediatric midline high-grade astrocytomas.

Figure 1

Open in a new tab

Distribution of mutations and alterations in 40 pediatric midline high-grade astrocytomas (Midline HGAs) and 42 cortical high-grade astrocytomas and high-grade gliomas (Cortical HGA-HGGs) described in the study. Mutations (mut) were identified with whole-exome sequencing (WES) where available and are indicated by coloured boxes according to the legend in the right panel. Amplifications (amp) and losses were identified utilizing DNA methylation profile derived copy number variant (Methyl-CNV) analysis and are indicated where available by coloured boxes according to the right panel legend. SC = spinal cord, 4th V = 4th ventricle. Boxes coloured in light grey indicate samples for which data is not available. Detailed information regarding tumor samples included herein can be found in Table S1, with specific variants and transcript accessions presented in Table S6.

The pattern of somatic mutations in specific genes or gene pathways revealed striking features (Figure 1, Figure S1). Recurrent somatic mutations in the activin A receptor, type I (ACVR1) occurred in 5/39 midline HGAs, and overlapped with K27M mutants (5/5), mainly H3.1 K27M (4/5) rather than H3.3 K27M or wild-type H3 tumors (p=0.0003, Fisher’s exact test, Figure S1). The mutated amino acid residues in ACVR1 have previously been reported as germline mutations causing fibrodysplasia ossificans progressiva (FOP), an inherited musculoskeletal disease9–14. The substitutions p.R206H, p.G328E, and p.G356D result in ligand-independent activation of the kinase leading to increase of bone morphogenetic protein (BMP) signaling and increased phospho-SMAD1/5/8 production in tissues10, while the novel mutations p.G328V or p.R258G mutations are predicted to exert gain-of-function effects similar to the previously described mutations p.R258S or p.G328E/W/R, based on physico-chemical properties of the mutant residues and predicted protein structure (Figure 4 and Table S2)11,12. ACVR1, also known as ALK2, is a type I receptor of the mammalian TGF-β signalling family with critical developmental roles in the mouse embryo13 and in early left-right patterning14. Investigation of ALK2 pathway activation using phospho-SMAD1/5/8 immunohistochemical staining in DIPG samples showed positive staining only in tumors with mutant ACVR1 (4/4), including in the novel G328V and R258G DIPG mutants (Figure 2). In addition, in the primary cell line DIPGIV, which carries H3.1 K27M and the novel ACVR1 mutation G328V, we demonstrate increased phospho-SMAD1/5/8 levels compared to normal human astrocytes and a significantly increased expression of BMP response-element containing genes ID1-3 and of SNAI1 compared to KNS42, an H3.3 G34V mutant ACVR1 wild-type cell line (Figure 4). These genes are early response genes induced following active BMP2 signaling and SMAD1/5/8 downstream effectors15,16.

Figure 4. Mutations identified in ACVR1 are associated with activation of downstream SMAD signaling pathways.

Figure 4

Open in a new tab

a, Distribution of mutations identified in ACVR1 (n=5) demonstrating their impact of amino acid substitutions on the kinase domain of the protein. b, Immunoblotting analysis of phospho-SMAD1/5/8 levels in ACVR1 wild-type NHA cells grown in 10% FBS, or NHA and DIPGIV (ACVR1 G328V) serum starved for 1 hour in serum/growth factor-free media (0%). c, Quantitative PCR (qPCR) analysis of downstream BMP effectors ID1, ID2, ID3 and SNAI1 expression in ACVR1 mutant (DIPGIV) and wild-type GBM cell line (KNS42). Values represented are fold changes calculated using the 2−ΔΔCt method, normalized to ACTB expression in calibrator NHA cells. P-values calculated using two-tailed T-test for significance, with error bars indicated as standard deviation from two replicates.

Figure 2. Increased levels of phospho-SMAD1/5/8 in ACVR1 mutant pediatric HGAs.

Figure 2

Open in a new tab

Immunohistochemical analysis of midline pediatric HGAs harbouring ACVR1 mutations identified in this study (n=4, left panels) demonstrate increased nuclear positivity of phospho-SMAD1/5/8 compared to pediatric midline HGAs wild-type for ACVR1 mutation (n=3) and control brain samples (right panels), with total-SMAD1 staining shown in each case as positive control. Scale bars indicate 20X=20μm. Clinicopathologic and molecular characteristics of tumor samples are presented in Table S1.

We identified mutations in FGFR1 in association with H3.3 K27M mutation in 4/39 cases: 3 thalamic HGAs also having NF1 mutations (previously reported in8) and one DIPG (Figure 1). These FGFR1 mutations occurred at hotspot residues in the tyrosine kinase domain of the FGFR1 receptor. They have been shown to lead to its constitutive activation in a subset of thalamic pilocytic astrocytoma, a grade I tumor which rarely progresses to higher-grade astrocytomas, and in NIH 3T3 cells8. Comparison with our dataset of non-midline cortical pHGAs indicated tumors carrying H3.1, ACVR1 or FGFR1 mutations to be exclusive to midline HGA (p=0.0040, Fisher’s exact test, Figure 1, Figure S1b).

TP53 mutations occurred in 29/39 samples, largely in combination with H3.3 K27M (23/27) as reported previously1,2 and to a much lesser extent with H3.1 K27M (1/5) (Figure 1, Figure S1b). One DIPG sample (mHGA1) had H3.3 K27M as well as a splicing mutation in CHEK2, a gene associated with Li-Fraumeni syndrome similar to TP53 (Figure 1, Tables S1, S6). Three midline samples displayed no detectable mutation across H3F3A/HIST1H3B/HIST1H3C, even when sequenced at extremely high depth (average read depth across all tumors/genes >20,000X, Table S3). They carried alterations described in adult HGA in the form of combined TP53 mutation, PDGFRA amplification, and CDKN2A loss (mHGA35), PDGFRA and EGFR amplifications and NF1 loss (mHGA36), or the previously identified FGFR1-TACC1 fusion17,18 (mHGA37) (Figure 1, Tables S1, S3–S4).

Somatic PDGFRA mutations were identified in a small subset of midline pHGA (5/39, Table S1). Notably, PDGFRA amplification occurred in 3/12 treatment-naïve DIPGs (Figure S2, Table S1). Interestingly, a cerebellar HGA sample where multiple biopsies were taken from different anatomical loci showed PDGFRA amplification in only one tumor site, while strikingly similar somatic mutation and DNA methylation profiles across all tumor sites were observed (Figure S3). This supports the view that PDGFRA amplifications precede therapy but can be further promoted by it as described in radiation-induced supra-tentorial HGA19. Other growth factor receptors and components described to have focal gains/losses in DIPG6,7 or HGA (MET, RB, PARP1) showed low incidence in our dataset (Figure 1, Tables S1, S4–S6) while recurrent mutations in components of the PI3K pathway, predicted to activate AKT signaling, were present in 10/39 midline HGA. Regarding alterations affecting telomere length, ATRX mutations were identified in 9 midline tumors, exclusively in H3.3 K27M-TP53 mutant samples and affected older children1 while no TERT-promoter mutations, assessed using targeted sequencing, were identified in midline pHGAs (Table S1).

K-to-M substitutions in histone variants at residue K27 have recently been shown to inhibit SET-domain containing histone methyltransferases20, possibly accounting for the specific DNA methylation pattern we observed in H3.3 K27M non-brainstem tumors4. When clustering our samples by global DNA methylation, all samples with a given H3 mutation (K27, G34 or absent) clustered together, but did not group based on tumor location within the brain, the particular H3 gene mutated, and additional partner mutations or structural alterations identified (Figure 3 and Figure S5–S6). The global epigenetic profile is thus strongly associated with the alteration in the histone 3 variant mark. Interestingly, TP53 alteration was associated with increased broad copy number changes as we previously showed2 (average of 32 events per sample) indicating a level of genomic instability. This number was significantly higher in H3.3 G34R/V mutant samples compared to other groups (Table S7), in keeping with global hypomethylation identified in these tumors4.

Figure 3. Clustering analysis of global DNA methylation profiles of 98 pediatric high-grade astrocytomas.

Figure 3

Open in a new tab

Global DNA methylation clustering analysis of pediatric HGAs distributed across the brain demonstrates similar impact on epigenomic dysregulation caused by K27M mutation regardless of age, brain location, associated mutation or the particular histone 3 variant affected. The top 10,000 most variable normalized methylation β-values were utilized for UPGMA clustering with colour key and scale in top left panel, and indicated mutations of interest in inset left panel. Robustness was assessed utilizing multiscale bootstrapping in Figure S6. Detailed information regarding sample clinicopathologic characteristics are included in Table S1 with methylation-derived CNVs presented in Tables S4–5.

The gain-of-function alterations in three growth factor receptors, ACVR1, FGFR1 and PDGFRA, associate with H3 K27M variants in midline HGA. These mutations are not seen concurrently, and ACVR1 and FGFR1 mutations are mutually exclusive with TP53 alterations and privilege specific locations within the midline of the brain. ACVR1 mutations do not seem to correlate with differential survival in patients with HGA, although no definite conclusions can be drawn based on our limited sample size (Figure S4). The lack of reported CNS tumor development in FOP patients or Acvr1/Alk2 mouse models suggests that aberrant activation of this pathway is not sufficient for tumorigenesis10 and may act in concert with H3 K27M mutations and other alterations we identify in the PI3K pathway in these tumors to induce tumorigenesis (Figure 1). Interestingly, ACVR1 is mainly expressed at E14 in the cortex of mouse embryos and at very low level in the brainstem, and aberrantly active ACVR1 leads to increased ventralization of zebrafish embryos11,12,21. Aberrant ACVR1 signaling in the brainstem, a midline structure for a gene involved in left-right patterning, may potentially specify a patterning defect in DIPG. FGFR signaling regulates neural progenitor maintenance and the development of the ventral midbrain22. Similar to ACVR1 mutations, the FGFR1 gain-of-function mutations we identify only lead to grade I astrocytomas if present on their own in a tumor8, and are commonly found in association with NF1 and H3.3 K27M mutations in midline HGA. Mutations in H3 variants in pediatric HGAs mirror IDH-mutations in their requirement for additional alterations to potentially induce HGA. G34R/V mutations are found in H3F3A and not in other histone genes to date, and invariably associate with TP53 and ATRX mutations in the cortex. In midline pHGA, K27M mutations arise in H3 variants that differ based on age and tumor location. They associate with TP53 mutations, or with activated neuro-developmental growth factor receptor pathways through distinct hits in ACVR1, FGFR1 or PDGFRA to achieve tumorigenesis. These alterations in growth factor receptors and in members of the PI3K pathway offer previously unforeseen therapeutic possibilities in a deadly cancer, while the observed level of genomic instability calls for caution in the choice of adjuvant therapy whenever possible. Importantly, we show that small needle pre-therapy biopsies can reliably identify the mutational landscape in HGA. This will allow for tailoring available therapies to results obtained following stereotactic biopsy in children affected by this fatal brain tumor, while effort continues to be made to target K27M mutations.

Online Methods

Patient samples and consent

All samples were obtained with informed consent after approval of the Institutional Review Board of the respective hospitals they were treated in and were independently reviewed by senior pediatric neuropathologists (SA, KLL) according to the WHO guidelines. Samples were obtained from the Montreal Children’s Hospital (Montreal, McGill University Health Centre), Boston Children’s Hospital (Boston, Harvard University), Brain Tumor Toronto Bank (BTTB, Toronto, University Health Network) and collaborators in Hungary, in addition to previously midline high-grade astrocytoma published samples in2,5 with sequencing data (n=11) and4,5 with previously published DNA methylation data (n=89) for a total of 98 tumors included herein for DNA methylation and copy number variant (CNV) analysis. DIPG biopsy samples obtained pre-therapy were from DFCI protocol 10–321 (n=12), a prospective phase II biopsy study of newly diagnosed DIPG. The protocol is IRB approved through the Dana-Farber Harvard Cancer Center IRB, FDA IND #111,882, ClinicalTrials.gov Identifier NCT01182350, has local institutional approval at all participating sites and informed consent was obtained from all parents. Additional pediatric midline high-grade astrocytoma samples were from needle biopsies or partial resections prior to treatment (n=28). Sequencing and clinical data for this cohort is presented in Table S1, and methylation-derived CNVs in genes of interest are presented in Tables S4–5.

Whole-exome DNA sequencing

Standard genomic DNA extraction methods were performed according to described company protocols (Qiagen). Paired-end library preparations were carried out using the Nextera Rapid Capture Exome kit according to instructions from the manufacturer (Illumina) from 50 ng of total starting genomic DNA. Sequencing was performed in rapid run mode with 100 bp paired end reads on Illumina HiSeq 2000. We removed adaptor sequences and quality trimmed reads using the Fastx toolkit (http://hannonlab.cshl.edu/fastx_toolkit/) and then used a custom script to ensure that only read pairs with both mates present were subsequently used. Reads were aligned to hg19 with BWA 0.5.923, and indel realignment was done using the GATK24. Duplicate reads were then marked using Picard (http://picard.sourceforge.net/) and excluded from downstream analyses. We assessed coverage of consensus coding sequence (CCDS) bases using the GATK, which showed that the majority of samples had >92% of CCDS bases covered by at least 10 reads, and >88% of CCDS bases covered by at least 20 reads.

For each sample, single nucleotide variants (SNVs) and short insertions and deletions (indels) were called using samtools mpileup25 with the extended base alignment quality (BAQ) adjustment (-E), and were then quality filtered to require at least 20% of reads supporting the variant call. Variants were annotated using both Annovar26 and custom scripts to identify whether they affected protein coding sequence, and whether they had previously been seen in the 1000 genomes dataset (Nov. 2011), the NHLBI GO exomes, or in approximately ~1000 exomes previously sequenced at our center. Useful links:

Variants in candidate genes of interest described herein in the midline HGA cohort (n=39) are detailed in Table S6.

MiSeq targeted high-depth DNA sequencing of H3F3A, HIST1H3B and HIST1H3C

Genomic DNA from midline HGA samples was utilized for high-depth sequencing of H3F3A, HIST1H3B and HIST1H3C genes to investigate the frequency of K27M reads in samples, notably those wild-type by whole-exome sequencing and high-resolution melting assays. Midline HGAs (n=24) were sequenced using the MiSeq sequencing platform (Illumina) and generated an average coverage of > 20,000X of the analogous K27M base change across the three histone variants (more specifically, for H3F3A gene, the average coverage is > 12,000X, for HIST1H3B it is > 12,000X, and for HIST1H3C the coverage is >35,000X). The reads were mapped to the reference genome (Human hg19) using bwa genome aligner23. The alignment files were fed to mpileup tool from SamTools package25 in order to find all the variations without any filter applied by the conventional variant callers. An in-house parser program was developed to extract different variations at the desired positions (in this case, K27) in the mapped paired reads covering the histone genes H3F3A, HIST1H3B and HIST1H3C. The results are provided in Table S3.

RNA sequencing

RNA was extracted from patient tumor mHGA37 using the Qiagen RNeasy lipid tissue Mini Kit (Qiagen) according to instructions from the manufacturer. Library preparation was performed with ribosomal RNA (rRNA) depletion methods according to instruction from the manufacturer (Epicentre) to achieve greater coverage of mRNA and other long non-coding transcripts. Paired-end sequencing was performed on the Illumina HiSeq 2000 platform.

RNA-seq fusion analysis

RNA-seq Fastq files were utilized for fusion analysis using the deFuse software algorithm 27 according to indicated settings. Alorithmic output was then analyzed for high-confidence fusion transcripts, which were then re-constructed in silico utilizing online bioinformatics tools and databases including Basic Local Alignment Search Tool, BLAST (http://blast.ncbi.nlm.nih.gov/Blast), Ensembl (http://useast.ensembl.org/index.html), UniProt (http://www.uniprot.org/) and the UCSC Genome Browser (http://genome.ucsc.edu) to assess impact on putative fusion proteins and compare them with existing, previously described fusions8,17.

TERT promoter mutation sequencing

Characterized mutations in the TERT promoter, C228T, and C250T variants with G>A nucleotide substitutions at genomic positions 1,295,228 and 1,295,250 (hg19) respectively28 were sequenced using the Sanger method in midline samples (n=14) and cortical samples (n=10) using the following cycling conditions: 96°C for 1min, 96°C for 10s, 60°C for 5s, 72°C for 1s, 72°C for 30s repeated for 33 cycles. Primer sequences are detailed in Table S8.

DNA methylation analysis

Extracted tumor DNA was analyzed for genome-wide DNA methylation patterns utilizing the HumanMethylation450 BeadChip platform according to instructions from the manufacturer (Illumina) and analyzed as described in4,5. From the selection of probes on the array, we removed probes from sex chromosomes (chrX and Y) as well as those located at sites with documented SNPs (as per dbSNP: http://www.ncbi.nlm.nih.gov/SNP/). Methylation values were normalized utilizing the Subset-quantile Within Array Normalization (SWAN) procedure provided in the R package minfi29. We performed hierarchical clustering using the 10,000 most variable sites. Distance was assessed using d = 1 − r, where r is the Pearson product-moment coefficient. Clustering was done using average linkage (UPGMA) and was validated for robustness of the procedure via multiscale resampling (1,000 iterations) using the R package pvclust 30 (Figure S6).

Copy number variant (CNV) detection

In order to assay CNV in our samples, we used a methylation-based method and controls implemented in the R/Bioconductor packages CopyNumber450k and CopyNumber450kData, respectively. CNV analysis for copy number gains and losses in previously described genes of interest is presented in Tables S4 and S5 respectively. Gross genomic aberrations were assessed using UCSC genome banding and our in-house algorithm described above. Bands covered with >90% significantly amplified or deleted segments were counted as “abnormal” and summed for each sample. Samples were grouped together by mutation type and a t-test was performed to assess statistically significant differences in aberration count between mutation subgroups. CNV analysis for broad areas of genomic instability is included in Table S7.

Cell lines and Western blotting

hTert-immortalized normal human astrocytes (NHA) were obtained from Dr Abhijit Guha (Labatt Brain Tumour Centre, Toronto, Canada) and were grown in DMEM supplemented with 10% FBS as previously described. DIPGIV primary cell line (kind gift of Dr. Michelle Monje, Stanford University, USA) and KNS42 (purchased from the JCRB) were grown as previously described31. Cell lines were serum/growth factor starved for 1 hour prior to protein extraction. Proteins were extracted in Tris-NaCl-EDTA lysis buffer as previously described 32 and blotting was performed for phosphoSMAD1/5/8 (Cell Signaling Technology, #9511 at 1:500 dilution in 5% bovine serum albumin (BSA) solution) on total lysates from NHA cells and from DIPGIV cells as previously described12. Beta-actin (13E5) (Cell Signaling Technology) was used as loading control.

Quantitative PCR for BMP target genes

Quantitative PCR (qPCR) was performed to assess for levels of activity downstream of the ACVR1/ALK2 receptor in total RNA extracted from DIPGIV (H3,1K27M-ACVR1 G328V mutant) and KNS42 (H3.3G34V ACVR1 wild-type) and NHA cells grown in DMEM supplemented with 10% FBS using ID1, ID2, ID3 and SNAI1 using primer sequences detailed in Table S8. Briefly, total RNA was extracted using the miRNeasy mini kit (Qiagen) according to manufacturer’s instructions with purity and integrity assessed utilizing Nanodrop (Thermo-Fisher) and Experion (Biorad) methodologies. 100ng of RNA was used for reverse-transcription using the iScript RT Supermix (BioRad) following manufacturer’s instructions. Real-time PCR was run on a Lightcycler 96 (Roche) with the SsoFast Evagreen SuperMix kit (BioRad). Cycling conditions were: 95°C for 30 sec followed by 40 cycles 95°C for 5sec/60 °C for 20 sec. Fold change values were calculated utilizing the 2−ΔΔCt method with ACTB expression and NHA cells used as the calibrator.

Immunohistochemistry

Immunohistochemistry (IHC) was performed on formalin-fixed, paraffin-embedded (FFPE) slides from patients with (n=4) or without ACVR1 mutation (n=3) and control normal brain (n=1) to assess activation/phosphorylation of Smad1/5/8 downstream of the ACVR1/ALK2 receptor (reviewed in 33). IHC was carried out as previously described34 utilizing anti-phospho-Smad1/5/8 (Cell Signaling Technology, #9511) and anti-Smad1 (Invitrogen #38-5400). IHC processing and imaging was performed blind of ACVR1 mutation status with representative images presented in Figure 2.

Fluorescence in-situ hybridization for PDGFRA amplification

Fluorescence in-situ hybridization (FISH) was performed as previously described in35,36. Briefly, FISH was performed utilizing 4μm thick tissue sections from a subset of midline HGAs, with a BAC probe directed towards the PDGFRA genomic locus at chromosomal region 4q12 (RP11-231c18, green) and a probe directed against 4p11.2-4q11.1 (CEN4, red) as a control to visualize chromosome 4. Scoring is included in Table S1, and representative images in Figure S2.

Supplementary Material

Supplementary Figures
Supplementary Tables

Acknowledgments

The authors would like to express their sincere gratitude toward all staff at the McGill University and Genome Quebec Innovation Centre for excellent technical expertise, library preparation and sequencing. This work was performed within the context of the I-CHANGE consortium (International Childhood Astrocytoma iNntegrated Genomics and Epigenomics consortium) and supported by funding from Genome Canada, Genome Quebec, The Institute for Cancer Research of the Canadian Institutes for Health Research (CIHR) McGill University and the Montreal Children’s Hospital Foundation. It was also supported by the Hungarian Scientific Research Fund (OTKA) Contract No. T-04639, the National Research and Development Fund (NKFP) Contract No. 1A/002/2004 (PH, MG) and TÁMOP-4.2.2A-11/1/KONV-2012-0025 (AK, LB). N. Jabado is a member of the Penny Cole lab and the recipient of a Chercheur Clinicien Senior Award. J. Majewski holds a Canada Research Chair (tier 2). We acknowledge the support of the Zach Carson DIPG Fund at the Dana-Farber Cancer Institute (DFCI), the Ellie Kavalieros Fund at DFCI, the Mikey Czech Foundation, the Prayer From Maria Foundation, the Hope for Caroline Fund at DFCI, the Ryan Harvey DIPG Fund at DFCI, the Stop&Shop Pediatric Brain Tumor Program at DFCI and the Pediatric Brain Tumor Clinical and Research Fund at DFCI. A.M.F. is supported by a studentship from the CIHR, as well as an award from the CIHR Systems Biology Training Program at McGill University. D. Bechet is supported by a studentship from the T.D trust/Montreal Children’s Hospital Foundation and N. Gerges by a studentship from Cedars. N.D.J. is supported by an award from the McGill Integrated Cancer Research Training Program.

Footnotes

Accession codes

Whole-exome sequencing data and DNA methylation data can be accessed through the European Genome-phenome Archive (EGA) at the following accession number: EGAS00001000720.

Author contributions

AMF, N. Gerges, POF, D. Bechet, DF, LR, AC, AL, SA, ZD performed experiments. AMF, SPC, JS, HN, NDJ, AL, SA, ZD, PMS analyzed the data and produced figures and tables. DTWJ, DS, PJ, TT, SG, MN, AB, LG, D. Bowers, JRL, JR, TA, SB, RG, GJ, KC, N. Gupta, MDP, ASC, BE, LC, AK, LB, PH, MG, J. Myseros, HM, SA, SMP provided tissue samples. KLL, J. Majewski, NJ, MWK provided project leadership and designed the study. All authors contributed to the final manuscript.

Conflict of Interest

The authors declare that they have no competing financial interests.

References

  • 1.Khuong-Quang DA, et al. K27M mutation in histone H3.3 defines clinically and biologically distinct subgroups of pediatric diffuse intrinsic pontine gliomas. Acta Neuropathol. 2012;124:439–47. doi: 10.1007/s00401-012-0998-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Schwartzentruber J, et al. Driver mutations in histone H3.3 and chromatin remodelling genes in paediatric glioblastoma. Nature. 2012;482:226–31. doi: 10.1038/nature10833. [DOI] [PubMed] [Google Scholar]
  • 3.Wu G, et al. Somatic histone H3 alterations in pediatric diffuse intrinsic pontine gliomas and non-brainstem glioblastomas. Nat Genet. 2012 doi: 10.1038/ng.1102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Sturm D, et al. Hotspot mutations in H3F3A and IDH1 define distinct epigenetic and biological subgroups of glioblastoma. Cancer Cell. 2012;22:425–37. doi: 10.1016/j.ccr.2012.08.024. [DOI] [PubMed] [Google Scholar]
  • 5.Fontebasso AM, et al. Mutations in SETD2 and genes affecting histone H3K36 methylation target hemispheric high-grade gliomas. Acta Neuropathol. 2013;125:659–69. doi: 10.1007/s00401-013-1095-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Paugh BS, et al. Genome-wide analyses identify recurrent amplifications of receptor tyrosine kinases and cell-cycle regulatory genes in diffuse intrinsic pontine glioma. J Clin Oncol. 2011;29:3999–4006. doi: 10.1200/JCO.2011.35.5677. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Zarghooni M, et al. Whole-genome profiling of pediatric diffuse intrinsic pontine gliomas highlights platelet-derived growth factor receptor alpha and poly (ADP-ribose) polymerase as potential therapeutic targets. J Clin Oncol. 2010;28:1337–44. doi: 10.1200/JCO.2009.25.5463. [DOI] [PubMed] [Google Scholar]
  • 8.Jones DT, et al. Recurrent somatic alterations of FGFR1 and NTRK2 in pilocytic astrocytoma. Nat Genet. 2013;45:927–32. doi: 10.1038/ng.2682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Shore EM, et al. A recurrent mutation in the BMP type I receptor ACVR1 causes inherited and sporadic fibrodysplasia ossificans progressiva. Nat Genet. 2006;38:525–7. doi: 10.1038/ng1783. [DOI] [PubMed] [Google Scholar]
  • 10.Fukuda T, et al. Generation of a mouse with conditionally activated signaling through the BMP receptor, ALK2. Genesis. 2006;44:159–67. doi: 10.1002/dvg.20201. [DOI] [PubMed] [Google Scholar]
  • 11.Chaikuad A, et al. Structure of the bone morphogenetic protein receptor ALK2 and implications for fibrodysplasia ossificans progressiva. J Biol Chem. 2012;287:36990–8. doi: 10.1074/jbc.M112.365932. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Fukuda T, et al. A unique mutation of ALK2, G356D, found in a patient with fibrodysplasia ossificans progressiva is a moderately activated BMP type I receptor. Biochem Biophys Res Commun. 2008;377:905–9. doi: 10.1016/j.bbrc.2008.10.093. [DOI] [PubMed] [Google Scholar]
  • 13.Mishina Y, Crombie R, Bradley A, Behringer RR. Multiple roles for activin-like kinase-2 signaling during mouse embryogenesis. Dev Biol. 1999;213:314–26. doi: 10.1006/dbio.1999.9378. [DOI] [PubMed] [Google Scholar]
  • 14.Kishigami S, et al. BMP signaling through ACVRI is required for left-right patterning in the early mouse embryo. Dev Biol. 2004;276:185–93. doi: 10.1016/j.ydbio.2004.08.042. [DOI] [PubMed] [Google Scholar]
  • 15.Miyazono K, Maeda S, Imamura T. BMP receptor signaling: transcriptional targets, regulation of signals, and signaling cross-talk. Cytokine Growth Factor Rev. 2005;16:251–63. doi: 10.1016/j.cytogfr.2005.01.009. [DOI] [PubMed] [Google Scholar]
  • 16.Korchynskyi O, ten Dijke P. Identification and functional characterization of distinct critically important bone morphogenetic protein-specific response elements in the Id1 promoter. J Biol Chem. 2002;277:4883–91. doi: 10.1074/jbc.M111023200. [DOI] [PubMed] [Google Scholar]
  • 17.Singh D, et al. Transforming fusions of FGFR and TACC genes in human glioblastoma. Science. 2012;337:1231–5. doi: 10.1126/science.1220834. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Zhang J, et al. Whole-genome sequencing identifies genetic alterations in pediatric low-grade gliomas. Nat Genet. 2013;45:602–12. doi: 10.1038/ng.2611. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Paugh BS, et al. Integrated molecular genetic profiling of pediatric high-grade gliomas reveals key differences with the adult disease. J Clin Oncol. 2010;28:3061–8. doi: 10.1200/JCO.2009.26.7252. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Lewis PW, et al. Inhibition of PRC2 activity by a gain-of-function H3 mutation found in pediatric glioblastoma. Science. 2013;340:857–61. doi: 10.1126/science.1232245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Shen Q, et al. The fibrodysplasia ossificans progressiva R206H ACVR1 mutation activates BMP-independent chondrogenesis and zebrafish embryo ventralization. J Clin Invest. 2009;119:3462–72. doi: 10.1172/JCI37412. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Lahti L, Peltopuro P, Piepponen TP, Partanen J. Cell-autonomous FGF signaling regulates anteroposterior patterning and neuronal differentiation in the mesodiencephalic dopaminergic progenitor domain. Development. 2012;139:894–905. doi: 10.1242/dev.071936. [DOI] [PubMed] [Google Scholar]
  • 23.Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25:1754–60. doi: 10.1093/bioinformatics/btp324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.McKenna A, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20:1297–303. doi: 10.1101/gr.107524.110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Li H, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics. 2009;25:2078–9. doi: 10.1093/bioinformatics/btp352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010;38:e164. doi: 10.1093/nar/gkq603. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.McPherson A, et al. deFuse: an algorithm for gene fusion discovery in tumor RNA-Seq data. PLoS Comput Biol. 2011;7:e1001138. doi: 10.1371/journal.pcbi.1001138. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Horn S, et al. TERT promoter mutations in familial and sporadic melanoma. Science. 2013;339:959–61. doi: 10.1126/science.1230062. [DOI] [PubMed] [Google Scholar]
  • 29.Maksimovic J, Gordon L, Oshlack A. SWAN: Subset-quantile within array normalization for illumina infinium HumanMethylation450 BeadChips. Genome Biol. 2012;13:R44. doi: 10.1186/gb-2012-13-6-r44. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Suzuki R, Shimodaira H. Pvclust: an R package for assessing the uncertainty in hierarchical clustering. Bioinformatics. 2006;22:1540–2. doi: 10.1093/bioinformatics/btl117. [DOI] [PubMed] [Google Scholar]
  • 31.Monje M, et al. Hedgehog-responsive candidate cell of origin for diffuse intrinsic pontine glioma. Proc Natl Acad Sci U S A. 2011;108:4453–8. doi: 10.1073/pnas.1101657108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Siegel PM, Ryan ED, Cardiff RD, Muller WJ. Elevated expression of activated forms of Neu/ErbB-2 and ErbB-3 are involved in the induction of mammary tumors in transgenic mice: implications for human breast cancer. EMBO J. 1999;18:2149–64. doi: 10.1093/emboj/18.8.2149. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Schmierer B, Hill CS. TGFbeta-SMAD signal transduction: molecular specificity and functional flexibility. Nat Rev Mol Cell Biol. 2007;8:970–82. doi: 10.1038/nrm2297. [DOI] [PubMed] [Google Scholar]
  • 34.Kleinman CL, et al. Fusion of TTYH1 with the C19MC microRNA cluster drives expression of a brain-specific DNMT3B isoform in the embryonal brain tumor ETMR. Nat Genet. 2014;46:39–44. doi: 10.1038/ng.2849. [DOI] [PubMed] [Google Scholar]
  • 35.Ramkissoon LA, et al. Genomic analysis of diffuse pediatric low-grade gliomas identifies recurrent oncogenic truncating rearrangements in the transcription factor MYBL1. Proc Natl Acad Sci U S A. 2013;110:8188–93. doi: 10.1073/pnas.1300252110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Firestein R, et al. CDK8 is a colorectal cancer oncogene that regulates beta-catenin activity. Nature. 2008;455:547–51. doi: 10.1038/nature07179. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Figures
NIHMS650089-supplement-Supplementary_Figures.pdf (2.4MB, pdf)
Supplementary Tables
NIHMS650089-supplement-Supplementary_Tables.xlsx (128.9KB, xlsx)

RESOURCES