Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Sep 1.
Published in final edited form as: Mol Cancer Res. 2017 May 18;15(9):1243–1254. doi: 10.1158/1541-7786.MCR-16-0389

Histone H3.3K27M represses p16 to accelerate gliomagenesis in a murine model of DIPG

Francisco J Cordero 1,2, Zhiqing Huang 5, Carole Grenier 5, Xingyao He 7, Guo Hu 1, Roger McLendon 2, Susan K Murphy 2,5, Rintaro Hashizume 6,7, Oren Becher 1,2,3,4,8,9,**
PMCID: PMC5581686  NIHMSID: NIHMS878484  PMID: 28522693

Abstract

Diffuse Intrinsic Pontine Glioma (DIPG) is a highly aggressive pediatric brainstem tumor genetically distinguished from adult GBM by the high prevalence of the K27M mutation in the histone H3 variant H3.3 (H3F3A). This mutation reprograms the H3K27me3 epigenetic landscape of DIPG by inhibiting the H3K27-specific histone methyltransferase EZH2. This globally reduces H3K27me2/3, critical repressive marks responsible for cell fate decisions, and also causes focal gain of H3K27me3 throughout the epigenome. To date the tumor-driving effects of H3.3K27M remain largely unknown. Here it is demonstrated that H3.3K27M cooperates with PDGF-B in vivo, enhancing gliomagenesis and reducing survival of p53 WT and knockout murine models of DIPG. H3.3K27M expression drives increased proliferation of tumor-derived murine neurospheres, suggesting that cell cycle deregulation contributes to increased malignancy in mutant tumors. RNA sequencing (RNA-Seq) on tumor tissue from H3.3K27M expressing mice indicated global upregulation of PRC2 target genes, and a subset of newly repressed genes enriched in regulators of development and cell proliferation. Strikingly, H3.3K27M induced targeted repression of the p16/ink4a (CDKN2A) locus, a critical regulator of the G0/G1 to S phase transition. Increased levels of H3K27me3 were observed at the p16 promoter; however, pharmacological reduction of methylation at this promoter did not rescue p16 expression. While DNA methylation is also present at this promoter, it is not K27M-dependent. Intriguingly, inhibition of DNA methylation restores p16 levels and is cytotoxic against murine tumor cells. Importantly, these data reveal that H3.3K27M-mediated p16 repression is an important mechanism underlying the proliferation of H3.3K27M tumor cells as in vivo cdkn2a knockout eliminates the survival difference between H3.3K27M and H3.3WT tumor-bearing mice.

Keywords: H3.3K27M, p16/Ink4a, brainstem glioma, DIPG, PRC2

Introduction

Despite over half a century of research directed toward the development of an efficacious treatment for Diffuse Intrinsic Pontine Glioma (DIPG), a rare and deadly tumor arising within the pons region of the brainstem predominantly in children, no therapies have improved the dismal survival rates. DIPG patients have a median survival of less than one year after diagnosis and less than 20% of patients survive past 2 years. The diffuse and infiltrating nature of the tumor coupled with its extremely sensitive location precludes surgical resection. Currently the standard regimen of radiation therapy provides only temporary relief from symptoms, and no chemotherapy has shown efficacy beyond radiation alone (1, 2).

Recent large-scale analyses of DIPG patient samples has allowed for major advances in our understanding of the driving mechanisms underlying DIPG. These studies have identified the most commonly occurring genetic alterations, which include H3.3/H3.1 K27M mutations (80%), mutations in TP53 (77%), and amplification of PDGFRA (36%) (3–11). The H3 histone mutations not only differentiate pediatric from adult gliomas, but also define distinct gene expression profiles between cortical and brainstem gliomas specific to the particular amino acid tail residue mutations (12). The H3.3K27M mutation has increasingly gained interest due to its high frequency, specific spatiotemporal occurrence, and gain-of-function mechanism of action. H3.3K27M histones act as potent inhibitors of H3K27 di- and tri-methylation (H3K27me2/3) by binding to the SET domain of EZH2, the histone methyltransferase component of the Polycomb repressive complex 2 (PRC2) causing global reduction of these repressive marks (13). The H3.3K27M-mediated global changes comprise a specific epigenetic signature that may hold valuable insight into the driving mechanisms behind this aggressive disease (14, 15).

ChIP-seq and whole genome bisulfite sequencing studies of K27M mutant DIPG and pediatric high-grade glioma patient samples identified focal regions retain or gain H3K27me3 along with the global loss of H3K27me3 and subsequent DNA hypomethylation profile (15). The CDKN2A locus was included as a site of novel PRC2-dependent gain of promoter H3K27me3 (14). Homozygous deletion of the CDKN2A locus, which harbors both the p16/INK4A and p19/ARF transcripts, though a common genetic alteration across all human cancer types, rarely occurs in DIPG (16). Although the tumor suppressor function of p16, an endogenous CDK4/6 inhibitor, has been thought to remain predominantly uncompromised in DIPG, phase I clinical trials are underway assessing the safety of CDK4/6 inhibitors in treating affected children (17). To date however no work has demonstrated whether this novel increase in H3K27me3 at the CDKN2A locus is directly mediated by the H3.3K27M oncohistone to regulate its expression or whether repression of this locus contributes to gliomagenesis.

To address the role of H3.3K27M in DIPG, we developed an in vivo brainstem glioma mouse model incorporating the H3.3K27M mutation and platelet-derived growth factor-B (PDGFB) activation, in both p53 WT and knockout settings, using the RCAS/tv-a avian retroviral system (13, 18). We demonstrate that the expression of H3.3K27M cooperates with PDGF signaling to increase proliferation and tumor grade in vivo in both p53 WT and deficient models. Intriguingly, H3.3K27M causes global gene de-repression as expected, but also novel gene repression. We find that p16 is a target for aberrant H3.3K27M-induced repression and while there is a localized gain in promoter H3K27me3, only inhibition of DNA methylation rescues gene expression. Importantly, genetic knockout of cdkn2a abolishes the survival difference between mice harboring H3.3K27M- and H3.3WT- expressing tumors. Additionally, H3.3K27M expressing tumor cells show significantly increased sensitivity to CDK4/6 inhibition by palbociclib compared to H3.3WT controls. This establishes the H3.3K27M-induced repression of p16 as a critical mechanism contributing to the acceleration of gliomagenesis by the mutation and highlights regulation of the G1-S transition as a promising therapeutic node in DIPG.

Materials and Methods

In vivo mouse model experiments

All animals were maintained in accordance with the Duke Animal Care and Use Committee under the approved protocol (A162-16-07). The following mice were used: Nestin Tv-a; p53fl/fl (C57BL/6J background) as previously described (19). Nestin Tv-a; CDKN2Afl/fl mice (C57BL/6J background) were obtained by crossing nestin tv-a mice with cdkn2afl/fl mice, provided by Ron Depinho. Mice were monitored for tumor growth by weight loss and neurological symptoms.

Cell lines and culture

Murine tumors were isolated and enzymatically digested as previously described (19). GFP-positive single cells were sorted by FACS and cultured in Neurocult media with proliferation supplement (Stem Cell Technologies) as neurospheres. Cells were split using Accutase (Innovative Cell Technologies) and all experiments were performed using primary (non-thawed) lines at low passage (under passage 10). For BrdU analysis, single cells were seeded into 96-well plates in biological and experimental triplicate, cultured for 72hrs, and then assessed for BrdU incorporation using the Cell Proliferation ELISA Kit (Roche) per the manufacturer’s recommendation. For drug treatments of both murine and human DIPG lines, single cells were seeded into 96-well plates in biological and experimental triplicate and treated with drugs or vehicle (0.1% DMSO) the following day. Primary pediatric human glioma cell lines, SF8628, and DIPG007 (HSJD-DIPG-007), contain H3.3K27M mutations, SU-DIPG-IV contains H3.1K27M, and all lines were obtained from Dr. Rintaro Hashizume at Northwestern University, Dr. Michelle Monje at Stanford University, and Dr. Angel Montero Carcaboso at Hospital Sant Joan de Déu Barcelona, Spain, in accord with institutionally approved protocol at each institution and the cell culture models have been previously described(20–22). SF8628 cells derived from surgical biopsy are maintained as an exponentially growing monolayer in complete medium consisting of Dulbecco’s modified Eagle’s medium (DMEM, GIBCO 11965, Invitrogen, Carlsbad, California) supplemented with 10% fetal bovine serum with penicillin-streptomycin and plasmocin. SU-DIPG IV cell culture derived from DIPG autopsy tissue are grown as tumor neurospheres in Tumor Stem Media (TSM) consisting of DMEM/F12 (Invitrogen, Carlsbad, CA), Neurobasal(-A) (Invitrogen, Carlsbad, CA), B27(-A) (Invitrogen, Carlsbad, CA), human-bFGF (20ng/ml) (Shenandoah Biotech, Warwick, PA), human-EGF (20ng/ml) (Shenandoah, Biotech, Warwick, PA), human PDGF-AB (20ng/ml) (Shenandoah, Biotech, Warwick, PA) and heparin (10ng/ml). DIPG007 (HSJD-DIPG-007) cells are derived from the autopsy are maintained as an exponentially growing monolayer in TSM media supplemented with 5% fetal bovine serum. Human cell histone mutational status was determined using Sanger sequencing for the H3F3A and HIST1H3B genes. Human cell cultures were validated by DNA fingerprinting using short tandem repeat (STR) analysis and checked for mycoplasma contamination. Human lines were cultured for 2–3 passages (2 weeks) following thaw for experiments. Other experimental details are as described in the text. Inhibitors used were GSK343, GSK126, EPZ-6438, Decitabine, azacitidine, and palbociclib (Selleck). No mycoplasma testing regimen was performed on murine cell lines as they are early passage tumor-derived cells.

Generation of murine brainstem gliomas

All animal studies were performed in accordance with the Duke University Animal Care and Use Committee and Guide for the Care and Use of Laboratory Animals (protocol # A214-13-08). The RCAS/TVA system was used to generate murine brainstem gliomas. DF1 virus producing cells were purchased from ATCC, cultured in DMEM (ATCC) supplemented with 10% FBS, 2mM L-glutamine, 100 units/mL penicillin and 100µg/mL streptomycin, and incubated at 39°C and 5% CO2. Cells were transfected with RCAS plasmids (RCAS-PDGF-B, RCAS-Cre, RCAS-H3.3WT, RCAS-H3.3K27M, RCAS-p53shRNA) using X-TremeGENE 9 (Roche) per the manufacturer’s instructions. 1µL (105 cells) of DF1 cells (ATCC Manassas, VA) expressing equal ratios of RCAS virus-producing cells were injected intracranially into the brainstem of postnatal day 3–4 nestin tv-a (Ntv-a) p53fl/fl mice to generate brainstem gliomas as previously described (19). Mice were euthanized with CO2 upon appearance of symptoms of brain tumor development (weight loss, lethargy, head tilt, or hydrocephalus) in accordance with Duke University IACUC protocol. Brain tissue from these animals was extracted and either fixed in 10% neutral buffered formalin and paraffin embedded or used to generate tumor cell lines.

In vitro infection of brainstem progenitors with RCAS viruses

Normal brainstem progenitors from nestin-tv-a mice were isolated and infected with concentrated RCAS-H3.3K27M-HA, RCAS-H3.3WT-HA, or RCASY (empty vector) viruses as previously described and grown under neurosphere conditions (23).

Tumor grading

Tumor samples fixed in 10% formalin for 24hrs were embedded in paraffin by the Duke Pathology Core and cut into 5µM sections using a Leica RM2235 microtome. H&E staining was performed using standard protocols. Tumor grading was performed by a blinded neuropathologist (RM).

Immunohistochemistry analysis

Formalin fixed brains were paraffin embedded by Duke Pathology Core Services. Sections were cut 5µm thick using a Leica RM2235 Microtome. Immunohistochemistry was performed using an automated processor (Discovery XT, Ventana Medical Systems, Inc.) Antibodies used are provided in the supplemental material. Quantification was performed using Metamorph software. We analyzed 10 different high-powered fields (40x), and quantified total nuclear area of positive staining to total nuclear area.

Immunofluorescence studies

Tumor-bearing mice were injected with EdU (10mg/kg) 4hrs prior to sacrifice, brains were fixed in 10% formalin, paraffin embedded and cut into 5µM sections. Slides were deparaffinized in xylenes and rehydrated with decreasing ethanol solutions. EdU staining was performed following the protocol for Click-iT EdU Alexa Fluor 594 Imaging Kit (Invitrogen #C10339). Quantification was performed using Metamorph software. We analyzed 10 different high-powered fields (40x), and quantified total EdU positive staining to total nuclear area (DAPI stained).

Western blots

Western blots on histone-extracted proteins were performed as previously described (13). Antibodies used are provided in the supplementary material. Quantification of band intensities was performed using Image J software.

MRI Tumor volume analysis

T2-weighted scans of the brain were acquired using a cryogenic coil (Cryoprobe, Bruker Biospin, Billerica, MA) and the following parameters: repetition time TR=1.5s, echo time TE=22ms, bandwidth 75kHz, flip angle 90, matrix: 160×160, 100 slices, field-of-view 1.6×1.6×1.0cm. Tumor volumes were obtained from manual segmentation.

RNA-seq analysis

Read-pairs were aligned to the mouse genome (mm9) by Expression Analysis. The number of read-pairs per gene was determined based upon the NCBIM37 gene annotation from the ENSEMBL database (24). Genes that did not have at least 10 read-pairs in any single library were excluded from further analysis. Normalization and differential expression analysis was carried out using the DESeq bioconductor package (default parameters, GSE98765) (25).

Reverse transcription, Real-time Quantitative PCR (qRT-PCR) analysis

Total RNA was isolated using RNeasy kit (QIAGEN) per the manufactures protocol. cDNA was synthesized from total mRNA using Superscript II and OligodT primers (Invitrogen). qRT-PCR TaqMan primers were used for murine p16, and murine p19 and sequences are provided in the supplementary material. Relative gene expression levels were generated using the ΔΔCt method (26).

Flow Cytometry

Cell cycle analysis by flow cytometry was performed by fixing cells in 70% ethanol on ice for 15mins. Cells were washed, stained with propidium iodide labeling solution (BD), and transferred to FACS tubes. Cell cycle analysis was performed by the Duke Cancer Center Flow Cytometry Core.

Chromatin Immunoprecipitation followed by quantitative PCR (ChIP-qPCR)

Tumor neurospheres were dissociated using Accutase (Innovative Cell Technologies) and counted using the Scepter automated cell counter (Millipore). Five million cells were fixed in 1% formaldehyde for 7mins at room temperature, quenched with the addition of 125mM glycine for 5mins and washed with 1xPBS before storing at −80°C. Cells were resuspended in 500µL lysis buffer (10mM HEPES, 0.5% NP-40, 1.5mM MgCl2, 10mM KCl) with protease inhibitors and incubated on ice for 10mins. After centrifugation at 5,000RPM for 5mins the nuclear pellet was lysed in 500µL nuclear lysis buffer (50mM Tris, 1%SDS, 10mM EDTA) with protease inhibitors for 15mins. Extracts were sonicated, cleared by centrifugation, and diluted to 1 million cells per mL in dilution buffer (0.01%SDS, 1.1% Triton X-100, 1.2mM EDTA, 167mM NaCl, 16.7mM Tris). Antibody incubation (4µg) was performed overnight. ChIP antibodies used are located in the supplemental material. 40µL of 50% Protein A/G beads (GE Healthcare) were added to lysates and rotated at 4C for 3hrs. Protein bound beads were sequentially washed with wash I (20mM Tris, 150mM NaCl, 2mM EDTA, 1% Triton X-100, 0.1%SDS), wash II (20mM Tris, 500mM NaCl, 2mM EDTA, 1% Triton X-100, 0.1% SDS), wash III (10mM Tris, 250mM LiCl, 1mM EDTA, 1% NP-40, 1% deoxycholate), then washed twice with TE. DNA was eluted with 1% SDS, 0.1M NaHCO3 and reverse cross-linked by incubation at 65C with RNase overnight. Eluants were then treated with proteinase K and DNA was purified using QIAGEN PCR purification kit. DNA amplification was performed by qPCR as described above.

Pyrosequencing

Primers for CDKN2a pyrosequencing assays were designed using PSQ assay design software version 1.0.6 (Qiagen). The primer sequences and PCR conditions are provided in the supplementary material. Seven microliters PCR products were used for pyrosequencing following the protocol from the manufacturer (Qiagen).

Statistical Analysis

Statistical analysis was performed using Graphpad Prism (Version 6.0). All survival curves were analyzed by Log-Rank (Mantel-Cox) test. IHC and IF analyses were performed using the Mann-Whitney test. IC50 and growth curve determinations and significance calculations were found by nonlinear regression analysis. All other assays were performed using either paired or unpaired standard t-test.

Results

H3.3K27M cooperates with PDGF signaling to increase tumor grade and proliferation in vivo

Previous studies have shown that H3.3K27M expression correlates with increased disease aggressiveness in patients and increased proliferation and stem cell characteristics in vitro (3, 27). In addition, some evidence suggests a nestin-positive, hedgehog responsive progenitor within the brainstem is a candidate for the cell of origin for DIPG (28). Therefore we utilized the RCAS/tv-a system to express the H3.3K27M mutation in nestin-expressing progenitor cells of the neonatal murine brainstem to model DIPG in vivo and investigate the direct pro-tumorigenic mechanisms of H3.3K27M. We found that H3.3K27M expression alone did not induce tumors by histological analysis of the brains (Supplemental Table 1).

PDGFB overexpression alone in our nestin tv-a murine model normally induces low-grade brainstem gliomas with the mice typically remaining asymptomatic (29). We examined the combined expression of PDGFB with H3.3K27M, H3.3WT, or empty vector to determine the impact of H3.3K27M on survival and tumor grade. All PDGFB; empty vector control mice survived to the 12-week endpoint as expected. Similarly, the majority of the PDGFB; H3.3WT control mice survived to the 12-week endpoint. Interestingly, 22% of the PDGFB; H3.3K27M expressing mice developed tumor symptoms before the 12-week endpoint that met our criteria for sacrifice (Figure 1A). Histological analysis of both PDGFB; empty vector, and PDGFB; H3.3WT brains showed only grade II tumors, while the PDGFB; H3.3K27M cohort demonstrated 47% grade II, 47% grade III and 6% grade IV tumor incidences (Figure 1B). The grade III-IV gliomas within the PDGFB; H3.3K27M group displayed increased microvascular proliferation, mitotic figures, and overall increased cellular density as detected by H&E staining (Figure 1C). Consistent with these findings, PDGFB; H3.3K27M tumors harbored significantly increased EdU-positive cycling cells as compared to controls (Figure 1C–D). Tumor volumes analyzed using MRI scans also complemented these results, revealing a trend for larger tumors with H3.3K27M expression (Supplemental Figure 1A). Thus, H3.3K27M increases proliferation, raises tumor grade, and reduces latency of PDGF-B driven, p53 WT brainstem glioma in mice.

Figure 1. H3.3K27M cooperates with PDGF signaling to increase tumor malignancy.

Figure 1

Open in a new tab

A. Nestin tv-a mice were injected with RCAS-PDGFB and either RCAS-Y, RCAS-H3.3WT, or RCAS-H3.3K27M. Kaplan-Meier survival curve of all three groups showing a significant difference between the PDGFB; H3.3K27M and PDGFB; RCASY groups (N= 15 for PDGF-B; empty vector, N= 50 for PDGF-B; H3.3K27M, p= 0.048). B. Graph of tumor grades for PDGFB; RCASY (n= 7), PDGFB; H3.3WT (n=13), and PDGFB; H3.3K27M (n= 15). C. Representative H&E (40x) stained murine tumors driven by PDGFB signaling in combination with 1.) RCASY empty vector 2.) H3.3WT, or 3.) H3.3K27M (top panels) and Immunofluorescence (IF) images (40x) of these groups with EdU (red) injection 4hrs prior to sacrifice to mark proliferating cells. Total nuclei stained with DAPI (blue). D. IF image quantification of total EdU positive nuclear area over total nuclear area for PDGFB; RCASY (n=6), PDGFB; H3.3WT (n=4) and PDGFB; H3.3K27M (n= 5) groups show a significantly higher percentage of EdU positive nuclei in PDGFB; H3.3K27M compared to both PDGFB; RCASY (p= 0.017) and PDGFB; H3.3WT (p= 0.016).

H3.3K27M accelerates tumorigenesis in a high-grade murine brainstem glioma model

We generated high-grade murine models of DIPG by adding RCAS-Cre mediated p53 knockout to our PDGFB driven tumors and expressing H3.3WT or H3.3K27M with either HA- or GFP-tags. Tumors were invasive, localized to the brainstem region, highly proliferative, and showed high expression of olig2, nestin, and the H3.3WT or H3.3K27M constructs by IHC (Figure 2A). Power analysis for the HA-tagged model indicated that the accurate detection of a minimum 5-day survival difference would require at least 40 mice per group. With this experimental design, we found that H3.3K27M-HA mice exhibited significantly shorter tumor latency with a median survival of 36 days compared to 41 days in the H3.3WT-HA group (Figure 2B and Supplementary Table 1). The GFP-tagged constructs showed similar results, with H3.3K27M mutant mice succumbing to tumor an average of 8 days earlier compared to H3.3WT-GFP mice (36 days and 44 days, respectively, Supplemental Figure 1B). Analysis of H3.3WT-HA and H3.3K27M-HA tumors showed no difference in tumor cell proliferation by phosphorylated H3 (pH3) IHC or EdU immunofluorescence (Supplemental Figure 1C,D).

Figure 2. H3.3K27M cooperates with PDGFB and p53 knockout to decrease survival and increase tumor cell proliferation.

Figure 2

Open in a new tab

A. Immunohistochemistry of high-grade H3.3K27M tumor model. 1x H&E shows lower brainstem tumor localization. Tumors show high expression of Olig2, GFP (histone tag), pH3, and nestin with low H3K27me3 at 40x. B. Nestin tv-a; p53fl/fl mice were injected with RCAS-PDGFB, RCAS-Cre, and either RCAS-H3.3WT-HA (n=40) or RCAS-H3.3K27M-HA (n=55). Kaplan-Meier curve of high grade PDGFB, p53 knockout tumors including either H3.3WT-HA or H3.3K27M-HA expression. K27M group shows significantly reduced survival (p= 0.0065). C. Histone extraction western blot analysis of in vitro cultured tumor derived neurospheres for H3K27me3, GFP (fused to H3.3 RCAS constructs), and total H3 proteins. D. BrdU incorporation proliferation assay shows significantly increased proliferation in isolated H3.3K27M-GFP expressing tumor derived neurosphere lines (n= 5 each, p= 0.003). E. Cell cycle analysis was performed on propidium iodide stained H3.3WT or H3.3K27M mutant murine glioma lines (n= 3 each). K27M-GFP expressing lines showed a significantly reduced percentage of cells in the G0/G1 stage (p= 0.006) with corresponding increases in S- and G2/M phases (p= 0.02 and p=0.01, respectively).

We next derived purified primary tumor lines using fluorescence-activated cell sorting (FACS) to isolate the GFP-tagged H3.3WT or H3.3K27M populations and cultured the cells short term in serum-free conditions for in vitro analysis. The GFP-purified neurospheres allow us to study the effects of H3.3K27M in isolation of the tumor bulk while better retaining in vivo tumor properties in an in vitro setting (30). We first performed western blots to determine the level of exogenous histone overexpression and analyze H3K27me3 levels. We confirmed that H3.3K27M expressing tumors harbored the characteristic global decrease in H3K27me3 by western blot. Additionally, the exogenous GFP-tagged histones were expressed at near equal levels between H3.3K27M and H3.3WT controls and showed comparable expression to endogenous histone levels (Figure 2C, Supplemental Figure 1E). BrdU incorporation assays show that H3.3K27M cells proliferate at a significantly higher rate as compared to the H3.3WT-GFP control tumor lines (Figure 2D). Furthermore, cell cycle analysis comparing H3.3K27M and H3.3WT controls revealed a significant shift in the H3.3K27M lines toward the S and G2/M phases with a corresponding reduction in the G0/G1 phase cells (Figure 2E). This further indicates a higher percentage of proliferating cells in H3.3K27M mutant lines. Taken together these data show that expression of H3.3K27M increases tumor cell proliferation in PDGFB driven brainstem gliomas independent of p53 status.

H3.3K27M induces localized repression of p16 expression

To elucidate potential mechanisms underlying H3.3K27M-induced increased malignancy and cell cycle progression, we extracted tumor tissue from PDGFB; p53−/−; empty vector control and PDGFB; p53−/−; H3.3K27M expressing high-grade tumors and performed RNA-Sequencing (RNA-Seq) analysis. We found a total of 224 significantly differentially expressed genes with 182 upregulated and 42 downregulated genes in the H3.3K27M group compared to the controls (Figure 3A, Supplementary Material). Unsupervised hierarchical clustering and principal component analysis showed H3.3K27M group tumors have a distinct gene expression profile compared to the PDGFB; p53−/−; empty vector controls (Figure 3A,B). Gene set enrichment analysis demonstrated a significant upregulation of PRC2 regulated genes from embryonic and neural progenitor (Supplemental Figure 2A). Gene ontology (GO) using DAVID software indicated upregulated genes were enriched in neuronal function and morphogenesis as reported in DIPG patient studies (Figure 3C) (15). Additionally, we used GSEA analysis to compare our mouse tumors to DIPG patient data and found varying levels of similarities (Supplemental Figure 2B).

Figure 3. H3.3K27M induces repression of p16 expression.

Figure 3

Open in a new tab

A. Unsupervised hierarchical clustering of differentially regulated genes (padj <0.05) between H3.3K27M overexpressing gliomas and PDGFB; p53-KO; empty vector controls (n= 3 each). B. Principal component analysis comparing PDGFB; p53 −/−; H3.3K27M and PDGFB; p53 −/−; empty vector tumors. C. GO pathway analysis results using DAVID software of upregulated and downregulated genes identified by RNA-Seq. D. qPCR validation of p16 and p19 transcript levels from murine tumor derived in vitro grown neurosphere lines show H3.3K27M induces significant repression of p16 compared to H3.3WT controls (n= 3 each, p < 0.0001)

Intriguingly, known PRC2 target genes were also among the downregulated genes in H3.3K27M tumors, suggesting residual PRC2-induced repression despite inhibition of EZH2 methyltransferase function by H3.3K27M. GO analysis of the downregulated genes revealed enrichment for developmental and cell cycle regulatory genes (Figure 3C). The critical tumor suppressor locus Cdkn2a is among the H3.3K27M-repressed loci (Figure 3C). As the cdkn2a locus harbors both the p16/ink4a and p19/Arf alternate transcripts, we utilized qPCR to analyze both transcripts and found specific p16 downregulation in both in vitro cultured tumor cells as well as ex-vivo infected brainstem progenitor cells (Figure 3D, Supplemental Figure 2C). Thus, these results suggest the H3.3K27M mutation specifically represses p16 without altering p19 expression levels, potentially contributing to the accelerated cell cycle progression and increased proliferation seen in our mutant tumor cells.

Aberrant increase of H3K27me3 is localized to the p16 promoter

As the CDKN2A locus is a known PRC2 target we next analyzed H3K27me3 levels at the p16 promoter as a possible H3.3K27M-induced silencing mechanism (31, 32). To test this we designed ChIP-qPCR primers for the p16 promoter, as well as the p19 promoter, p16 exon 1α, and p16 exon 2 (Figure 4A). We then performed ChIP-qPCR for H3K27me3 on our in vitro GFP-purified tumor lines. We used the hoxa11, a site that is highly repressed by PRC2 in adult cells, and the β-actin promoter as positive and negative controls for our ChIP-qPCR analysis respectively, and found high H3K27me3 levels in both H3.3K27M and H3.3WT cells at hoxa11, with very low levels at β-actin (Supplemental Figure 3A). Analysis at the p16 promoter revealed that H3.3K27M significantly increased H3K27me3 levels at the p16 promoter with no significant difference at the p19 promoter (Figure 4B). H3K27me3 analysis at p16 exon 1α and exon 2 showed trends for increased methylation (Supplemental Figure 3B). Of note, ChIP for the GFP tag fused to H3.3WT and H3.3K27M constructs showed both exogenous histones integrated into the p16 promoter region at equal levels, and are unaffected by EZH2 inhibition, suggesting that the difference observed for H3K27me3 is not a result of reduced occupancy of the mutant oncohistone at this locus (Figure 4C).

Figure 4. H3.3K27M-induced gain of H3K27me3 at the p16 promoter.

Figure 4

Open in a new tab

A. Schematic of Cdkn2a locus showing locations of ChIP primers for p16 and p19 promoters (region 1 and 2), as well as p16 exon 1α and exon 2 (regions 3 and 4). Green bars indicate CpG island locations B. ChIP-qPCR of H3K27me3 at the p16 and p19 promoters in cultured, high-grade murine tumor cells (n=3) C. ChIP-qPCR for GFP tagged histone constructs localized to the p16 promoter after a 7 day 1µM GSK343 or DMSO vehicle treatment (n=3) D. Western blot for H3K27me3 and total H3 of murine tumor derived neurospheres after a 7 day treatment with either vehicle or 1µM GSK343. E. ChIP-qPCR of H3K27me3 at the p16 promoter of cultured tumor cells after a 7 day, 1µM GSK343 or DMSO vehicle treatment (*= p< 0.05, ** = p<0.002). F. qPCR for p16 expression levels in response to GSK343 or vehicle of in vitro murine tumor cell lines. (n=3 per group). G. Cell count assay on murine tumor cell lines (n= 3 each) treated with 1µM GSK343 for 7 days. Growth curves were significantly different between H3.3K27M+vehicle and both H3.3WT+ GSK343 and H3.3WT+ vehicle (p= 0.005 and p= 0.0009, respectively). There was also a significant difference between H3.3K27M+GSK343 and both H3.3WT+GSK343 and H3.3WT+vehicle (p=0.006 and p= 0.001, respectively).

We next treated H3.3WT or H3.3K27M murine tumor lines with two potent EZH2 inhibitors, GSK343 and EPZ6438, to test whether the increased H3K27me3 levels at the p16 promoter regulates p16 expression (33, 34). Both drugs induced global reductions in H3K27me3 levels after 7days of 1µM treatment (Figure 4D, Supplemental Figure 3C). Additionally, GSK343 treatment led to a specific reduction in promoter H3K27me3 at the p16 locus (Figure 3E). Interestingly, we measured p16 promoter H3K4me3 levels with and without EZH2 inhibition by ChIP and detected the mark in both H3.3WT and H3.3K27M cells with no significant difference in the vehicles. However there was a significant difference in the response of H3.3K27M cells to EZH2 inhibition, inducing increased H3K4me3 levels at the p16 promoter (Supplemental Figure 4D). Surprisingly, while the EZH2 inhibitors reduced p16 H3K27me3 to near background levels and increased promoter H3K4me3, there was no corresponding rescue of p16 expression after the 7 day or with an extended 21-day treatment (Figure 4F, Supplemental Figure 4E). Accordingly, EZH2 inhibition with GSK343, EPZ6438, and an additional EZH2 inhibitor GSK126, did not affect proliferation of either H3.3K27M or H3.3K27M cells, with no observed IC50 up to 10µM (Figure 4G and Supplemental Figure 3F). Taken together, these data suggest that there are additional regulators involved in the H3.3K27M-induced inhibition of p16.

Inhibition of DNA methylation rescues p16 expression in H3.3K27M cells

Earlier studies provide evidence suggesting PRC2 components interact with DNA methylation machinery, recruiting them to sites of H3K27me3 silencing (35). To determine whether the increased p16 promoter H3K27me3 in H3.3K27M cells coincides with DNA methylation, we designed pyrosequencing primers for the region within the p16 promoter overlapping the ChIP-qPCR primers (Figure 5A). While 6 out of 8 H3.3K27M and only 3 out of 9 H3.3WT tumor-derived neurosphere lines exhibited higher than 20% promoter DNA methylation, there was no significant difference in the average percent methylation between the groups (Figure 5B, Supplemental Table 2). Additionally, there was no apparent increase in localization of DNMT1, which is responsible for the maintenance of DNA methylation during replication, to the p16 promoter in the H3.3K27M-expressing tumor cells (Supplemental Figure 4A). Notably, the pharmacological inhibition of EZH2 did not affect DNA methylation levels at the p16 promoter of both H3.3K27M and H3.3WT tumor cells (Supplemental Figure 4B).

Figure 5. The role of DNA methylation in H3.3K27M-induced p16 repression.

Figure 5

Open in a new tab

A. Schematic of CpG islands (green bars) within the murine cdkn2a locus and pyrosequencing primer locations (blue bars). B. Pyrosequencing results for DNA methylation at the p16 promoter of in vitro H3.3WT or H3.3K27M murine tumor cells (n=3, p= 0.15). C. Percent DNA methylation levels at the p16 promoter of H3.3WT and H3.3K27M murine tumor lines after 1µM decitabine treatment for 72hrs (n=3, p= 0.017 for H3.3K27M vehicle vs decitabine). D. p16 transcript expression in murine cell lines after 72hrs 1µM decitabine treatment on murine cell lines (n=3, p= 0.016 for H3.3K27M vehicle vs decitabine).

We next treated our in vitro murine tumor cells and cultured H3.3K27M mutant DIPG patient lines with the potent DNA methyltransferase inhibitor decitabine to examine whether a reduction in promoter DNA methylation could rescue p16 expression. Treatment of tumor lines with 1µM decitabine for 72 hours significantly reduced p16 promoter DNA methylation specifically in H3.3K27M cells in both mouse and human DIPG lines (Figure 5C, Supplemental Figure 4C). Strikingly, this treatment significantly rescued p16 mRNA expression to near WT levels in the H3.3K27M mouse cells and resulted in increased p16 protein expression in human DIPG lines (Figure 5D, Supplemental Figure 4D, Supplemental Table 3). Additionally, decitabine treatment exhibited a potent anti-proliferative effect independent of the H3.3K27M mutation in both mouse and human lines. Treatment with a different DNA methyltransferase, azacitidine, showed the same effect in mouse tumor cells (Supplemental Figure 4E). Collectively, these data suggest that DNA methylation plays a role in regulating p16 expression in H3.3K27M cells.

Loss of p16/ink4a is central to H3.3K27M-mediated increased tumor cell proliferation and malignancy

The p16 tumor suppressor is an important regulator of cell cycle entry by inhibiting cyclin dependent kinases 4 and 6(CDK4/6) from phosphorylating RB protein leading to inhibition of S-phase entry (36). In order to determine whether the p16/RB pathway contributes to the K27M-mediated increased cell proliferation, we utilized palbociclib, a potent and specific CDK4/6 inhibitor FDA-approved for the treatment of breast cancer to specifically target the pathway (37, 38). We have previously shown that palbociclib significantly increases survival in a murine brainstem glioma model induced with PDGF-B in cdkn2a knockout mice (19). Therefore tumor cells expressing the mutant histone may exhibit increased sensitivity to this drug due to the reduced p16 levels. Indeed, the IC50 for the H3.3K27M expressing cells was significantly lower as compared to the H3.3WT controls (0.18µM, and 1.5µM, respectively, Figure 6A). Cell cycle analysis of H3.3WT and H3.3K27M tumor cells treated with 2.5µM palbociclib for 48 hours showed no effect of the drug on H3.3WT cells, while the percentage of H3.3K27M cells undergoing S-phase entry was significantly reduced with the treatment (Figure 6B).

Figure 6. Cdkn2a is central to the H3.3K27M-induced increased tumor cell proliferation and decreased tumor latency.

Figure 6

Open in a new tab

A. IC50 determination of the CDK4/6 inhibitor palbociclib on H3.3WT-GFP or H3.3K27M-GFP murine tumor derived lines using BrdU incorporation. Cell lines were incubated with palbociclib for 48hrs before the assay was performed using the following concentrations: 0nM, 1nM, 10nM, 100nM, 500nM, 1µM, and 5µM. The IC50s for the H3.3WT and H3.3K27M lines were calculated to be 1.5µM and 0.18µM, respectively (n=3 per group, p= 0.002). B. Cell cycle analysis summary of H3.3K27M-GFP and H3.3WT-GFP FACS purified tumor derived neurospheres treated with 2.5µM palbociclib for 48hrs (* = p< 0.05, ** = p< 0.009). Additionally, H3.3WT+ PD vs H3.3K27M+ vehicle in S- and G2/M phases, and H3.3WT+ vehicle vs H3.3K27M+PD are statistically significant (p< 0.009, p< 0.05, and p< 0.05, respectively). C. Nestin tv-a; cdkn2afl/fl mice were injected with RCAS-PDGFB, RCAS-p53shRNA, and either RCAS-H3.3WT (n=17) or H3.3K27M (n= 22), with or without RCAS-Cre (n=6 for both). Kaplan-Meier survival shows significant survival differences between H3.3K27M vs H3.3WT p= 0.016, H3.3K27M vs H3.3K27M+Cdkn2a−/− p= 0.0007, H3.3K27M vs H3.3WT+ Cdkn2a−/− p= 0.0001, H3.3WT vs H3.3K27M+ Cdkn2a−/− p< 0.0001, and H3.3WT vs H3.3WT+Cdkn2a−/− p< 0.0001.

Finally, we incorporated a Cdkn2a floxed allele into our RCAS murine tumor model in order to determine whether reduced p16 expression is a key component of H3.3K27M-mediated accelerated tumor development. We induced tumors in nestin tv-a; cdkn2afl/fl mice by injection of RCAS viruses containing PDGFB, p53shRNA, and either the H3.3WT-GFP or H3.3K27M-GFP constructs along with RCAS-Cre or empty vector with the caveat that p19/Arf is also deleted in these mice. Power analysis indicated 20 mice per group were required to determine the impact of H3.3K27M in this model. We observed significantly reduced survival in the H3.3K27M-GFP group, as seen in our other H3.3K27M expressing models, with a median survival of 37 days compared to 46 days in the H3.3WT-GFP group (Figure 6C). Interestingly, the cdkn2a knockout cohorts exhibited the most aggressive tumor development with both H3.3WT and H3.3K27M expressing groups (median survival 30 and 32 days, respectively) displaying significantly shorter survival than the cdkn2a WT groups. Importantly, the H3.3K27M-induced survival difference in this DIPG model was completely abolished with knockout of cdkn2a (Figure 6C). Of note, the reduced variability of the cdkn2a knockout models allows for sufficient power with n= 6 per group. Together with the increased sensitivity to CDK4/6 inhibition, this critically demonstrates that H3.3K27M-induced repression of p16 expression is a major factor in the accelerated tumor development and increased proliferating cells seen in our mutant models.

Discussion

Understanding how the hallmark H3.3K27M mutation contributes to DIPG development and progression is key in developing efficacious therapies. Our studies use genetic mouse modeling to demonstrate that H3.3K27M directly accelerates gliomagenesis by cooperating with PDGF signaling independent of p53. Furthermore, our work provides strong evidence that repression of p16 is an important component of H3.3K27M-driven gliomagenesis.

H3.3K27M cooperates with PDGF signaling and p53 loss to accelerate tumorigenesis

Previous patient studies have implicated that the presence of H3.3K27M mutation correlates with increased malignancy, decreased survival, PDGFRA amplifications, and TP53 alterations (3–5). These data suggest that there is a relationship between the tumor driving mechanisms of these alterations, and our work here along with other studies has provided evidence in favor of this (22, 27). Interestingly our IHC analysis of the high-grade model revealed no significant difference in the percentage of proliferating cells between H3.3K27M and H3.3WT tumors. This may be a consequence of the time point in which tissue was collected, as mice were sacrificed upon reaching euthanasia criteria, at a late stage of the disease. Once cultured however, we show H3.3K27M expression leads to significantly increased proliferation.

Our previous work revealed that H3.3K27M expression combined with p53 loss induced ectopic proliferating clusters, but was not sufficient to induce tumors (13). Our work here suggests the H3.3K27M mutation requires PDGF signaling as a mitogen to exert its pro-tumorigenic effects. It remains to be seen whether the histone mutations (either H3.1 or H3.3) induce similar effects in the context of other genetic alterations such as PPM1D mutations, ACVR1 mutations, or PI3K pathway alterations, all of which are found in DIPG (39, 40).

Epigenetic mechanisms play a role in H3.3K27M-induced p16 repression

RNA-Seq analysis of our murine tumors revealed a substantial number of differentially expressed genes, including most notably a subset of repressed genes in H3.3K27M tumors. Comparing expression results from our model to human DIPG published data shows mixed results by GSEA. This could be attributed to different species, and the complex factors affecting patient samples such as treatment effects, post-mortem collection, and human genetics. Nevertheless, our GO term pathways for the upregulated and downregulated genes showed enrichment in neuronal, and developmental pathways respectively, which have been reported in human studies. Importantly, the work by Chan et al. indirectly suggested that H3.3K27M expression induced specific p16 repression by gain of H3K27me3 using normal neural stem cells as controls. Our work adds to this by revealing that the p16 suppression is a direct result of H3.3K27M expression in vivo, as well as in vitro in both tumor-derived neurospheres and single construct expressing neuronal progenitors.

While CDKN2A alteration is common to many cancers including adult high-grade gliomas and pediatric high-grade gliomas in the cerebral cortex, it is rare in DIPG (9, 17, 41). However, our work reveals that expression of H3.3K27M may recruit epigenetic regulators to repress p16. We find that H3.3K27M induces a focal increase in p16 promoter H3K27me3, a canonically repressive epigenetic mark. A recent study has also noted the importance of H3.3K27M-induced p16 repression in DIPG using a different mouse model and human DIPG lines, but there are also notable differences (22). Mohammad et al observed that inhibition of EZH2 is sufficient to restore p16 expression and inhibit cell growth. However, we show that EZH2 inhibitors do not rescue p16 expression nor inhibit proliferation at a dose that is effective at reducing the H3K27me3 gain at the p16 promoter, in our H3.3K27M murine cells. Our results are in agreement with a recent report in which EZH2 inhibitors were shown to be ineffective against human DIPG lines in vitro suggesting additional repressive systems are involved (42). The discrepant observations could be due to cell-type specific differences or could be related to p53 status of the various models as suggested by Mohhammad et al, or other unknown factors, and highlights the complexity of studying H3.3K27M mutations.

Continued p16 repression after eliminating the promoter H3K27me3 gain suggests additional regulation acting dependent or independently of the repressive histone mark. H3K27me3 is known to recruit DNA methyltransferases (DNMTs) to drive de novo DNA methylation in certain cancers, and the Cdkn2a locus can be regulated by both H3K27me3 and DNA methylation of promoter CpG islands (31, 43, 44). The relationship between these epigenetic regulators is complex but interdependent (43, 45–48). This remains true in the context of H3.3K27M as tumors from pediatric glioma patients harboring these mutations exhibit global reduction of H3K27me3 and a DNA hypomethylation profile (15). Therefore it is possible that, in the context of H3.3K27M expression, regions of aberrant gains in H3K27me3 may also harbor increased DNA methylation providing an additional repression.

Supportive of this hypothesis here we find that regions of DNA methylation overlap with regions of increased H3K27me3 within the p16 promoter. Additionally, inhibition of DNA methylation rescued p16 expression specifically in H3.3K27M cells, although the mechanism remains unclear and warrants further study. Regulation of DNA methylation is complex but important to tumorigenesis, as recent evidence shows that hypoxia can induce loss of TET enzyme activity, which catalyzes DNA demethylation, and therefore drives DNA hypermethylation in tumors (49). However, several works have shown that PRC2 binding to H3.3K27M stalls the complex, which may increase recruitment of DNMTs to regions with H3K27M occupancy (14, 50, 51). Although we did not observe increased localization of DNMT1 to the p16 promoter, which is responsible for the maintenance of DNA methylation during replication, DNMT3A or DNMT3B, which are the only other mammalian DNMTs and perform de novo methylation of unmethylated, and can methylate hemimethylated DNA, may be responsible (52). Thus, further investigation is required to determine whether the H3.3K27M-induced increase in H3K27me3 acts as a recruiting platform for DNA methylation or whether other unknown factors are involved is needed.

H3.3K27M-induced repression of p16 drives accelerated gliomagenesis

Preferential repression of p16 is known to enhance mitogenic responsiveness and promote tumor progression (44). We find the H3.3K27M-induced p16 repression imparts increased sensitivity to the CDK4/6 inhibitor palbociclib compared to H3.3WT cells. Our data indicate that palbociclib, currently in clinical trials for the 10–30% of patients with genetic aberrations of the RB pathway, may be effective for a much larger percentage of DIPG patients than originally presumed.

Essential to this work, genetic knockout of p16 eliminated the H3.3K27M-induced survival difference between mutant and control mice, suggesting that repression of this gene is key in H3.3K27M-enhanced gliomagenesis. Together these findings support a mechanism whereby PDGF signaling and localized H3.3K27M-induced repression of p16 cooperate to accelerate tumor cell growth and increase malignancy in DIPG models. This work highlights that rescue of p16 activity through inhibitors of DNA methylation, CDK4/6, or a combination strategy as a promising avenue for DIPG therapy that may impart additional efficacy to radiation alone.

Supplementary Material

1
2
3
4
5
6
7
8
9

Implications.

This study shows that H3.3K27M mutation and PDGF signaling act in concert to accelerate gliomagenesis in a genetic mouse model and identifies repression of p16 tumor suppressor as a target of H3.3K27M highlighting the G1-S cell cycle transition as a promising therapeutic avenue.

Acknowledgments

The authors thank Dr. Michelle Monje and Dr. Angel Montero Carcaboso for sharing their DIPG human cell lines, David Corcoran and the Duke Genomic Analysis and Bioinformatics Core for their help with RNA-Seq analysis, Yasheng Gao for microscopy support, the Duke Cancer Center Flow Cytometry Core for their aid with FACS and cell cycle analysis, and the Armstrong, Linardic, and Wechsler labs for discussions. They also thank Katherine Misuraca and Emily Miller for their help with editing and discussions. The RCAS-p53shRNA was generously provided by the Holland lab. MRI was performed at the Duke Center for In Vivo Microscopy, an NIBIB National Biomedical Technology Resource Center (4 P41 EB015897).

Grant Support

This work was supported by NIH grant NS093986-01 (F.J. Cordero), K02 award NS086917 (OJB), the Damon Runyon Cancer Research Foundation (OJB), and the Rory David Deutsch Foundation (OJB).

References

  • 1.Donaldson SS, Laningham F, Fisher PG. Advances toward an understanding of brainstem gliomas. J Clin Oncol. 2006;24:1266–72. doi: 10.1200/JCO.2005.04.6599. [DOI] [PubMed] [Google Scholar]
  • 2.Hargrave D, Bartels U, Bouffet E. Diffuse brainstem glioma in children: critical review of clinical trials. Lancet Oncol. 2006;7:241–8. doi: 10.1016/S1470-2045(06)70615-5. [DOI] [PubMed] [Google Scholar]
  • 3.Khuong-Quang DA, Buczkowicz P, Rakopoulos P, Liu XY, Fontebasso AM, Bouffet E, 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]
  • 4.Schwartzentruber J, Korshunov A, Liu XY, Jones DT, Pfaff E, Jacob K, 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]
  • 5.Sturm D, Witt H, Hovestadt V, Khuong-Quang DA, Jones DT, Konermann C, 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]
  • 6.Wu G, Broniscer A, McEachron TA, Lu C, Paugh BS, Becksfort J, et al. Somatic histone H3 alterations in pediatric diffuse intrinsic pontine gliomas and non-brainstem glioblastomas. Nat Genet. 2012;44:251–3. doi: 10.1038/ng.1102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Castel D, Philippe C, Calmon R, Le Dret L, Truffaux N, Boddaert N, et al. Histone H3F3A and HIST1H3B K27M mutations define two subgroups of diffuse intrinsic pontine gliomas with different prognosis and phenotypes. Acta Neuropathol. 2015;130:815–27. doi: 10.1007/s00401-015-1478-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Wu G, Diaz AK, Paugh BS, Rankin SL, Ju B, Li Y, et al. The genomic landscape of diffuse intrinsic pontine glioma and pediatric non-brainstem high-grade glioma. Nat Genet. 2014;46:444–50. doi: 10.1038/ng.2938. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Paugh BS, Qu C, Jones C, Liu Z, Adamowicz-Brice M, Zhang J, 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]
  • 10.Puget S, Philippe C, Bax DA, Job B, Varlet P, Junier MP, et al. Mesenchymal transition and PDGFRA amplification/mutation are key distinct oncogenic events in pediatric diffuse intrinsic pontine gliomas. PLoS One. 2012;7:e30313. doi: 10.1371/journal.pone.0030313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Paugh BS, Zhu X, Qu C, Endersby R, Diaz AK, Zhang J, et al. Novel oncogenic PDGFRA mutations in pediatric high-grade gliomas. Cancer Res. 2013;73:6219–29. doi: 10.1158/0008-5472.CAN-13-1491. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Bjerke L, Mackay A, Nandhabalan M, Burford A, Jury A, Popov S, et al. Histone H3.3. mutations drive pediatric glioblastoma through upregulation of MYCN. Cancer Discov. 2013;3:512–9. doi: 10.1158/2159-8290.CD-12-0426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Lewis PW, Muller MM, Koletsky MS, Cordero F, Lin S, Banaszynski LA, 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]
  • 14.Chan KM, Fang D, Gan H, Hashizume R, Yu C, Schroeder M, et al. The histone H3.3K27M mutation in pediatric glioma reprograms H3K27 methylation and gene expression. Genes Dev. 2013;27:985–90. doi: 10.1101/gad.217778.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Bender S, Tang Y, Lindroth AM, Hovestadt V, Jones DT, Kool M, et al. Reduced H3K27me3 and DNA hypomethylation are major drivers of gene expression in K27M mutant pediatric high-grade gliomas. Cancer Cell. 2013;24:660–72. doi: 10.1016/j.ccr.2013.10.006. [DOI] [PubMed] [Google Scholar]
  • 16.Sturm D, Bender S, Jones DT, Lichter P, Grill J, Becher O, et al. Paediatric and adult glioblastoma: multiform (epi)genomic culprits emerge. Nat Rev Cancer. 2014;14:92–107. doi: 10.1038/nrc3655. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Paugh BS, Broniscer A, Qu C, Miller CP, Zhang J, Tatevossian RG, 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]
  • 18.von Werder A, Seidler B, Schmid RM, Schneider G, Saur D. Production of avian retroviruses and tissue-specific somatic retroviral gene transfer in vivo using the RCAS/TVA system. Nat Protoc. 2012;7:1167–83. doi: 10.1038/nprot.2012.060. [DOI] [PubMed] [Google Scholar]
  • 19.Barton KL, Misuraca K, Cordero F, Dobrikova E, Min HD, Gromeier M, et al. PD-0332991, a CDK4/6 inhibitor, significantly prolongs survival in a genetically engineered mouse model of brainstem glioma. PLoS One. 2013;8:e77639. doi: 10.1371/journal.pone.0077639. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Grasso CS, Tang Y, Truffaux N, Berlow NE, Liu L, Debily MA, et al. Functionally defined therapeutic targets in diffuse intrinsic pontine glioma. Nat Med. 2015;21:827. doi: 10.1038/nm0715-827a. [DOI] [PubMed] [Google Scholar]
  • 21.Hashizume R, Smirnov I, Liu S, Phillips JJ, Hyer J, McKnight TR, et al. Characterization of a diffuse intrinsic pontine glioma cell line: implications for future investigations and treatment. J Neurooncol. 2012;110:305–13. doi: 10.1007/s11060-012-0973-6. [DOI] [PubMed] [Google Scholar]
  • 22.Mohammad F, Weissmann S, Leblanc B, Pandey DP, Hojfeldt JW, Comet I, et al. EZH2 is a potential therapeutic target for H3K27M-mutant pediatric gliomas. Nat Med. 2017 doi: 10.1038/nm.4293. [DOI] [PubMed] [Google Scholar]
  • 23.Misuraca KL, Barton KL, Chung A, Diaz AK, Conway SJ, Corcoran DL, et al. Pax3 expression enhances PDGF-B-induced brainstem gliomagenesis and characterizes a subset of brainstem glioma. Acta Neuropathol Commun. 2014;2:134. doi: 10.1186/s40478-014-0134-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Flicek P, Amode MR, Barrell D, Beal K, Brent S, Carvalho-Silva D, et al. Ensembl 2012. Nucleic Acids Res. 2012;40:D84–90. doi: 10.1093/nar/gkr991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;11:R106. doi: 10.1186/gb-2010-11-10-r106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods. 2001;25:402–8. doi: 10.1006/meth.2001.1262. [DOI] [PubMed] [Google Scholar]
  • 27.Funato K, Major T, Lewis PW, Allis CD, Tabar V. Use of human embryonic stem cells to model pediatric gliomas with H3.3K27M histone mutation. Science. 2014;346:1529–33. doi: 10.1126/science.1253799. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Monje M, Mitra SS, Freret ME, Raveh TB, Kim J, Masek 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]
  • 29.Becher OJ, Hambardzumyan D, Walker TR, Helmy K, Nazarian J, Albrecht S, et al. Preclinical evaluation of radiation and perifosine in a genetically and histologically accurate model of brainstem glioma. Cancer Res. 2010;70:2548–57. doi: 10.1158/0008-5472.CAN-09-2503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM, et al. Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the phenotype and genotype of primary tumors than do serum-cultured cell lines. Cancer Cell. 2006;9:391–403. doi: 10.1016/j.ccr.2006.03.030. [DOI] [PubMed] [Google Scholar]
  • 31.Kotake Y, Cao R, Viatour P, Sage J, Zhang Y, Xiong Y. pRB family proteins are required for H3K27 trimethylation and Polycomb repression complexes binding to and silencing p16INK4alpha tumor suppressor gene. Genes Dev. 2007;21:49–54. doi: 10.1101/gad.1499407. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Jacobs JJ, Kieboom K, Marino S, DePinho RA, van Lohuizen M. The oncogene and Polycomb-group gene bmi-1 regulates cell proliferation and senescence through the ink4a locus. Nature. 1999;397:164–68. doi: 10.1038/16476. [DOI] [PubMed] [Google Scholar]
  • 33.Knutson SK, Kawano S, Minoshima Y, Warholic NM, Huang KC, Xiao Y, et al. Selective inhibition of EZH2 by EPZ-6438 leads to potent antitumor activity in EZH2-mutant non-Hodgkin lymphoma. Mol Cancer Ther. 2014;13:842–54. doi: 10.1158/1535-7163.MCT-13-0773. [DOI] [PubMed] [Google Scholar]
  • 34.Poirier JT, Gardner EE, Connis N, Moreira AL, de Stanchina E, Hann CL, et al. DNA methylation in small cell lung cancer defines distinct disease subtypes and correlates with high expression of EZH2. Oncogene. 2015;34:5869–78. doi: 10.1038/onc.2015.38. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Vire E, Brenner C, Deplus R, Blanchon L, Fraga M, Didelot C, et al. The Polycomb group protein EZH2 directly controls DNA methylation. Nature. 2006;439:871–4. doi: 10.1038/nature04431. [DOI] [PubMed] [Google Scholar]
  • 36.Harbour JW, Luo RX, Dei Santi A, Postigo AA, Dean DC. Cdk phosphorylation triggers sequential intramolecular interactions that progressively block Rb functions as cells move through G1. Cell. 1999;98:859–69. doi: 10.1016/s0092-8674(00)81519-6. [DOI] [PubMed] [Google Scholar]
  • 37.Flaherty KT, Lorusso PM, Demichele A, Abramson VG, Courtney R, Randolph SS, et al. Phase I, dose-escalation trial of the oral cyclin-dependent kinase 4/6 inhibitor PD 0332991, administered using a 21-day schedule in patients with advanced cancer. Clin Cancer Res. 2012;18:568–76. doi: 10.1158/1078-0432.CCR-11-0509. [DOI] [PubMed] [Google Scholar]
  • 38.Turner NC, Huang Bartlett C, Cristofanilli M. Palbociclib in Hormone-Receptor-Positive Advanced Breast Cancer. N Engl J Med. 2015;373:1672–3. doi: 10.1056/NEJMc1510345. [DOI] [PubMed] [Google Scholar]
  • 39.Buczkowicz P, Hoeman C, Rakopoulos P, Pajovic S, Letourneau L, Dzamba M, et al. Genomic analysis of diffuse intrinsic pontine gliomas identifies three molecular subgroups and recurrent activating ACVR1 mutations. Nat Genet. 2014;46:451–6. doi: 10.1038/ng.2936. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Taylor KR, Mackay A, Truffaux N, Butterfield YS, Morozova O, Philippe C, et al. Recurrent activating ACVR1 mutations in diffuse intrinsic pontine glioma. Nat Genet. 2014;46:457–61. doi: 10.1038/ng.2925. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Brennan CW, Verhaak RG, McKenna A, Campos B, Noushmehr H, Salama SR, et al. The somatic genomic landscape of glioblastoma. Cell. 2013;155:462–77. doi: 10.1016/j.cell.2013.09.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Wiese M, Schill F, Sturm D, Pfister S, Hulleman E, Johnsen SA, et al. No Significant Cytotoxic Effect of the EZH2 Inhibitor Tazemetostat (EPZ-6438) on Pediatric Glioma Cells with Wildtype Histone 3 or Mutated Histone 3.3. Klin Padiatr. 2016;228:113–7. doi: 10.1055/s-0042-105292. [DOI] [PubMed] [Google Scholar]
  • 43.Schlesinger Y, Straussman R, Keshet I, Farkash S, Hecht M, Zimmerman J, et al. Polycomb-mediated methylation on Lys27 of histone H3 pre-marks genes for de novo methylation in cancer. Nat Genet. 2007;39:232–6. doi: 10.1038/ng1950. [DOI] [PubMed] [Google Scholar]
  • 44.Sharpless NE, Bardeesy N, Lee KH, Carrasco D, Castrillon DH, Aguirre AJ, et al. Loss of p16Ink4a with retention of p19Arf predisposes mice to tumorigenesis. Nature. 2001;413:86–91. doi: 10.1038/35092592. [DOI] [PubMed] [Google Scholar]
  • 45.Gal-Yam EN, Egger G, Iniguez L, Holster H, Einarsson S, Zhang X, et al. Frequent switching of Polycomb repressive marks and DNA hypermethylation in the PC3 prostate cancer cell line. Proc Natl Acad Sci U S A. 2008;105:12979–84. doi: 10.1073/pnas.0806437105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Keshet I, Schlesinger Y, Farkash S, Rand E, Hecht M, Segal E, et al. Evidence for an instructive mechanism of de novo methylation in cancer cells. Nat Genet. 2006;38:149–53. doi: 10.1038/ng1719. [DOI] [PubMed] [Google Scholar]
  • 47.Ohm JE, McGarvey KM, Yu X, Cheng L, Schuebel KE, Cope L, et al. A stem cell-like chromatin pattern may predispose tumor suppressor genes to DNA hypermethylation and heritable silencing. Nat Genet. 2007;39:237–42. doi: 10.1038/ng1972. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Widschwendter M, Fiegl H, Egle D, Mueller-Holzner E, Spizzo G, Marth C, et al. Epigenetic stem cell signature in cancer. Nat Genet. 2007;39:157–8. doi: 10.1038/ng1941. [DOI] [PubMed] [Google Scholar]
  • 49.Thienpont B, Steinbacher J, Zhao H, D’Anna F, Kuchnio A, Ploumakis A, et al. Tumour hypoxia causes DNA hypermethylation by reducing TET activity. Nature. 2016;537:63–8. doi: 10.1038/nature19081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Justin N, Zhang Y, Tarricone C, Martin SR, Chen S, Underwood E, et al. Structural basis of oncogenic histone H3K27M inhibition of human polycomb repressive complex 2. Nat Commun. 2016;7:11316. doi: 10.1038/ncomms11316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Jiao L, Liu X. Structural basis of histone H3K27 trimethylation by an active polycomb repressive complex 2. Science. 2015;350:aac4383. doi: 10.1126/science.aac4383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Jackson M, Krassowska A, Gilbert N, Chevassut T, Forrester L, Ansell J, et al. Severe global DNA hypomethylation blocks differentiation and induces histone hyperacetylation in embryonic stem cells. Mol Cell Biol. 2004;24:8862–71. doi: 10.1128/MCB.24.20.8862-8871.2004. [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

1
NIHMS878484-supplement-1.tif (3.5MB, tif)
2
NIHMS878484-supplement-2.tif (2.8MB, tif)
3
NIHMS878484-supplement-3.tif (2.5MB, tif)
4
NIHMS878484-supplement-4.tif (2.2MB, tif)
5
NIHMS878484-supplement-5.docx (114.3KB, docx)
6
NIHMS878484-supplement-6.docx (108.9KB, docx)
7
NIHMS878484-supplement-7.pdf (36.3KB, pdf)
8
NIHMS878484-supplement-8.pdf (33.3KB, pdf)
9
NIHMS878484-supplement-9.pdf (109.9KB, pdf)

RESOURCES