Acta Neuropathologica

, Volume 127, Issue 6, pp 881–895

Comparative multidimensional molecular analyses of pediatric diffuse intrinsic pontine glioma reveals distinct molecular subtypes

Authors

  • Amanda M. Saratsis
    • Division of NeurosurgeryAnn & Robert H. Lurie Children’s Hospital of Chicago
    • Center for Genetic Medicine Research, Children’s National Medical Center
  • Madhuri Kambhampati
    • Center for Genetic Medicine Research, Children’s National Medical Center
  • Kendall Snyder
    • Center for Genetic Medicine Research, Children’s National Medical Center
  • Sridevi Yadavilli
    • Center for Genetic Medicine Research, Children’s National Medical Center
  • Joseph M. Devaney
    • Center for Genetic Medicine Research, Children’s National Medical Center
    • Department of Integrative Systems BiologyGeorge Washington University School of Medicine and Health Sciences
  • Brennan Harmon
    • Center for Genetic Medicine Research, Children’s National Medical Center
  • Jordan Hall
    • Center for Genetic Medicine Research, Children’s National Medical Center
  • Eric H. Raabe
    • Division of Neuro-PathologyJohns Hopkins University School of Medicine
    • Division of Pediatric OncologyJohns Hopkins University School of Medicine
  • Ping An
    • Division of Neuro-PathologyJohns Hopkins University School of Medicine
    • Neurobiology DepartmentCollege of Basic Medical Sciences, China Medical University
  • Melanie Weingart
    • Division of Neuro-PathologyJohns Hopkins University School of Medicine
  • Brian R. Rood
    • Division of OncologyCenter for Cancer and Immunology Research, Children’s National Medical Center
  • Suresh N. Magge
    • Division of NeurosurgeryChildren’s National Medical Center
  • Tobey J. MacDonald
    • Department of PediatricsEmory University School of Medicine
  • Roger J. Packer
    • Division of NeurologyCenter for Neuroscience Research, Children’s National Medical Center
    • Brain Tumor Institute, Center for Neuroscience and Behavioral Medicine, Children’s National Medical Center
    • Center for Genetic Medicine Research, Children’s National Medical Center
    • Department of Integrative Systems BiologyGeorge Washington University School of Medicine and Health Sciences
Original Paper

DOI: 10.1007/s00401-013-1218-2

Cite this article as:
Saratsis, A.M., Kambhampati, M., Snyder, K. et al. Acta Neuropathol (2014) 127: 881. doi:10.1007/s00401-013-1218-2

Abstract

Diffuse intrinsic pontine glioma (DIPG) is a highly morbid form of pediatric brainstem glioma. Here, we present the first comprehensive protein, mRNA, and methylation profiles of fresh-frozen DIPG specimens (n = 14), normal brain tissue (n = 10), and other pediatric brain tumors (n = 17). Protein profiling identified 2,305 unique proteins indicating distinct DIPG protein expression patterns compared to other pediatric brain tumors. Western blot and immunohistochemistry validated upregulation of Clusterin (CLU), Elongation Factor 2 (EF2), and Talin-1 (TLN1) in DIPGs studied. Comparisons to mRNA expression profiles generated from tumor and adjacent normal brain tissue indicated two DIPG subgroups, characterized by upregulation of Myc (N-Myc) or Hedgehog (Hh) signaling. We validated upregulation of PTCH, a membrane receptor in the Hh signaling pathway, in a subgroup of DIPG specimens. DNA methylation analysis indicated global hypomethylation of DIPG compared to adjacent normal tissue specimens, with differential methylation of 24 genes involved in Hh and Myc pathways, correlating with protein and mRNA expression patterns. Sequencing analysis showed c.83A>T mutations in the H3F3A or HIST1H3B gene in 77 % of our DIPG cohort. Supervised analysis revealed a unique methylation pattern in mutated specimens compared to the wild-type DIPG samples. This study presents the first comprehensive multidimensional protein, mRNA, and methylation profiling of pediatric brain tumor specimens, detecting the presence of two subgroups within our DIPG cohort. This multidimensional analysis of DIPG provides increased analytical power to more fully explore molecular signatures of DIPGs, with implications for evaluating potential molecular subtypes and biomarker discovery for assessing response to therapy.

Keywords

Diffuse intrinsic pontine glioma (DIPG)Brainstem gliomaProteomicsHistone H3Myc oncogeneHedgehog

Supplementary material

401_2013_1218_MOESM1_ESM.pdf (333 kb)
Online Resource 1 Pediatric Brain Tumor Tissue Protein Expression Profiles. Principal component analysis (PCA) of tumor protein profiles (fold change values of protein expression in tumor compared to average expression across all normal brain tissue) demonstrates independent clustering of DIPG tumor specimens (navy blue) from other tumor types (PDF 332 kb)
401_2013_1218_MOESM2_ESM.pdf (6.1 mb)
Online Resource 2 Pediatric Brain Tumor Tissue Protein Expression Profiles. Protein expression profiles of brain tumor tissue specimens (n=31), reported as fold change in expression for each tumor vs. average expression across all normal tissue specimens (PDF 6242 kb)
401_2013_1218_MOESM3_ESM.pdf (2.2 mb)
Online Resource 3 Pairwise DIPG Tissue Protein Expression Profiles. Tissue protein expression profiles of DIPG patients (n=10), reported as fold change in expression for each tumor vs. normal specimen pair (PDF 2252 kb)
401_2013_1218_MOESM4_ESM.pdf (227 kb)
Online Resource 4 Comparative Analysis of DIPG Protein Expression Profiles. Filtered ANOVA results (p-value<0.05 and FC -2 or 2) comparing protein expression profiles of DIPGs by molecular subgroup (pairwise tumor vs. normal fold change values), reported as fold change expression values in Subgroup 1 vs. Subgroup 2 (PDF 227 kb)
401_2013_1218_MOESM5_ESM.pdf (401 kb)
Online Resource 5 Comparison of DIPG CSF, Fresh Frozen, and Formalin Fixed Paraffin Embedded (FFPE) Tissue Proteomes. Listing of detected proteins (Swiss-Prot IDs) in DIPG CSF, Fresh Frozen and FFPE Tissue, with overlap between datasets (PDF 401 kb)
401_2013_1218_MOESM6_ESM.pdf (208 kb)
Online Resource 6 Spectral Representation of TLN1, EF2 and CLU. Tissue protein profiles of specimens from Patient IDs 6 and 7 were inspected for the presence of spectra representing TLN1, EF2 and CLU. Each spectra representing the parent ion (MS) was further searched to confirm the presence of identified peptide in the amino acid sequence (MS/MS). Each spectra represents one positively identified peptide of a given protein as identified by b and y ions listed in accompanying tables (PDF 207 kb)
401_2013_1218_MOESM7_ESM.pdf (1.1 mb)
Online Resource 7 Pediatric Brain Tumor Tissue mRNA Expression Profiles. mRNA expression profiles of brain tumor tissue specimens (n=28), reported as fold change in expression for each tumor vs. average expression across all normal tissue specimens (ANOVA, p-value Dx <0.05 and FC -2 or 2) (PDF 1088 kb)
401_2013_1218_MOESM8_ESM.pdf (257 kb)
Online Resource 8 Unsupervised clustering of Tumor mRNA profiles. Unsupervised hierarchical clustering of tumor tissue mRNA expression profiles (fold change values of mRNA expression in tumor compared to average expression across all normal brain tissue specimens) identified differential expression of 1,249 genes. DIPGs cluster independently from other tumor types but closest to supratentorial astrocytomas (ANOVA, p-value<0.05 and FC-2 or 2): (*) H3.3 K27M, (**) H3.1 K27M, ( ) H3 wild type (PDF 257 kb)
401_2013_1218_MOESM9_ESM.pdf (260 kb)
Online resource 9 Unsupervised hierarchical clustering of high-grade astrocytoma mRNA expression values (tumor vs. average normal FC values: DIPG n=13, Yellow; pediatric supratentorial GBM n=4, Green) demonstrate unique profiles by tumor location and Histone 3 K27M mutation status (H3.1 Red, H3.3 Orange, H3 wild type Gray) (PDF 260 kb)
401_2013_1218_MOESM10_ESM.pdf (3 mb)
Online Resource 10 Pairwise DIPG Tissue mRNA Expression Profiles (Part 1 of 3). mRNA expression profiles of DIPG tissue (n=8), reported as fold change in expression for each tumor vs. normal specimen pair (PDF 3109 kb)
401_2013_1218_MOESM11_ESM.pdf (3.1 mb)
Online Resource 10 Pairwise DIPG Tissue mRNA Expression Profiles (Part 2 of 3). mRNA expression profiles of DIPG tissue (n=8), reported as fold change in expression for each tumor vs. normal specimen pair (PDF 3200 kb)
401_2013_1218_MOESM12_ESM.pdf (2.9 mb)
Online Resource 10 Pairwise DIPG Tissue mRNA Expression Profiles (Part 3 of 3). mRNA expression profiles of DIPG tissue (n=8), reported as fold change in expression for each tumor vs. normal specimen pair (PDF 2924 kb)
401_2013_1218_MOESM13_ESM.pdf (2.9 mb)
Online Resource 11 Comparative Analysis of DIPG mRNA Expression Profiles. Comparison of mRNA expression profiles (pairwise tumor vs. normal fold change values) by molecular subgroups revealed through unsupervised clustering, reported as fold change expression values in Subgroup 1 vs. Subgroup 2 (PDF 2965 kb)
401_2013_1218_MOESM14_ESM.pdf (223 kb)
Online Resource 12 Unsupervised clustering of DIPG mRNA profiles. Unsupervised hierarchical clustering of DIPG tissue mRNA expression profiles revealed two distinct DIPG subgroups, Subgroup 1 (red) and Subgroup 2 (blue), with differential expression of 158 genes (ANOVA, p- value<0.05, FC<-2 or >2): (*) H3.3 K27M, (**) H3.1 K27M, ( ) H3 wild type (PDF 223 kb)
401_2013_1218_MOESM15_ESM.pdf (217 kb)
Online Resource 13 Role of TLN1 and CLU in the Canonical Actin Cytoskeletal Signaling Pathway. Functional pathways analysis of DIPG tissue protein profiles revealed Actin Cytoskeletal Signaling as the top pathway of interaction. Dense interaction with validated tumor tissue proteins Clusterin (CLU) and Talin-1 (TLN1) is observed. Protein expression levels are depicted as up- (red) or down- (green) regulated fold change values (tumor vs. normal) (PDF 217 kb)
401_2013_1218_MOESM16_ESM.pdf (156 kb)
Online Resource 14 Differential Expression of Genes involved in Myc and Hh Pathways Identified through Functional Pathway Analysis of DIPG Subgroups. Functional pathway analysis of gene expression profiles from DIPG and normal tissue implicated differential expression of genes related to Myc and Hh signaling pathways between DIPG Subgroup 1 and 2. Gene expression fold changes are represented as average of pairwise tumor vs. normal expression values across specimens in a given subgroup, listed in parentheses following each gene (PDF 155 kb)
401_2013_1218_MOESM17_ESM.pdf (375 kb)
Online Resource 15 Functional Analysis of Molecular Profiles Reveals Differential expression of Myc and Hh Signaling Pathways between DIPG Subgroups.Top Panel Comparative functional analysis of DIPG tumor tissue molecular profiles revealed differential activity of regulatory molecules Myc (top) and GLI1 (Hh pathway molecule, bottom) between Subgroup 1 and 2. Gene expression in respective molecular networks depicted as fold change values (pairwise tumor vs. normal) in a Myc (PID 3) and Hh patient (PID 13). Heat map representation of mRNA dysregulation observed in Myc (top) and Hh (bottom) networks are depicted. Bottom Panel Hedgehog (Hh) pathway analysis depicts mRNA dysregulation in a Hh patient (PID 13). mRNA expression levels are depicted as up- (red) or down- (green) regulated fold change values (tumor vs. normal)(PDF 375 kb)
401_2013_1218_MOESM18_ESM.pdf (211 kb)
Online Resource 16 Sanger Sequencing of H3F3A and HIST1H3B Mutation in DIPG Specimens. Sanger sequencing chromatogram demonstrating H3F3A and HIST1H3B mutation encoding pLys27Met (K27M) substitution in DIPG tumor tissue. (*) indicates A→T substitution in mutant specimens (PDF 210 kb)
401_2013_1218_MOESM19_ESM.pdf (241 kb)
Online Resource 17 Supervised hierarchical clustering of DIPG protein expression profiles (n=14, tumor vs. average normal FC values) revealed significantly 112 differentially expressed proteins by Histone 3 K27M mutation status (ANOVA, p-value<0.05 and FC<-2 or >2, H3.1: Red, H3.3: Orange, H3 wild type: Gray) (PDF 240 kb)
401_2013_1218_MOESM20_ESM.pdf (371 kb)
Online Resource 18 Principle component representation of high-grade astrocytoma mRNA expression profiles (DIPG n=13: yellow, pediatric supratentorial GBM n=3: green, tumor vs. average normal FC values) representing 2,274 significantly differentially expressed molecules detected on supervised analysis by tumor location and demonstrating unsupervised profile clustering by Histone 3 K27M mutation status (ANOVA, p-value<0.05 and FC<-2 or >2, H3.1 Red, H3.3 Orange, H3 wild type Gray) (PDF 371 kb)
401_2013_1218_MOESM21_ESM.pdf (341 kb)
Online Resource 19 Differential Patterns of mRNA Expression in DIPG tissue based on H3.3 K27M Mutation Status. Top Panel Supervised clustering of pairwise mRNA profiles by H3.3 mutation status, demonstrating 345 differentally expressed genes between mutant (n=5) and wild type (n=4) patients (ANOVA, p value <0.05, FC>-2 or <2). H3.1 mutation denoted by (*). Bottom Panel Pathways analysis of dysregulated genes identified in analysis above (Top Panel). mRNA expression levels are provided in parentheses as up- (red) and down- (green) regulated fold change values (mutant vs. wild type)(PDF 341 kb)
401_2013_1218_MOESM22_ESM.pdf (178 kb)
Online Resource 20 Functional Analysis of Gene Dysregulation Based on Histone H3.3 K27M Status. Top networks of molecular interaction and biological function were identified via functional pathway analysis of differentially expressed genes in DIPG by H3.3 K27M status. mRNA expression levels are provided in parentheses as fold change values (pairwise tumor vs. normal) (PDF 178 kb)
401_2013_1218_MOESM23_ESM.pdf (300 kb)
Online Resource 21 Immunohistochemical Staining of Adult GBM Tissue for ATRX. Adult GBM tissue was stained as a positive control for ATRX expression. Scale bar = 100μm) (PDF 300 kb)
401_2013_1218_MOESM24_ESM.pdf (18.1 mb)
Online Resource 22 DIPG Tissue DNA Methylation Profiles. DNA methylation profiles of DIPG patients (n=9), reported as fold change in methylation in tumor vs. normal tissue (ANOVA, p-value with FDR<0.05 and FC <-3 or >3) (PDF 18576 kb)
401_2013_1218_MOESM25_ESM.pdf (39.8 mb)
Online Resource 23 Pairwise DIPG Tissue DNA Methylation Profiles (Part 1 of 2). DNA methylation profiles of DIPG patients (n=9), reported as fold change in methylation for each tumor vs. normal specimen pair (PDF 40708 kb)
401_2013_1218_MOESM26_ESM.pdf (36.5 mb)
Online Resource 23 Pairwise DIPG Tissue DNA Methylation Profiles (Part 2 of 2). DNA methylation profiles of DIPG patients (n=9), reported as fold change in methylation for each tumor vs. normal specimen pair (PDF 37359 kb)
401_2013_1218_MOESM27_ESM.pdf (920 kb)
Online Resource 24 Comparative Analysis of DIPG DNA Methylation Profiles by Molecular Subgroup. Filtered ANOVA results ( p-value with FDR <0.05 and FC <-2 or >2) of DNA methylation profiles (pairwise tumor vs. normal fold change values) compared by DIPG subgroup, reported as fold change methylation values in Subgroup 1 vs. Subgroup 2 (PDF 920 kb)
401_2013_1218_MOESM28_ESM.pdf (176 kb)
Online Resource 25 Patterns of DNA Methylation Supervised by DIPG Subgroups. Supervised comparison of genome-wide methylation profiles between DIPG Subgroup 1 and 2 revealed 786 differentially methylated loci. Top canonical pathways (top section) and networks of molecular interactions (bottom section) were identified, with differentially methylated genes between the two subgroups. Fold changes are listed in parentheses following each gene (PDF 176 kb)
401_2013_1218_MOESM29_ESM.pdf (201 kb)
Online Resource 26 Differential DNA Methylation of Genes Involved in Myc and Hh Pathways Between DIPG Subgroups. A total of 29 differentially methylated genes pathways between DIPG subgroups 1 and 2 were identified through unsupervised analysis of methylation status of 486 CpG loci related to Myc and Hh signaling. DNA methylation (CpG) sites for each gene are presented along with gene (column 3) and protein (column 4) expression values. For methylation analysis, we used average pairwise tumor vs. normal fold change values in Subgroup 1 relative to Subgroup 2; for gene and protein expressions, we used average pairwise tumor vs. normal fold change values by subgroup. Fold changes are listed in parentheses following each gene (PDF 201 kb)
401_2013_1218_MOESM30_ESM.pdf (1.5 mb)
Online Resource 27 Comparative Analysis of Tissue DNA Methylation Profiles by H3.3 Mutation Status. Filtered ANOVA results (p-value<0.05 and FC<-3 or >3) of DIPG DNA methylation profiles compared by H3.3 K27M mutation status, reported as fold change methylation values in H3.3 mutant vs. wild type patients (PDF 1555 kb)
401_2013_1218_MOESM31_ESM.pdf (320 kb)
Online Resource 28 Differential Patterns of DNA Methylation in DIPG Tissue by H3.3 K27M Mutation Status. Supervised analysis of DNA methylation profiles based on H3.3 K27M mutation status (n=9 pairs, pairwise tumor vs. normal fold change values) revealed 645 differentially methylated loci (ANOVA, p-value <0.05 and FC <-3 or >3). Genome wide mapping of differences in DNA methylation patterns between H3.3 K27M mutants and wild type tissue specimens is depicted. X-axis=Chromosomes 1-Y. Y-axis=Fold change in DNA methylation of identified gene across all related CpG sites in mutant relative to wild type tissue (PDF 320 kb)
401_2013_1218_MOESM32_ESM.pdf (785 kb)
Online Resource 29 Comparative Analysis of Pairwise DIPG DNA Methylation Profiles by H3.3 Mutation Status. Filtered ANOVA results (p-value<0.05 and FC<-3 or >3) of pairwise DIPG DNA methylation profiles (pairwise tumor vs. normal fold change values) compared by H3.3 K27Mmutation status, reported as fold change methylation values in H3.3 mutant vs. wild type patients (PDF 784 kb)
401_2013_1218_MOESM33_ESM.pdf (291 kb)
Online Resource 30 Supervised Clustering of Methylation in DIPG tissue based on H3F3A Mutation Status. Top Panel Supervised analysis of DNA methylation profiles based on H3.3 mutation status revealed 645 differentially methylated loci. The heat map depicted represents the 212 loci with similar methylation pattern in all specimens based on H3.3 mutation status. H3.1 Mutant denoted by (*). Bottom Panel Functional analysis of loci hypomethylated in all H3.3 mutants mapped to cell death inhibition (see Online Resource 31). DNA methylation, listed in parentheses as fold change values (pairwise tumor vs. normal) in mutant specimens relative to wild type specimens, and p-value are provided for each molecule (PDF 290 kb)
401_2013_1218_MOESM34_ESM.pdf (174 kb)
Online Resource 31 Patterns of DNA Methylation in DIPG Supervised by H3.3 K27M Mutation Status. Pathway analysis of supervised H3.3 mutation status data identified top biological functions and molecular networks of interaction. Fold changes are listed in parentheses following each gene (PDF 174 kb)

Copyright information

© Springer-Verlag Berlin Heidelberg 2013