Open AccessTBIO-19. INTEGRATED GENOMIC, PROTEOMIC AND PHOSPHOPROTEOMIC ANALYSIS OF SEVEN TYPES OF PEDIATRIC BRAIN CANCER
Author(s) -
Francesca Petralia,
Nicole Tignor,
Dmitri Rykunov,
Boris Revas,
Shrabanti Chowdhury,
Azra Krek,
Pichae Raman,
Jiayi Ji,
Yuankun Zhu,
Wanli Ma,
Xiaoyu Song,
David Fenyö,
Steven P. Gygi,
Richard G. Ivey,
Antonio Iavarone,
Jeffrey R. Whiteaker,
Antonio Colaprico,
Alexey I. Nesvizhskii,
Henry Rodriguez,
Amanda G. Paulovich,
Tara Hiltke,
Adam Resnick,
Pei Wang,
Brian R. Rood
Publication year - 2020
Publication title -
neuro-oncology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 4.005
H-Index - 125
eISSN - 1523-5866
pISSN - 1522-8517
DOI - 10.1093/neuonc/noaa222.846
Subject(s) - proteomics , biology , proteome , atypical teratoid rhabdoid tumor , computational biology , phosphoproteomics , ependymoma , astrocytoma , proteogenomics , brain tumor , genomics , glioma , cancer research , bioinformatics , genome , medulloblastoma , pathology , kinase , medicine , genetics , protein kinase a , protein phosphorylation , gene
We performed a comprehensive proteogenomic analysis across seven childhood brain tumors for a deeper understanding of their functional biology. Whole genome sequencing, RNAseq, quantitative proteomic profiling and phosphoproteomics were performed on 219 fresh frozen tumor samples representing the histologic diagnoses of: low grade astrocytoma (93), ependymoma (32), high grade astrocytoma (26), medulloblastoma (22), ganglioglioma (18), craniopharyngioma (16) and atypical teratoid rhabdoid tumor (12). Unsupervised clustering analysis based on proteomics data reveals eight clusters with distinct protein profiles and pathway activities. While some clusters coincide with histologic diagnoses, a couple of clusters appear to be a mixture of different diagnoses, including one cluster consisting of “aggressive” tumors characterized by poor survival and high stemness scores. By integrating proteomic data with RNAseq and WGS data, we characterize the impact of mutations (H3K27M, BRAFV600E, BRAF fusion) and CNVs upon the proteome across various diagnoses. Multiomics based kinase-substrate association analysis and co-expression network analysis reveal targetable active kinase networks within these tumors. Proteomic data reveals unique biology associated with H3K27M mutation status in HGG and BRAF aberrations in LGG. Characterization of the tumor microenvironment through deconvolution analyses based on multi-omics data reveals 5 distinct tumor clusters associated with different populations of infiltrating immune cells and the relative activity of the immune system based upon the expression of pro-inflammation or immunosuppressive markers. This study reports the first large-scale deep comprehensive proteogenomic analysis crossing traditional histologic boundaries to uncover foundational pediatric brain tumor biology including functional insight that helps drive translational efforts.