GENE-19. DEEP PROTEOMIC SURVEY ACROSS SEVEN CHILDHOOD BRAIN TUMORS. (23rd April 2019)

• Record Type: Journal Article
• Authors: Wang, Pei; Tignor, Nicole; Gygi, Steven; Ivey, Richard; Paulovich, Amanda; Nescizhskii, Alexey; Thakurta, Sanjukta Guha; Whiteaker, Jeffery; Kennedy, Jacob; Voytovich, Uliana; Ding, Li; Wang, Liang Bo; Raman, Pichai; Zhu, Yuankun; Hiltke, Tara; Resnick, Adam; Rood, Brian
• Is Part Of: Neuro-oncology. Volume 21(2019)Supplement 2
• Page Start: ii85
• Page End: ii85
• Publication Date: 2019-04-23
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/noz036.090
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: Genomic characterization has allowed for the differentiation of different tumor types based upon the abundance of gene transcripts. However, owing to the many layers of regulation between transcript and the post-translationally modified protein, the functional moiety of the cell, it has been challenging to extrapolate biology from these transcriptional differences. We hypothesized that a comparative analysis of the proteome and phosphoproteome across 7 childhood brain tumors would yield a deeper understanding of the differences in their functional biology. We performed tandem mass tag labeling and triple mass spectrometry of 226 fresh frozen tumor samples collected in a single institution representing the histologic diagnoses of: high grade astrocytoma (27), low grade astrocytoma (97), ganglioglioma (20), ependymoma (32), medulloblastoma (22), atypical teratoid rhabdoid tumor (12), and craniopharyngioma (16). Among these samples were 22 pairs from pre/post recurrence. Across this sample set, we quantified 9155 proteins and 13632 phospho sites. After data preprocessing including normalization, batch correction and missing value imputation, we performed consensus clustering and showed that protein profiles can effectively distinguish major histology types. Additional regression based differential analyses revealed groups of proteins and pathways showing distinct activity patterns across different histologic types. Further leveraging the WGS data, we characterized the functional impact of mutation and fusion features of the glial tumors. Specifically, focusing on the largest cohort, low grade glioma, we analyzed the proteomic ramifications of BRAF status: V600E, BRAF-fusion and wild type. Moreover, from the primary and recurrent tumor pairs, we discovered the upregulation of proteins associated with immune evasion in more advanced LGG tumors. The incorporation of the proteomic dimension into large scale efforts at tumor characterization adds functional insight that can help drive translational efforts.