| Literature DB >> 30877103 |
Mary Hoffman1,2,3, Aaron H Gillmor1,2,3, Daniel J Kunz4,5,6, Michael J Johnston2,3, Ana Nikolic1,2,3, Kiran Narta1,2,3, Mehdi Zarrei7,8,9, Jennifer King2,3, Katrina Ellestad2,3, Ngoc Ha Dang2,3, Florence M G Cavalli10, Michelle M Kushida10, Fiona J Coutinho10, Yuankun Zhu11, Betty Luu10, Yussanne Ma12, Andrew J Mungall12, Richard Moore12, Marco A Marra12, Michael D Taylor10, Trevor J Pugh13, Peter B Dirks8,10, Douglas Strother14, Lucie Lafay-Cousin14, Adam C Resnick11, Stephen Scherer7,8,9, Donna L Senger2,3,15, Benjamin D Simons4,5,16, Jennifer A Chan2,3,17, A Sorana Morrissy1,2,3, Marco Gallo18,2,3,19.
Abstract
Pediatric glioblastoma (pGBM) is a lethal cancer with no effective therapies. To understand the mechanisms of tumor evolution in this cancer, we performed whole-genome sequencing with linked reads on longitudinally resected pGBM samples. Our analyses showed that all diagnostic and recurrent samples were collections of genetically diverse subclones. Clonal composition rapidly evolved at recurrence, with less than 8% of nonsynonymous single-nucleotide variants being shared in diagnostic-recurrent pairs. To track the origins of the mutational events observed in pGBM, we generated whole-genome datasets for two patients and their parents. These trios showed that genetic variants could be (i) somatic, (ii) inherited from a healthy parent, or (iii) de novo in the germlines of pGBM patients. Analysis of variant allele frequencies supported a model of tumor growth involving slow-cycling cancer stem cells that give rise to fast-proliferating progenitor-like cells and to nondividing cells. Interestingly, radiation and antimitotic chemotherapeutics did not increase overall tumor burden upon recurrence. These findings support an important role for slow-cycling stem cell populations in contributing to recurrences, because slow-cycling cell populations are expected to be less prone to genotoxic stress induced by these treatments and therefore would accumulate few mutations. Our results highlight the need for new targeted treatments that account for the complex functional hierarchies and genomic heterogeneity of pGBM. SIGNIFICANCE: This work challenges several assumptions regarding the genetic organization of pediatric GBM and highlights mutagenic programs that start during early prenatal development.Graphical Abstract: http://cancerres.aacrjournals.org/content/canres/79/9/2111/F1.large.jpg. ©2019 American Association for Cancer Research.Entities:
Mesh:
Substances:
Year: 2019 PMID: 30877103 PMCID: PMC7282886 DOI: 10.1158/0008-5472.CAN-18-3441
Source DB: PubMed Journal: Cancer Res ISSN: 0008-5472 Impact factor: 12.701