Transposon Mutagenesis Identifies Genes That Transform Neural Stem
Total Page:16
File Type:pdf, Size:1020Kb
Transposon mutagenesis identifies genes that transform neural stem cells into glioma-initiating cells Hideto Kosoa,b, Haruna Takedaa,c, Christopher Chin Kuan Yewa, Jerrold M. Warda, Naoki Nariaid, Kazuko Uenod, Masao Nagasakid, Sumiko Watanabeb, Alistair G. Ruste, David J. Adamse, Neal G. Copelanda,f, and Nancy A. Jenkinsa,f,1 aDivision of Genetics and Genomics, Institute of Molecular and Cell Biology, Agency for Science, Technology and Research, Singapore 138673; bDivision of Molecular and Developmental Biology, Institute of Medical Science, University of Tokyo, Tokyo 108-8639, Japan; cDepartment of Microbiology, University of Tokyo, Tokyo 113-0033, Japan; dDepartment of Integrative Genomics, Tohoku Medical Megabank Organization, Tohoku University, Sendai 980-8579, Japan; eExperimental Cancer Genetics, Wellcome Trust Sanger Institute, Hinxton, Cambridge CB10 1HH, United Kingdom; and fCancer Research Program, Methodist Hospital Research Institute, Houston, TX 77030 AUTHOR SUMMARY Glioblastoma multiforme (GBM) is the by random chance, and therefore are most common form of malignant brain most likely to harbor genes responsible cancer in adults, whose mean survival is for cancer progression. We identified 1 y. Evidence indicates the presence of 140 and 148 CIS genes in immortalized self-renewing, stem-like cells within hu- cells and tumors, respectively (Fig. P1). man gliomas (1). Neural stem cells Thirty-four genes were common to (NSCs) are considered the cell of origin both immortalized cells and tumors, of GBM, the normal cell that acquires the and the rest were specific to either, first glioma-promoting mutation(s), suggesting different mechanisms for im- because the adult brain has very few mortalization and tumor formation. proliferating cells capable of accumulat- Analysis of CIS genes from immortalized ing mutations required for gliomagenesis. lines identified signaling pathways that To identify these genetic alterations, we involved cytoskeletal organization and used mobile DNA segments called transcriptional regulation, which play transposons that can affect gene function key roles in regulating self-renewal by inserting in or near genes, thus pro- vs. differentiation of NSCs. In contrast, Fig. P1. Transformation of NSCs into cancer- viding an unbiased, high-throughput initiating cells for mesenchymal GBM. Immor- 34 CIS genes, which were common, method for identifying genes important talization-CIS genes (e.g., Gli3, Nf1) promote enriched genes associated with mitosis for cancer (2). Here, we used this ap- immortalization of NSCs. Immortalized astroglial and cell division. proach to identify genes and signaling cells show a large repertoire of insertions re- We also identified genes specificto pathways that are able to transform NSCs presented by different colors. Transposons are tumors. Analysis of these genes associ- into cancer-initiating cells for GBM. continuously jumping to other sites in the genome ated them with the receptor tyrosine The mutagenic Sleeping Beauty (SB) after transplantation, and transplanted cells kinase (RTK) signaling pathway, which transposons, which were designed to elicit become cancer-initiating cells by randomly ac- is the most frequently mutated pathway loss-of-function mutations as well as gain- quiring new insertions in tumor CIS genes (e.g., identified in human GBM. The proto- Met, Pdgfrb, Gab1) and then clonally expand to of-function mutations, were mobilized oncogene Met and the tumor suppressor fi form tumors with characteristics of mesenchymal speci cally in the NSC compartment by GBM. Each tumor showed a distinct combination of gene Nf1 are most frequently mutated using genetically engineered mice (3). CIS genes even though all tumors were derived in the mouse tumors. Frequent NF1 NSCs were expanded in vitro, induced to from a single line. Most of the immortalization CIS mutations and increased MET expres- differentiate, and serially passaged to se- genes, with the exception of 34 genes, are lost sion are characteristic of mesenchymal lect for the immortalized cells that exhibit during tumor formation. The 114 tumor CIS genes GBM (4), indicating that the mouse an unlimited proliferative potential in were newly acquired during tumor development. tumors genetically resemble mesen- culture. The frequency of immortaliza- These gene sets are enriched in processes associated chymal GBM. with distinct signaling pathways. tion was higher in cells undergoing active We next examined the combinations SB transposition than in cells that were of CIS genes in tumors derived from not. Gene expression profiling of the immortalized cells showed a single immortalized line. Notably, each tumor showed a distinct that these cells were significantly enriched for genes differentially combination of insertional mutations in RTK pathway genes, expressed in astroglial cells, which represent an immature even though all tumors were derived from a single line, in- stage of the astrocyte lineage. DNA analysis showed that con- dicating that transplanted cells become tumorigenic by acquiring tinuously mobilized transposons generated a large repertoire of immortalized cells with unique combinations of insertions (Fig. P1). More than half of the immortalized lines with active Author contributions: H.K., N.G.C., and N.A.J. designed research; H.K. and H.T. performed SB transposition induced tumors after s.c. transplantation, research; N.N., K.U., M.N., S.W., A.G.R., and D.J.A. contributed new reagents/analytic whereas only one of six immortalized lines that lacked active SB tools; H.K., C.C.K.Y., J.M.W., N.N., and A.G.R. analyzed data; and H.K., N.G.C., and transposition was tumorigenic. Analysis of DNA microarrays N.A.J. wrote the paper. showed that tumors were significantly enriched for genes specific The authors declare no conflict of interest. for the mesenchymal subtype of GBM, consistent with an This is a Contributed submission. astroglial origin for mesenchymal GBM (4). Data deposition: The array data reported in this paper have been deposited in the Gene We next sequenced the transposon insertion sites from Expression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo/ (accession no. GSE3689). 25 immortalized lines and 67 tumors and identified common 1To whom correspondence should be addressed. E-mail: [email protected]. insertion sites (CISs), which are regions in the genome that See full research article on page E2998 of www.pnas.org. harbor a higher number of transposon insertions than predicted Cite this Author Summary as: PNAS 10.1073/pnas.1215899109. 17746–17747 | PNAS | October 30, 2012 | vol. 109 | no. 44 www.pnas.org/cgi/doi/10.1073/pnas.1215899109 Downloaded by guest on September 24, 2021 new insertions in genes that promote tumor development, fol- These analyses will undoubtedly aid in the identification of PNAS PLUS lowed by clonal expansion and the generation of tumors with potential therapeutic targets for this deadly disease. distinct combinations of mutations in tumor CIS genes (Fig. P1). Functional analysis of the tumor CIS genes identified in our 1. Singh SK, et al. (2004) Identification of human brain tumour initiating cells. Nature 432 screen, such as Met, Pdgfrb, and Gab1, showed that these genes (7015):396–401. play functionally overlapping roles in tumor development. 2. Copeland NG, Jenkins NA (2010) Harnessing transposons for cancer gene discovery. Nat fi Rev Cancer 10(10):696–706. Together, these ndings suggest that tumor formation in this 3. Dupuy AJ, et al. (2009) A modified sleeping beauty transposon system that can be model system mimics the evolutionary processes now thought to used to model a wide variety of human cancers in mice. Cancer Res 69(20): generate many human cancers, in which the pathway to tumor 8150–8156. development has a branched architecture reminiscent of Dar- 4. Verhaak RG, et al.; Cancer Genome Atlas Research Network (2010) Integrated genomic ’ analysis identifies clinically relevant subtypes of glioblastoma characterized by win s iconic evolutionary tree (5). Our dataset of GBM-causing abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 17(1):98–110. genes provides a rich resource for cross-species comparative 5. Gerlinger M, et al. (2012) Intratumor heterogeneity and branched evolution revealed analyses of forthcoming human sequencing data from GBMs. by multiregion sequencing. N Engl J Med 366(10):883–892. MEDICAL SCIENCES Koso et al. PNAS | October 30, 2012 | vol. 109 | no. 44 | 17747 Downloaded by guest on September 24, 2021.