. . _:b463924222 . . _:b11235251 . . _:b11235259 "We adapted the centroid-based classifier in order to use these orthologs genes [>>10<<]. This new classifier was re-trained and re-tested on the original TCGA dataset, with the loss in classification power was less than 10% (Table S1). The new classifier also retained the capability to sort the same TCGA dataset into four" . _:b463924239 . . _:b11235261 "From this study we compiled a list of 369 genes that were derived from 5 gene lists as published [>>43<<], which represented signatures of five cell types:" . _:b463924199 . . _:b11235289 "the two mouse models we used the limma package [86]\u2013[88] within R/Bioconductor framework [89]\u2013[93], due to its high sensitivity and increased control for false positive rate compared with other variance modeling strategies [94]\u2013[>>97<<]. For this specific analysis we normalized the mouse samples using gcRMA due to its optimal combination between precision and accuracy [98], [99]." . _:b463924273 . _:b11235328 . . _:b11235270 "PIC retrovirus was titered to 106/ml using Mouse Embryonic Fibroblasts (MEFs) prepared from stop-floxed YFP mice [>>77<<]. Brain surgery procedures in mice were as described in Methods S1. For WM targeting, we used the coordinates of 2.1 mm lateral, 2.2 mm rostral and 1.8 mm deep, (2.1 mm+2.2 mm+1.8 mm) with bregma as the reference point. For targeting" . . _:b463924282 . _:b11235300 "The question of which cells can give rise to GBM, and where in the adult brain such cells reside, has long been an area of intense research and controversy [1], [2], [49], [>>50<<]. Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [18]\u2013[20], [25], [26], [51], [52]." . . _:b11235231 . _:b11235312 . _:b463924232 . _:b463924268 . . _:b11235266 "Not surprisingly, many of these processes are tightly regulated by p53 [47], [>>48<<]. Interestingly, 195 genes on the DGE list were not identified as being related to p53 by the Ingenuity-based analysis, suggesting that such a relationship has not previously been established. We propose that these 195 genes are ideal" . . _:b11235337 "Using activated EGFR combined with p53 deletions, Persson et al. provided evidence that OPCs can be induced to form tumors resembling human oligodendroglioma [>>27<<]. The differences in tumor types seen in the previous studies and our study could be due to the nature of the genetic alterations introduced and/or the age of the animal when these alterations are delivered. Interestingly, it was shown" . . _:b463924283 . _:b463924231 . . _:b11235272 "Mice harboring floxed tumor suppressors and stop-floxed reporters were generated by breeding the following strains: floxed Pten mice [78], floxed p53 mice [>>79<<], stop-floxed YFP mice [77] and stop-floxed luciferase mice (Thomas Ludwig, unpublished)." . _:b463924249 . . . . . _:b463924197 . . _:b11235293 "GSEA [100], [101] of the murine neural lineage genes [>>43<<] in the integrated mouse and human GBM data set was conducted as described [10] and implemented in a matlab script [102]." . _:b463924229 . . _:b463924228 . _:b11235234 . _:b463924212 . . _:b463924231 . . . _:b463924230 . . . . _:b463924225 . . _:b463924224 . . _:b463924227 . _:b11235339 . _:b463924211 . _:b11235239 . _:b463924226 . _:b463924237 . _:b11235341 . _:b463924218 . _:b463924236 . _:b11235341 "These tumors were widely distributed in the brain and showed a spectrum of phenotypes that resembled the different subtypes of high grade gliomas as described [7], [>>71<<]. They did not find a significant correlation between the types of genetic alterations used to initiate tumors and the subtypes of tumors that formed. However, they found an intriguing correlation with tumor location, suggesting that tumor" . _:b463924239 . _:b11235267 _:b11235272 . _:b11235267 _:b11235273 . _:b11235267 _:b11235274 . . _:b11235273 . _:b11235267 _:b11235275 . _:b463924238 . _:b11235267 _:b11235276 . _:b11235267 _:b11235277 . _:b11235267 _:b11235278 . _:b11235267 _:b11235279 . _:b463924233 . _:b463924291 . _:b11235334 "Similarly, several experimental models have shown that PDGF stimulation can induce the formation of gliomas [11], [12], [14], [17], [22], [>>42<<]. In particular, one study using the RCAS/tv-a system has shown that PDGF is sufficient to induce neonatal OPCs to form gliomas that resemble low-grade oligodendrogliomas [21]. In our model the tumors show the histological features of GBM." . _:b463924236 . _:b463924232 . _:b11235242 "Among the adult glial progenitor populations, the best characterized are the oligodendrocyte progenitor cells (OPCs) that express PDGFR\u03B1, NG2 and Olig2 [>>28<<], [30]\u2013[32]." . _:b11235267 _:b11235268 . _:b11235267 _:b11235269 . _:b11235267 _:b11235270 . _:b463924195 . _:b11235267 _:b11235271 . _:b463924235 . _:b11235267 _:b11235288 . _:b11235267 _:b11235289 . _:b11235267 _:b11235290 . _:b11235267 _:b11235291 . _:b463924234 . _:b11235267 _:b11235292 . _:b11235267 _:b11235293 . _:b11235267 _:b11235294 . _:b11235267 _:b11235295 . _:b11235230 "Numerous studies have characterized the expression profiles and genetic alterations found in GBM [>>3<<]\u2013[9]. A recent comprehensive analysis of The Cancer Genome Atlas (TCGA) dataset has identified four distinct subtypes of GBM; Proneural, Neural, Classical and Mesenchymal [10]. Interestingly, each of these subtypes show an enrichment of" . _:b463924245 . _:b11235267 _:b11235280 . _:b11235267 _:b11235281 . _:b11235267 _:b11235282 . _:b11235267 _:b11235283 . . _:b11235267 _:b11235284 . _:b463924244 . _:b11235267 _:b11235285 . _:b11235344 "is applied to 59 TCGA Proneural GBM identified by Verhaak et al. classification system, 40 of 59 TCGA Proneural patients remained as Proneural, while the other 19 patients were classified as Proliferative (18) or Mesenchymal (1) [7], [>>10<<] (Figure 5B)." . _:b11235267 _:b11235286 . _:b11235267 _:b11235287 . _:b463924247 . _:b463924246 . . . . _:b11235244 "Among the adult glial progenitor populations, the best characterized are the oligodendrocyte progenitor cells (OPCs) that express PDGFR\u03B1, NG2 and Olig2 [28], [30]\u2013[>>32<<]. OPCs are widely distributed, both in the white matter and cortex, and comprise the largest population of cycling cells in the adult brain [33]\u2013[37]. Thus OPCs represent an abundant reservoir of potentially transformable cells." . _:b463924201 . _:b463924241 . _:b463924260 . _:b11235253 . . _:b463924240 . _:b11235310 "Retroviruses only infect dividing cells [>>56<<], and therefore, can be used to selectively target progenitor cells in the brain [57]\u2013[59]." . _:b463924265 . _:b463924243 . _:b11235240 "of cells in the adult brain that may have the capacity to form brain tumors, including neural stem cells in the subventricular zone (SVZ) [18], [25], [26] and more differentiated glial progenitors in the subcortical white matter [17], [>>28<<]. Furthermore, experiments using the RCAS/tv-a system have shown that progenitors with the capacity to form tumors are not restricted to the SVZ, but are distributed throughout the adult brain [29]." . _:b11235255 . _:b463924242 . _:b11235270 . _:b11235285 "These results also allow us to validate the classification results obtained using the modified ClaNC classifier [>>10<<], [85]. We performed this analysis for both the mouse samples and the human validation set (Table S1), and the error rate was again below 10% in the human set. With both methods, the classification of all samples matched." . _:b463924253 . _:b11235235 "A number of studies have used animal models to explore the process of gliomagenesis [11]\u2013[>>27<<]. Many of these models induced tumor formation by introducing genetic lesions into the embryonic or neonatal brain." . . _:b463924252 . _:b11235254 "Furthermore, these cells can be propagated in cultures using growth conditions that have been optimized for OPCs [17], [>>41<<]. Immunocytochemical analysis confirmed that these cells continue to express many OPC markers, including Olig2, PDGFR\u03B1, Nkx2.2, NG2 and Sox10 (Figure S2) as well as markers seen in glial progenitors and other progenitor populations, such" . _:b11235292 "GSEA [100], [>>101<<] of the murine neural lineage genes [43] in the integrated mouse and human GBM data set was conducted as described [10] and implemented in a matlab script [102]." . . . _:b11235315 . _:b463924255 . . _:b11235314 . _:b11235303 "Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [18]\u2013[20], [>>25<<], [26], [51], [52]." . _:b463924254 . _:b11235302 . _:b463924249 . _:b463924248 . _:b11235316 . _:b11235342 . . _:b463924251 . _:b11235343 . . _:b463924198 . _:b463924203 . _:b463924250 . _:b11235318 . _:b11235340 . _:b463924234 . _:b463924261 . . _:b463924260 . _:b11235320 . . _:b11235296 _:b11235300 . _:b463924263 . _:b11235296 _:b11235301 . _:b11235296 _:b11235302 . _:b11235278 "org/cgi-bin/cancer/publications/pub_paper.cgi? mode=view&paper_id=162&p=t) [>>7<<], [10], [44], [45]." . _:b11235296 _:b11235303 . _:b11235280 "org/cgi-bin/cancer/publications/pub_paper.cgi? mode=view&paper_id=162&p=t) [7], [10], [>>44<<], [45]." . . _:b463924262 . _:b11235322 . _:b11235296 _:b11235297 . _:b11235296 _:b11235298 . _:b11235296 _:b11235299 . _:b463924194 . _:b11235296 _:b11235308 . _:b11235323 . _:b463924257 . _:b11235296 _:b11235309 . _:b11235296 _:b11235310 . _:b11235296 _:b11235311 . _:b11235296 _:b11235304 . _:b463924256 . . . _:b463924227 . _:b11235296 _:b11235305 . _:b11235296 _:b11235306 . _:b11235296 _:b11235307 . _:b11235289 . _:b11235296 _:b11235316 . _:b463924259 . _:b11235296 _:b11235317 . _:b11235296 _:b11235318 . _:b11235296 _:b11235319 . . _:b11235296 _:b11235312 . _:b463924258 . _:b11235296 _:b11235313 . _:b11235296 _:b11235314 . _:b11235296 _:b11235315 . _:b11235296 _:b11235324 . _:b11235308 "Furthermore, evidence suggested that GBM arising outside the SVZ may have distinct growth characteristics and clinical outcomes [>>54<<], [55]. Therefore, characterizing the tumorigenic potential of cells that reside outside the SVZ is of both scientific and clinical interest." . _:b463924269 . _:b11235296 _:b11235325 . _:b11235296 _:b11235326 . _:b11235296 _:b11235327 . _:b11235296 _:b11235320 . _:b11235298 . _:b463924268 . _:b11235296 _:b11235321 . _:b11235296 _:b11235322 . _:b11235307 "However, radiographic data from patients suggests that the origins of GBM are not restricted to the SVZ and that many tumors arise in the subcortical white matter [>>53<<]. Furthermore, evidence suggested that GBM arising outside the SVZ may have distinct growth characteristics and clinical outcomes [54], [55]. Therefore, characterizing the tumorigenic potential of cells that reside outside the SVZ is of" . _:b11235250 "Mutations and loss of heterozygosity of p53 and amplification of PDGFR\u03B1 are most frequently seen in the Proneural subtype, while loss of Pten is observed throughout all subtypes [8]\u2013[>>10<<], [40]. However, much remains to be learned about how the cells of origin and genetic alterations interact to determine GBM phenotype." . _:b11235296 _:b11235323 . _:b463924223 . _:b463924271 . _:b11235237 "There are several different populations of cells in the adult brain that may have the capacity to form brain tumors, including neural stem cells in the subventricular zone (SVZ) [18], [>>25<<], [26] and more differentiated glial progenitors in the subcortical white matter [17], [28]." . . _:b463924270 . . . _:b463924265 . _:b463924293 . _:b463924264 . . _:b463924267 . _:b11235304 "Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [18]\u2013[20], [25], [>>26<<], [51], [52]." . _:b11235332 "Similarly, several experimental models have shown that PDGF stimulation can induce the formation of gliomas [11], [12], [14], [>>17<<], [22], [42]." . _:b463924266 . . _:b463924277 . _:b463924276 . . _:b463924279 . . _:b463924278 . _:b11235308 . _:b11235332 . _:b463924273 . _:b11235333 . _:b463924272 . _:b11235334 . _:b463924275 . _:b463924216 . _:b11235335 . . . . _:b463924274 . . _:b11235328 . . _:b463924272 . _:b463924285 . . _:b11235329 . _:b463924284 . _:b11235330 . _:b11235266 . _:b11235338 "Interestingly, it was shown that, in addition to a subset of GBM [>>10<<], oligodendrogliomas are also enriched in the Proneural signature [70]." . _:b463924287 . _:b11235283 . _:b11235331 . _:b463924286 . _:b11235337 . _:b11235282 . _:b11235340 . . _:b463924281 . _:b11235285 . _:b11235260 . . _:b11235341 . _:b463924280 . _:b11235284 . _:b11235261 . _:b11235342 . . _:b463924283 . . _:b11235343 . _:b463924282 . _:b11235336 . _:b463924293 . _:b11235337 . . . . _:b11235343 "system is applied to 59 TCGA Proneural GBM identified by Verhaak et al. classification system, 40 of 59 TCGA Proneural patients remained as Proneural, while the other 19 patients were classified as Proliferative (18) or Mesenchymal (1) [>>7<<], [10] (Figure 5B)." . _:b463924292 . "PMC0" . _:b11235338 . _:b463924295 . _:b11235321 . _:b11235241 "Furthermore, experiments using the RCAS/tv-a system have shown that progenitors with the capacity to form tumors are not restricted to the SVZ, but are distributed throughout the adult brain [>>29<<]. Among the adult glial progenitor populations, the best characterized are the oligodendrocyte progenitor cells (OPCs) that express PDGFR\u03B1, NG2 and Olig2 [28], [30]\u2013[32]. OPCs are widely distributed, both in the white matter and cortex," . . _:b11235339 . . _:b463924294 . _:b11235307 . _:b11235348 . _:b463924289 . . _:b463924252 . _:b463924213 . . _:b463924288 . . _:b463924291 . _:b11235231 "Numerous studies have characterized the expression profiles and genetic alterations found in GBM [3]\u2013[>>9<<]. A recent comprehensive analysis of The Cancer Genome Atlas (TCGA) dataset has identified four distinct subtypes of GBM; Proneural, Neural, Classical and Mesenchymal [10]. Interestingly, each of these subtypes show an enrichment of gene" . _:b463924275 . _:b463924290 . _:b11235294 . _:b463924243 . _:b11235227 _:b11235228 . _:b11235227 _:b11235229 . _:b11235344 . _:b11235256 "These cells likely represent uninfected glial progenitors that are being induced to proliferate via paracrine growth factor stimulation, as previously described [17], [>>22<<], [42]. However, the YFP- cells at 21 dpi accounted for only \u223C6% of total number of proliferating cells (data not shown). This indicates that the retrovirus infected cells, with Pten deleted, have a significant selective advantage, and" . . _:b463924229 "3"^^ . _:b11235227 _:b11235230 . _:b11235227 _:b11235231 . _:b11235246 "OPCs are widely distributed, both in the white matter and cortex, and comprise the largest population of cycling cells in the adult brain [33]\u2013[>>37<<]. Thus OPCs represent an abundant reservoir of potentially transformable cells." . . . _:b11235345 . _:b463924228 "3"^^ . _:b463924259 . _:b11235346 . _:b463924231 "3"^^ . _:b11235262 "We extended the GSEA analysis to three additional datasets [>>7<<], [44], [45]. Out of the combined 176 patients, 48 were classified as Proneural GBM (Verhaak RG, Hoadley KA and Hayes DN, personal communications)." . . _:b11235227 _:b11235240 . _:b11235227 _:b11235241 . _:b11235347 . _:b11235262 . _:b463924230 "3"^^ . _:b11235227 _:b11235242 . _:b11235227 _:b11235243 . _:b11235227 _:b11235244 . _:b463924279 . _:b11235227 _:b11235245 . . _:b11235227 _:b11235246 . _:b463924225 "3"^^ . _:b11235227 _:b11235247 . _:b11235227 _:b11235232 . _:b11235227 _:b11235233 . _:b463924224 "3"^^ . _:b11235227 _:b11235234 . _:b11235227 _:b11235235 . . _:b11235227 _:b11235236 . _:b11235227 _:b11235237 . _:b463924204 . _:b463924227 "3"^^ . _:b11235227 _:b11235238 . _:b11235227 _:b11235239 . . . _:b463924226 "3"^^ . . . _:b463924237 "2"^^ . _:b11235247 "Furthermore, these different progenitor cells are susceptible to the genetic alterations seen in the particular tumor subtype to which they give rise [>>38<<], [39]. Specific genetic lesions are also observed in different subtypes of GBM. Mutations and loss of heterozygosity of p53 and amplification of PDGFR\u03B1 are most frequently seen in the Proneural subtype, while loss of Pten is observed" . _:b11235227 _:b11235248 . . _:b463924236 "2"^^ . . _:b11235227 _:b11235249 . _:b11235227 _:b11235250 . _:b11235227 _:b11235251 . . _:b463924239 "2"^^ . _:b11235228 . _:b463924238 "2"^^ . . _:b463924233 "3"^^ . . _:b11235232 "A recent comprehensive analysis of The Cancer Genome Atlas (TCGA) dataset has identified four distinct subtypes of GBM; Proneural, Neural, Classical and Mesenchymal [>>10<<]. Interestingly, each of these subtypes show an enrichment of gene expression signatures from distinct neural lineages, implying that the expression patterns of the different subtypes may reflect the phenotype of their specific cells of" . _:b463924232 "3"^^ . _:b463924295 . _:b463924217 . _:b463924235 "2"^^ . _:b11235249 . . _:b11235279 . _:b463924234 "3"^^ . _:b463924245 "2"^^ . _:b463924244 "2"^^ . _:b463924247 "2"^^ . _:b463924246 "2"^^ . . . _:b11235250 . _:b463924241 "2"^^ . . _:b463924240 "2"^^ . _:b11235280 . _:b11235227 "introduction" . _:b463924243 "2"^^ . . . _:b463924242 "2"^^ . _:b463924255 . _:b463924288 . _:b11235340 "These tumors were widely distributed in the brain and showed a spectrum of phenotypes that resembled the different subtypes of high grade gliomas as described [>>7<<], [71]. They did not find a significant correlation between the types of genetic alterations used to initiate tumors and the subtypes of tumors that formed. However, they found an intriguing correlation with tumor location, suggesting that" . _:b463924245 . _:b463924289 . _:b463924290 . _:b463924253 "2"^^ . _:b463924291 . _:b463924200 . _:b463924292 . _:b463924293 . . _:b463924294 . _:b463924252 "2"^^ . _:b463924295 . _:b463924262 . _:b463924214 . _:b463924255 "2"^^ . _:b11235259 . _:b463924254 "2"^^ . _:b11235252 _:b11235253 . _:b11235252 _:b11235254 . _:b463924272 . _:b11235252 _:b11235255 . _:b463924273 . _:b463924274 . _:b11235258 . _:b463924249 "2"^^ . _:b463924275 . _:b463924233 . _:b463924276 . _:b463924277 . _:b11235252 _:b11235260 . _:b463924278 . _:b463924248 "2"^^ . _:b463924279 . _:b11235252 _:b11235261 . _:b11235252 _:b11235262 . _:b463924280 . . _:b11235296 "discussion" . _:b463924281 . _:b11235252 _:b11235263 . _:b11235252 _:b11235256 . _:b463924282 . _:b463924251 "2"^^ . _:b463924283 . _:b11235252 _:b11235257 . _:b11235252 _:b11235258 . _:b463924284 . _:b11235252 _:b11235259 . _:b463924285 . _:b463924286 . _:b463924250 "2"^^ . _:b463924287 . . _:b11235315 "Using the same approach as published [>>10<<], we showed that the mouse tumors are most closely related to the Proneural subtype." . _:b463924256 . _:b463924257 . _:b463924258 . _:b463924261 "2"^^ . _:b11235336 "In addition to PDGFR\u03B1 amplification, other genetic alterations, including p53 and IDH1 mutations, are strongly associated with Proneural GBM [>>10<<]. We propose that OPCs are particularly sensitive to the transforming effects of these genetic alterations." . _:b463924259 . _:b463924260 . _:b463924235 . _:b11235271 . _:b11235268 . _:b463924261 . _:b463924262 . _:b11235233 . _:b463924260 "2"^^ . _:b463924263 . _:b463924264 . _:b463924265 . _:b463924266 . _:b11235255 "These cells likely represent uninfected glial progenitors that are being induced to proliferate via paracrine growth factor stimulation, as previously described [>>17<<], [22], [42]." . _:b11235232 . _:b463924263 "2"^^ . _:b463924267 . _:b463924268 . _:b463924269 . _:b463924270 . _:b463924262 "2"^^ . _:b463924271 . _:b463924240 . _:b11235316 "In fact, 94 of the 210 (>45%) Proneural classifier genes are enriched in OPCs [>>10<<], [43] (Table S1)." . . _:b463924241 . _:b463924242 . _:b463924257 "2"^^ . _:b463924243 . _:b463924244 . _:b11235304 . _:b463924245 . _:b463924246 . _:b11235339 "A recent paper showed that genetically deleting combinations of tumor suppressors (Pten, Rb1 and p53) in the adult brain with an inducible Cre system (GFAP-CreER) will give rise to high grade astrocytomas [>>71<<]. These tumors were widely distributed in the brain and showed a spectrum of phenotypes that resembled the different subtypes of high grade gliomas as described [7], [71]." . _:b463924256 "2"^^ . _:b11235299 . _:b11235264 . _:b463924247 . _:b463924248 . _:b11235342 "Previously, using a list containing 35 genes, high grade glioma (HGG), including GBM, were classified into three groups: Proneural, Proliferative and Mesenchymal [>>7<<]. In this classification system, the Proliferative and Mesenchymal tumors were associated with poor prognosis, as defined by shorter patient survival, while Proneural tumors were associated with better prognosis. When the Phillips et al." . . _:b463924249 . _:b463924250 . _:b463924259 "2"^^ . _:b463924251 . . _:b463924252 . _:b463924253 . _:b463924254 . _:b463924258 "2"^^ . _:b463924255 . _:b463924224 . _:b463924225 . _:b463924226 . _:b463924269 "2"^^ . _:b463924227 . . _:b463924261 . _:b463924228 . _:b463924229 . _:b463924230 . _:b463924268 "2"^^ . _:b463924231 . _:b463924232 . . _:b11235251 "Mutations and loss of heterozygosity of p53 and amplification of PDGFR\u03B1 are most frequently seen in the Proneural subtype, while loss of Pten is observed throughout all subtypes [8]\u2013[10], [>>40<<]. However, much remains to be learned about how the cells of origin and genetic alterations interact to determine GBM phenotype." . _:b463924233 . _:b463924234 . _:b463924271 "2"^^ . _:b463924235 . _:b463924236 . _:b11235301 . _:b463924237 . _:b11235314 "GBM can be divided into subtypes on the basis of global gene expression [7], [>>10<<]. Using the same approach as published [10], we showed that the mouse tumors are most closely related to the Proneural subtype." . _:b463924238 . _:b463924270 "2"^^ . _:b11235333 . _:b463924239 . _:b11235267 "methods" . _:b463924265 "2"^^ . _:b11235298 "The question of which cells can give rise to GBM, and where in the adult brain such cells reside, has long been an area of intense research and controversy [1], [>>2<<], [49], [50]. Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [18]\u2013[20], [25], [26], [51], [52]." . . _:b463924264 "2"^^ . . _:b11235327 "PDGFR\u03B1 is selectively expressed in OPCs in the normal brain [60]\u2013[64], and the ability of PDGF signaling to stimulate the proliferation and migration of OPCs is well established [61], [>>64<<]\u2013[69]. Similarly, several experimental models have shown that PDGF stimulation can induce the formation of gliomas [11], [12], [14], [17], [22], [42]. In particular, one study using the RCAS/tv-a system has shown that PDGF is sufficient to" . _:b11235317 . _:b463924267 "2"^^ . . . _:b463924287 . _:b463924266 "2"^^ . _:b11235319 . _:b463924277 "2"^^ . _:b11235291 . . _:b463924226 . _:b463924276 "2"^^ . _:b463924209 . _:b11235243 "Among the adult glial progenitor populations, the best characterized are the oligodendrocyte progenitor cells (OPCs) that express PDGFR\u03B1, NG2 and Olig2 [28], [>>30<<]\u2013[32]. OPCs are widely distributed, both in the white matter and cortex, and comprise the largest population of cycling cells in the adult brain [33]\u2013[37]. Thus OPCs represent an abundant reservoir of potentially transformable cells." . _:b463924279 "2"^^ . . _:b11235276 . _:b463924240 . _:b463924278 "2"^^ . _:b11235278 . . _:b463924263 . _:b11235257 . _:b463924273 "2"^^ . _:b463924286 . _:b463924272 "2"^^ . . _:b463924275 "2"^^ . . _:b463924285 . _:b463924274 "2"^^ . _:b463924210 . _:b11235228 . _:b11235252 _:b11235264 . _:b463924285 "2"^^ . . _:b11235252 _:b11235265 . _:b11235252 _:b11235266 . _:b11235229 . . _:b463924284 "2"^^ . _:b11235230 . _:b11235322 "Amplification of the PDGFR\u03B1 receptor is most commonly seen in Proneural GBM [2], [>>10<<]. Furthermore, proteomic studies revealed that activation of PDGF signaling is associated with Proneural phenotype [40]. PDGFR\u03B1 is selectively expressed in OPCs in the normal brain [60]\u2013[64], and the ability of PDGF signaling to stimulate" . _:b463924287 "2"^^ . _:b11235313 . _:b11235231 . _:b463924286 "2"^^ . . _:b463924281 "2"^^ . _:b11235248 . _:b463924280 "2"^^ . . _:b11235272 . _:b463924283 "2"^^ . _:b11235290 "For this specific analysis we normalized the mouse samples using gcRMA due to its optimal combination between precision and accuracy [98], [>>99<<]. For the Ingenuity Pathway Analysis\u00AE (www.ingenuity.com), we used the genes lists from the differential expression analysis, at False Discovery Rate 0.1 cutoff, to identified biological function and diseases that were enriched. The" . _:b11235227 . _:b463924196 . _:b463924282 "2"^^ . _:b463924266 . _:b11235236 . . _:b463924293 "2"^^ . _:b11235237 . _:b463924292 "2"^^ . . _:b11235238 . . _:b463924295 "2"^^ . _:b11235239 . _:b11235271 "Mice harboring floxed tumor suppressors and stop-floxed reporters were generated by breeding the following strains: floxed Pten mice [>>78<<], floxed p53 mice [79], stop-floxed YFP mice [77] and stop-floxed luciferase mice (Thomas Ludwig, unpublished)." . _:b463924294 "2"^^ . _:b11235232 . _:b463924289 "2"^^ . _:b11235233 . _:b11235287 "To identify the statistically significant differentially expressed genes between the two mouse models we used the limma package [86]\u2013[88] within R/Bioconductor framework [>>89<<]\u2013[93], due to its high sensitivity and increased control for false positive rate compared with other variance modeling strategies [94]\u2013[97]." . _:b463924288 "2"^^ . _:b11235234 . _:b11235330 . . _:b463924291 "2"^^ . _:b11235235 . _:b463924290 "2"^^ . . _:b11235297 . _:b11235244 . _:b463924280 . _:b11235245 . _:b11235246 . _:b11235321 "Amplification of the PDGFR\u03B1 receptor is most commonly seen in Proneural GBM [>>2<<], [10]. Furthermore, proteomic studies revealed that activation of PDGF signaling is associated with Proneural phenotype [40]. PDGFR\u03B1 is selectively expressed in OPCs in the normal brain [60]\u2013[64], and the ability of PDGF signaling to" . _:b463924230 . _:b11235247 . . _:b11235234 "A number of studies have used animal models to explore the process of gliomagenesis [>>11<<]\u2013[27]. Many of these models induced tumor formation by introducing genetic lesions into the embryonic or neonatal brain." . _:b463924289 . _:b11235240 . _:b11235274 "Primary tumor preparation was generated as described in Methods S1 [>>17<<]. The tumor preparation was plated overnight on poly-lysine coated dishes. For serial transplantation studies cells were injected into NOD/SCID mice (as described above)." . _:b11235241 . _:b11235303 . . . _:b11235242 . . _:b11235268 "PDGF-IRES-Cre was generated by cloning human PDGF-B [>>42<<] and Cre [75] into pQXIX vector (Clontech). Cre-GFP was described [76]. VSVG pseudotyped retrovirus was generated as described in Methods S1." . _:b11235243 . _:b11235286 . _:b11235252 . _:b11235253 . _:b463924290 . _:b11235254 . _:b11235233 "each of these subtypes show an enrichment of gene expression signatures from distinct neural lineages, implying that the expression patterns of the different subtypes may reflect the phenotype of their specific cells of origin [>>10<<]." . _:b463924271 . _:b11235255 . _:b11235248 . _:b11235328 "PDGFR\u03B1 is selectively expressed in OPCs in the normal brain [60]\u2013[64], and the ability of PDGF signaling to stimulate the proliferation and migration of OPCs is well established [61], [64]\u2013[>>69<<]. Similarly, several experimental models have shown that PDGF stimulation can induce the formation of gliomas [11], [12], [14], [17], [22], [42]. In particular, one study using the RCAS/tv-a system has shown that PDGF is sufficient to" . . _:b11235254 . . _:b11235249 . _:b11235250 . _:b11235293 . _:b11235295 . _:b11235251 . _:b11235296 _:b11235332 . _:b11235296 _:b11235333 . _:b11235260 . _:b11235296 _:b11235334 . . _:b11235296 _:b11235335 . _:b11235296 _:b11235328 . _:b11235296 _:b11235329 . _:b11235305 "Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [18]\u2013[20], [25], [26], [>>51<<], [52]. However, radiographic data from patients suggests that the origins of GBM are not restricted to the SVZ and that many tumors arise in the subcortical white matter [53]. Furthermore, evidence suggested that GBM arising outside the" . _:b11235261 . _:b11235296 _:b11235330 . _:b11235329 . _:b11235296 _:b11235331 . _:b11235296 _:b11235340 . _:b11235281 "To identify the human to mouse orthologs, we utilized the gene level sequence based mapping method, which is based on the reciprocal best match as available from NCBI HomoloGene (build64) [81], [>>82<<]. The data was centered around zero by performing a z-score linear transformation of the expression values. Given the highly homogenous nature of the mouse samples, we used an across genes standardization, which provided an optimal" . _:b11235296 _:b11235341 . _:b11235262 . _:b11235296 _:b11235342 . _:b463924207 . _:b11235296 _:b11235343 . _:b11235296 _:b11235336 . . _:b11235263 . _:b11235296 _:b11235337 . . _:b11235296 _:b11235338 . _:b11235296 _:b11235339 . _:b11235296 _:b11235348 . _:b11235256 . . _:b11235296 _:b11235344 . _:b11235296 _:b11235345 . _:b11235257 . _:b463924256 . _:b11235296 _:b11235346 . _:b11235296 _:b11235347 . _:b11235269 "Cre-GFP was described [>>76<<]. VSVG pseudotyped retrovirus was generated as described in Methods S1." . _:b11235258 . _:b11235259 . _:b11235242 . _:b11235268 . _:b11235345 . _:b11235287 . _:b463924205 . . _:b463924269 . _:b11235269 . _:b11235247 . _:b463924253 . _:b463924244 . _:b11235240 . _:b11235270 . _:b11235238 . _:b11235271 . _:b463924219 . _:b11235243 . _:b11235264 . _:b11235265 . _:b463924278 . . _:b11235300 . _:b11235266 . . _:b11235277 . _:b11235267 . . _:b11235294 "GSEA [100], [101] of the murine neural lineage genes [43] in the integrated mouse and human GBM data set was conducted as described [>>10<<] and implemented in a matlab script [102]. Expression values were replaced by ranks and the reported Enrichment Score (ES) was computed as sum of the maximum deviations above and below zero in the random walk." . _:b11235245 "OPCs are widely distributed, both in the white matter and cortex, and comprise the largest population of cycling cells in the adult brain [>>33<<]\u2013[37]. Thus OPCs represent an abundant reservoir of potentially transformable cells." . _:b11235276 . _:b11235283 "In the first approach we modified the centroid classifier as published [>>10<<]. We re-computed the centroids using 723, instead of 840, classifier genes." . _:b11235277 . _:b11235324 . _:b11235252 . _:b11235256 . . _:b11235278 . . _:b463924198 "6"^^ . _:b11235279 . _:b11235258 "Of 840 genes that were used to classify human GBM from TCGA into 4 subtypes (Proneural, Neural, Classical and Mesenchymal) [>>10<<], 723 orthologs could be mapped to the mouse arrays (see Methods and Table S1)." . . _:b11235272 . _:b11235236 . _:b463924193 "16"^^ . . _:b463924196 "6"^^ . _:b11235273 . _:b11235277 "The dataset for different neural lineages in the mouse brains was downloaded from NCBI GEO with the access number GSE9566 [>>43<<]. The complete human GBM datasets were downloaded from TCGA website (cancergenome.nih.gov), NCBI GEO with the access number GSE4271, Rembrandt website (https://caintegrator.nci.nih.gov/rembrandt/) and Broad Institute (http://www." . _:b11235335 . _:b463924199 "5"^^ . _:b463924197 "6"^^ . _:b11235274 . "10.1371%2Fjournal.pone.0020041" . _:b11235275 . _:b11235324 "PDGFR\u03B1 is selectively expressed in OPCs in the normal brain [>>60<<]\u2013[64], and the ability of PDGF signaling to stimulate the proliferation and migration of OPCs is well established [61], [64]\u2013[69]." . _:b463924194 "8"^^ . _:b463924195 "6"^^ . _:b11235284 . . _:b11235319 "To further explore this issue, we expanded the analysis to compare the expression of genes that are enriched in OPCs to that of other cell types isolated from the mouse brain [>>43<<]. We showed that the mouse tumors and human Proneural GBM were both consistently enriched in OPC genes, but not in the signature genes of other cell types (oligodendrocytes, neurons, astrocytes or cultured astroglia). These results provide" . _:b11235285 . _:b11235276 "Raw data was then normalized and summarized by robust multichip average (RMA) [>>80<<]. The dataset for different neural lineages in the mouse brains was downloaded from NCBI GEO with the access number GSE9566 [43]. The complete human GBM datasets were downloaded from TCGA website (cancergenome.nih.gov), NCBI GEO with the" . . _:b11235286 . . _:b463924197 . _:b11235227 . _:b11235287 . _:b463924206 . . _:b463924196 . _:b463924205 "4"^^ . . _:b11235280 . _:b11235296 . _:b11235245 . _:b463924199 . _:b463924267 . _:b463924202 . _:b463924204 "5"^^ . . _:b11235281 . _:b463924198 . _:b463924207 "4"^^ . . _:b11235282 . _:b463924193 . _:b463924206 "4"^^ . _:b11235306 "Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [18]\u2013[20], [25], [26], [51], [>>52<<]. However, radiographic data from patients suggests that the origins of GBM are not restricted to the SVZ and that many tumors arise in the subcortical white matter [53]. Furthermore, evidence suggested that GBM arising outside the SVZ may" . _:b11235344 . _:b11235283 . _:b463924201 "5"^^ . _:b11235236 "There are several different populations of cells in the adult brain that may have the capacity to form brain tumors, including neural stem cells in the subventricular zone (SVZ) [>>18<<], [25], [26] and more differentiated glial progenitors in the subcortical white matter [17], [28]." . _:b463924193 . _:b11235292 . _:b463924195 . _:b11235346 "Recently, a report identified a group of TCGA GBM patients with a glioma-CpG island methylator phenotype (G-CIMP) [>>73<<]. The majority of these patients belongs to the Proneural subtype and is associated with longer survival [73], [74]." . _:b11235291 "GSEA [>>100<<], [101] of the murine neural lineage genes [43] in the integrated mouse and human GBM data set was conducted as described [10] and implemented in a matlab script [102]." . _:b463924200 "5"^^ . _:b11235293 . _:b11235329 "Similarly, several experimental models have shown that PDGF stimulation can induce the formation of gliomas [>>11<<], [12], [14], [17], [22], [42]." . _:b463924194 . _:b463924203 "5"^^ . _:b463924284 . _:b463924257 . _:b11235294 . _:b463924205 . _:b463924202 "5"^^ . . _:b11235241 . _:b11235295 . _:b463924204 . . _:b463924213 "4"^^ . _:b463924254 . _:b11235288 . _:b463924207 . _:b463924212 "4"^^ . _:b11235267 . _:b11235297 "The question of which cells can give rise to GBM, and where in the adult brain such cells reside, has long been an area of intense research and controversy [>>1<<], [2], [49], [50]. Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [18]\u2013[20], [25], [26], [51], [52]." . _:b11235289 . _:b463924206 . _:b463924215 "4"^^ . _:b11235290 . _:b463924201 . _:b463924214 "4"^^ . _:b11235311 "Retroviruses only infect dividing cells [56], and therefore, can be used to selectively target progenitor cells in the brain [>>57<<]\u2013[59]. Using a stop-floxed YFP reporter, we were able to identify and fate map the cells that had been infected by PDGF-IRES-Cre expressing retroviruses. The distinct advantages of this approach allow us not only to control the timing and" . _:b11235291 . _:b463924200 . _:b11235347 . _:b463924209 "4"^^ . . _:b11235300 . _:b463924203 . _:b11235346 . _:b463924208 "4"^^ . _:b463924237 . _:b11235301 . _:b463924277 . _:b463924202 . _:b11235263 . _:b463924211 "4"^^ . _:b11235299 "The question of which cells can give rise to GBM, and where in the adult brain such cells reside, has long been an area of intense research and controversy [1], [2], [>>49<<], [50]. Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [18]\u2013[20], [25], [26], [51], [52]." . _:b11235290 . _:b11235302 . _:b463924213 . _:b463924210 "4"^^ . _:b11235303 . _:b463924212 . _:b463924221 "3"^^ . _:b11235326 "PDGFR\u03B1 is selectively expressed in OPCs in the normal brain [60]\u2013[64], and the ability of PDGF signaling to stimulate the proliferation and migration of OPCs is well established [>>61<<], [64]\u2013[69]." . _:b11235334 . _:b11235238 "There are several different populations of cells in the adult brain that may have the capacity to form brain tumors, including neural stem cells in the subventricular zone (SVZ) [18], [25], [>>26<<] and more differentiated glial progenitors in the subcortical white matter [17], [28]." . _:b463924215 . _:b11235296 . _:b463924220 "3"^^ . _:b463924208 . _:b11235284 "The mis-classification rate in the original human validation set was below 10% compared to what was previously published [>>10<<]. In the second approach, we again used the human training set, and then averaged the expression values of each gene across samples belonging to the same GBM subtype. We computed the Pearson correlation coefficients between the given query" . _:b463924209 . _:b11235297 . _:b463924214 . _:b11235253 "Furthermore, these cells can be propagated in cultures using growth conditions that have been optimized for OPCs [>>17<<], [41]. Immunocytochemical analysis confirmed that these cells continue to express many OPC markers, including Olig2, PDGFR\u03B1, Nkx2.2, NG2 and Sox10 (Figure S2) as well as markers seen in glial progenitors and other progenitor populations," . _:b463924210 . . _:b463924223 "3"^^ . _:b463924211 . _:b463924274 . _:b463924212 . _:b463924213 . _:b11235298 . _:b463924228 . _:b463924209 . _:b463924214 . _:b463924222 "3"^^ . _:b463924215 . _:b463924216 . _:b11235286 "We used expression data from the murine brain transcriptome database as published [>>43<<], which includes Gene sets that are specifically enriched in the following cell types:" . _:b463924217 . _:b11235299 . _:b463924208 . _:b463924218 . _:b463924217 "3"^^ . _:b463924219 . _:b463924220 . _:b463924221 . _:b11235308 . _:b463924211 . . _:b463924222 . _:b463924216 "3"^^ . _:b463924223 . . . _:b463924193 . _:b11235309 . _:b11235264 "As a master transcriptional regulator, p53 can either activate or repress target gene expression [>>46<<]. Several genes on the DGE list, including Ccng1, Cdkn1a, Fas and Tnfrsf10, which contain binding sites for p53 in their promoters and whose expression is activated by p53, are expressed at higher levels in Ptenf/f tumors vs. Ptenf/f;" . _:b463924210 . _:b11235288 "between the two mouse models we used the limma package [86]\u2013[88] within R/Bioconductor framework [89]\u2013[93], due to its high sensitivity and increased control for false positive rate compared with other variance modeling strategies [>>94<<]\u2013[97]. For this specific analysis we normalized the mouse samples using gcRMA due to its optimal combination between precision and accuracy [98], [99]." . _:b463924194 . _:b463924219 "3"^^ . _:b463924195 . _:b463924196 . . _:b11235310 . _:b463924197 . _:b463924221 . _:b11235331 . _:b463924198 . _:b463924218 "3"^^ . _:b463924199 . _:b463924200 . _:b463924201 . _:b11235311 . _:b463924220 . _:b463924202 . _:b463924203 . _:b463924204 . _:b463924205 . _:b11235304 . _:b463924223 . _:b463924206 . _:b463924207 . _:b11235305 . _:b463924222 . _:b11235306 . _:b463924217 . _:b11235336 . _:b11235237 . _:b11235307 . _:b463924216 . . _:b463924224 . _:b11235273 "Mice harboring floxed tumor suppressors and stop-floxed reporters were generated by breeding the following strains: floxed Pten mice [78], floxed p53 mice [79], stop-floxed YFP mice [>>77<<] and stop-floxed luciferase mice (Thomas Ludwig, unpublished). The resulting mouse lines therefore are on mixed genetic backgrounds." . _:b11235316 . _:b463924219 . _:b11235305 . _:b11235338 . _:b11235317 . _:b463924218 . _:b463924258 . _:b11235318 . _:b463924264 . _:b11235327 . _:b11235295 "GSEA [100], [101] of the murine neural lineage genes [43] in the integrated mouse and human GBM data set was conducted as described [10] and implemented in a matlab script [>>102<<]. Expression values were replaced by ranks and the reported Enrichment Score (ES) was computed as sum of the maximum deviations above and below zero in the random walk." . _:b11235319 . _:b11235265 "Not surprisingly, many of these processes are tightly regulated by p53 [>>47<<], [48]. Interestingly, 195 genes on the DGE list were not identified as being related to p53 by the Ingenuity-based analysis, suggesting that such a relationship has not previously been established. We propose that these 195 genes are" . _:b11235326 . _:b11235228 "GBM is remarkably heterogeneous, and may actually represent several distinct entities with different cells of origin, different genetic lesions and different clinical behaviors [>>1<<], [2]. Numerous studies have characterized the expression profiles and genetic alterations found in GBM [3]\u2013[9]. A recent comprehensive analysis of The Cancer Genome Atlas (TCGA) dataset has identified four distinct subtypes of GBM;" . _:b463924220 . _:b11235312 . _:b11235325 . _:b11235313 . _:b11235313 "GBM can be divided into subtypes on the basis of global gene expression [>>7<<], [10]. Using the same approach as published [10], we showed that the mouse tumors are most closely related to the Proneural subtype." . _:b463924292 . _:b11235314 . _:b11235315 . _:b11235332 . _:b11235324 . _:b463924251 . _:b11235347 "The majority of these patients belongs to the Proneural subtype and is associated with longer survival [>>73<<], [74]. Interestingly, these G-CIMP+ patients are evenly distributed on our GSEA ranking (Table S1). Furthermore, removing these patients from the data set caused a similar reduction in median survival from both groups, but with limited" . _:b11235325 . _:b11235348 "The majority of these patients belongs to the Proneural subtype and is associated with longer survival [73], [>>74<<]. Interestingly, these G-CIMP+ patients are evenly distributed on our GSEA ranking (Table S1). Furthermore, removing these patients from the data set caused a similar reduction in median survival from both groups, but with limited overall" . _:b11235326 . . _:b11235327 . . _:b11235320 . _:b463924242 . _:b11235321 . _:b11235288 . _:b11235322 . _:b463924229 . _:b11235323 . . _:b11235335 "In particular, one study using the RCAS/tv-a system has shown that PDGF is sufficient to induce neonatal OPCs to form gliomas that resemble low-grade oligodendrogliomas [>>21<<]. In our model the tumors show the histological features of GBM. Furthermore, we show that these tumors have a Proneural phenotype and are also highly enriched in expressing OPC genes. While it is possible that PDGF signaling plays a role" . _:b11235239 "of cells in the adult brain that may have the capacity to form brain tumors, including neural stem cells in the subventricular zone (SVZ) [18], [25], [26] and more differentiated glial progenitors in the subcortical white matter [>>17<<], [28]. Furthermore, experiments using the RCAS/tv-a system have shown that progenitors with the capacity to form tumors are not restricted to the SVZ, but are distributed throughout the adult brain [29]." . . _:b463924241 . . . _:b463924270 . _:b11235257 "These cells likely represent uninfected glial progenitors that are being induced to proliferate via paracrine growth factor stimulation, as previously described [17], [22], [>>42<<]. However, the YFP- cells at 21 dpi accounted for only \u223C6% of total number of proliferating cells (data not shown). This indicates that the retrovirus infected cells, with Pten deleted, have a significant selective advantage, and quickly" . . _:b463924215 . _:b11235230 . . . _:b11235309 "Furthermore, evidence suggested that GBM arising outside the SVZ may have distinct growth characteristics and clinical outcomes [54], [>>55<<]. Therefore, characterizing the tumorigenic potential of cells that reside outside the SVZ is of both scientific and clinical interest." . _:b11235345 "However, the p53 pathway can be compromised by alterations other than mutations in p53 itself, such as MDM2 amplification [>>72<<]. Although our DGE list was identified by comparing tumors with and without p53 deletion as an initiating genetic lesion, this gene list likely captures the global changes in expression signatures that result from overall alterations in" . _:b11235244 . _:b11235248 "Furthermore, these different progenitor cells are susceptible to the genetic alterations seen in the particular tumor subtype to which they give rise [38], [>>39<<]. Specific genetic lesions are also observed in different subtypes of GBM. Mutations and loss of heterozygosity of p53 and amplification of PDGFR\u03B1 are most frequently seen in the Proneural subtype, while loss of Pten is observed throughout" . _:b11235279 "org/cgi-bin/cancer/publications/pub_paper.cgi? mode=view&paper_id=162&p=t) [7], [>>10<<], [44], [45]." . _:b11235309 . . _:b11235235 . _:b11235229 "GBM is remarkably heterogeneous, and may actually represent several distinct entities with different cells of origin, different genetic lesions and different clinical behaviors [1], [>>2<<]. Numerous studies have characterized the expression profiles and genetic alterations found in GBM [3]\u2013[9]. A recent comprehensive analysis of The Cancer Genome Atlas (TCGA) dataset has identified four distinct subtypes of GBM; Proneural," . _:b463924225 . . _:b463924276 . _:b11235260 "We then interrogated a murine dataset that was generated from neural lineages isolated from mouse brain [>>43<<]. From this study we compiled a list of 369 genes that were derived from 5 gene lists as published [43], which represented signatures of five cell types:" . _:b463924248 . _:b11235312 "Retroviruses only infect dividing cells [56], and therefore, can be used to selectively target progenitor cells in the brain [57]\u2013[>>59<<]. Using a stop-floxed YFP reporter, we were able to identify and fate map the cells that had been infected by PDGF-IRES-Cre expressing retroviruses. The distinct advantages of this approach allow us not only to control the timing and" . . _:b11235306 . _:b463924246 . _:b11235302 "Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [18]\u2013[>>20<<], [25], [26], [51], [52]." . _:b11235331 "Similarly, several experimental models have shown that PDGF stimulation can induce the formation of gliomas [11], [12], [>>14<<], [17], [22], [42]." . . _:b11235246 . _:b11235275 "For immunocytochemical staining, 3\u00D7104 cells per well were plated on 8 well culture chamber slides overnight. Cells were then fixed and stained the next day as described [>>17<<]." . . _:b11235311 . _:b463924281 . _:b11235252 "results" . _:b11235325 "PDGFR\u03B1 is selectively expressed in OPCs in the normal brain [60]\u2013[>>64<<], and the ability of PDGF signaling to stimulate the proliferation and migration of OPCs is well established [61], [64]\u2013[69]." . . . _:b11235330 "Similarly, several experimental models have shown that PDGF stimulation can induce the formation of gliomas [11], [>>12<<], [14], [17], [22], [42]." . _:b11235263 "We extended the GSEA analysis to three additional datasets [7], [>>44<<], [45]. Out of the combined 176 patients, 48 were classified as Proneural GBM (Verhaak RG, Hoadley KA and Hayes DN, personal communications)." . . _:b11235269 . _:b11235282 "We applied two approaches to classify the mouse samples, both of which used the mouse orthologs mapped from the original human classifier genes [>>10<<]. In the first approach we modified the centroid classifier as published [10]. We re-computed the centroids using 723, instead of 840, classifier genes." . _:b11235317 "In fact, 94 of the 210 (>45%) Proneural classifier genes are enriched in OPCs [10], [>>43<<] (Table S1)." . . _:b11235274 . _:b11235281 . _:b463924247 . _:b11235249 "Mutations and loss of heterozygosity of p53 and amplification of PDGFR\u03B1 are most frequently seen in the Proneural subtype, while loss of Pten is observed throughout all subtypes [>>8<<]\u2013[10], [40]." . _:b11235275 . . . _:b463924250 . _:b11235229 . . _:b11235301 "Many studies have provided evidence that neural stem cells in the SVZ have the capacity to form brain tumors [>>18<<]\u2013[20], [25], [26], [51], [52]." . . _:b11235265 . . _:b463924221 . _:b463924288 . _:b11235318 "It was previously shown that Proneural GBM express genes that are associated with the oligodendrocyte lineage [>>10<<]. To further explore this issue, we expanded the analysis to compare the expression of genes that are enriched in OPCs to that of other cell types isolated from the mouse brain [43]. We showed that the mouse tumors and human Proneural GBM" . . _:b11235292 . _:b463924238 . _:b11235333 "Similarly, several experimental models have shown that PDGF stimulation can induce the formation of gliomas [11], [12], [14], [17], [>>22<<], [42]. In particular, one study using the RCAS/tv-a system has shown that PDGF is sufficient to induce neonatal OPCs to form gliomas that resemble low-grade oligodendrogliomas [21]. In our model the tumors show the histological features" . _:b11235348 . _:b11235310 . _:b463924294 . _:b463924208 . _:b11235320 "Certain common genetic alterations tend to occur in specific GBM subtypes [>>10<<]. This raises the possibility that these genetic alterations are particularly effective at transforming the cell type that gives rise to the specific GBM subtype." . _:b11235323 "Furthermore, proteomic studies revealed that activation of PDGF signaling is associated with Proneural phenotype [>>40<<]. PDGFR\u03B1 is selectively expressed in OPCs in the normal brain [60]\u2013[64], and the ability of PDGF signaling to stimulate the proliferation and migration of OPCs is well established [61], [64]\u2013[69]. Similarly, several experimental models" .