ABSTRACT: Background: Signaling by receptor tyrosine kinases (RTK) is frequently dysregulated in gliomas. Inter-individual variability in the causes for dysregulated RTK signaling may have hampered the efficacy of targeted therapies. Using gene expression modules around key regulators in the RAS-RAF-MEK-MAPK cascade and in the phosphatidylinositol 3-kinase-AKT pathways, we developed a “RMPA” clustering scheme to distinguish gliomas with varying extents of RTK signaling. Results: We identified gene modules consistently co-expressed with NF1 (NF1-M), Sprouty (SPRY-M) and PTEN (PTEN-M) in gliomas. Their signatures enabled robust clustering of adult diffuse gliomas of WHO grades II-IV into RMPAhigh and RMPAlow phenotypes in a morphology-independent manner. In five independent data sets from three continents containing more than 1500 adult diffuse gliomas, RMPAhigh gliomas were associated with poor prognosis while RMPAlow gliomas were not. The RMPAhigh and RMPAlow glioma subtypes showed distinct levels of the activities of RAS-RAF-MEK-MAPK cascade and PI3K-AKT pathway and harbored unique sets of genomic alterations in the RTK signaling-related genes. The RMPAhigh gliomas contained large numbers of immature vessel cells and tumor associated macrophages and both cell types expressed high levels of pro-angiogenic RTKs including MET, VEGFR1, KDR, EPHB4 and NRP1. Conclusion: Inter-glioma variability in RTK signaling activities can be defined using the RMPA clustering scheme. The combined signatures of NF1-M, SPRY-M and PTEN-M reflect RTK signaling activity both in the glioma cells and in the glioma microenvironment. Our data show that RTK signaling in the glioma microenvironment may play a pivotal role in glioma progression. Transcriptome data from 22 fresh gliomas (2 astrocytoma II, 1 oligodendrocytoma II, 7 oligoastrocytoma II, 4 anaplastic oligoastrocytoma III and 8 GBM) were obtained using Affymetrix Human Gene 1.0 ST Array. Unsupervised hierarchical clustering of the expression data for the SPRY-M, NF1-M and PTEN-M was performed on these transcriptome data to identify samples with RMPAhigh or RMPAlow signature.
Project description:Primary glioblastoma, representing over 90% of adult glioblastoma, develop rapidly without preexisting lower-grade glioma. We have generated a mouse model of primary glioblastoma driven by a single p53 mutation. These p53-mutant gliomas lose the syntenic region of human chromosome 10q, which is mapped to mouse chr19 and chr7. Loss of mouse chr19, containing Pten, activates PI3K/Akt signaling. Rictor/mTORC2 deletion inhibits Akt signaling, causing a significant delay in p53-mutant driven glioma formation. Unexpectedly, Rictor/mTORC2 loss promotes p53-mutant driven medulloblastomas with unique features of pediatric SHH medulloblastoma. Mechanistically, Rictor/mTORC2 loss inhibits the generation of glioma precursor cells from neural stem/progenitor cells in the adult brain, while causing a delay in differentiation of granule cell precursors in the developing brain, a cell-of-origin of SHH medulloblastoma. Overall design: In this study we generate mouse models with (1) p53 loss (2) p53 loss combined with gemline Nf1 mutation (3) a conditional cis-p53/Nf1 mutation. We profiled the resultant tumors as well as normal forebrain and cerebellum as controls.
Project description:Gioma stem cells from Nf1-associated optic gliomas and NSCs from third ventricle zone show differencies in gene expression patterns and physiological functions. We used microarrays to identify differential gene expression between glioma stem cells and the control counterparts including the wild type and Nf1-/- neural stem cells from the third ventricle. Overall design: Cultured glioma stem cells were isolated from mouse optic gliomas. Wild-type and Nf1-/- TVZ NSCs were generated from postnantal day 1 Nf1flox/flox pups. We sought to obtain neural stem cells from three independent cultures. There are 9 samples in total. 3 wild-type TVZ NSCs, 3 Nf1-/- TVZ NSCs, 3 glioma stem cells.
Project description:The concept that solid tumors are maintained by a productive interplay between neoplastic and non-neoplastic elements has gained traction with the demonstration that stromal fibroblasts and immune system cells dictate cancer development and progression. While less studied, brain tumor (glioma) biology is likewise influenced by non-neoplastic immune system cells (macrophages and microglia) which interact with neoplastic glioma cells to create a unique physiological state (glioma ecosystem) distinct from that found in the normal tissue. To explore this neoplastic ground state, we leveraged several preclinical mouse models of neurofibromatosis type 1 (NF1) optic glioma, a low-grade astrocytoma whose formation and maintenance requires productive interactions between non-neoplastic and neoplastic cells, and employed whole tumor RNA-sequencing and mathematical deconvolution strategies to characterize this low-grade glioma ecosystem as an aggregate of cellular and acellular elements. Using this approach, we demonstrate that optic gliomas generated by altering the germline Nf1 gene mutation, the glioma cell of origin, or the presence of co-existing genetic alterations represent molecularly-distinct tumors. However, these optic glioma tumors share a 25-gene core signature, not found in normal optic nerve, that is normalized by microglia inhibition (minocycline), but not conventional (carboplatin) or molecularly-targeted (rapamycin) chemotherapy. Lastly, we identify a genetic signature conferred by Pten reduction and corrected by PI3K inhibition. This signature predicts progression-free survival in patients with either low-grade or high-grade glioma. Collectively, these findings support the concept that gliomas are composite ecological systems whose biology and response to therapy may be best defined by examining the tumor as a whole. Overall design: We profiled homogenized optic nerve, optic glioma tissue, and cell lines from mouse models and used a template-switching-based library construction protocol to faciliate low-input RNA-Seq.
Project description:miRNA expression analysis of mouse glioma and PNET. Glioma was developed from SVZ cells through conditional codeletion of Pten/p53 or Rb/p53; while PNET was developed by codeletion of Rb/p53. Overall design: Pten/p53 codeleted SVZ cells always developed gliomas, while Rb/p53 cells mainly gave rise to PNET, but a small amout of Rb/p53 mice developed gliomas. Three groups: Pten/p53 glioma, Rb/p53 glioma and Rb/p53 PNET; biological control: each RNA sample was from a mouse tumour, three samples in each group.
Project description:Mosaic Analysis with Double Markers (MADM) based glioma mouse model, which homozygously lacks Tp53 and Nf1, spontaneously developed gliomas at the post-natal 90-120 days. Tp53 and Nf1 are among the most frequently mutated genes in human glioma patients. Investigating the expression changes of genes induced by inactivation of Tp53 and Nf1 can be a clue to clarify the mechanism of gliomagenesis. We examined the expressions of glioma in MADM mouse at post-natal 150 days (n=3) and of normal brain in Tp53 and Nf1 wild type mouse at post-natal 150 days (n=2). We used SurePrint G3 Mouse GE 8×60K array slides (G4858A, Agilent Technologies). Overall design: Microarray experiment was performed as single-color hybridization on SurePrint G3 Mouse GE 8 x 60K array.
Project description:Solid cancers develop within a supportive microenvironment that promotes tumor formation and continued growth through the elaboration of mitogens and chemokines. Within these tumors, monocytes (macrophages and microglia) represent rich sources of these stromal factors. Leveraging a genetically-engineered mouse model of neurofibromatosis type 1 (NF1) low-grade brain tumor (optic glioma), previous studies have demonstrated that microglia are important for glioma formation and maintenance. To identify the tumor-associated microglial factors that support glioma growth (gliomagens), we employed a comprehensive large scale discovery effort using optimized advanced RNA-sequencing methods. Candidate gliomagens were prioritized to identify potential secreted or membrane-bound proteins, which were next validated by quantitative RT-PCR and RNA FISH following minocycline-mediated microglial inactivation in vivo. Using these selection criteria, Ccl5 was identified as a highly expressed chemokine in both genetically engineered Nf1 mouse and human optic gliomas. As a candidate gliomagen, recombinant Ccl5 increased Nf1-deficient optic nerve astrocyte growth in vitro. Importantly, consistent with its critical role in maintaining tumor growth, Ccl5 inhibition with neutralizing antibodies reduced Nf1 mouse optic glioma growth in vivo. Collectively, these findings establish Ccl5 as critical stromal growth factor in low-grade glioma maintenance relevant to future microglia-targeted therapies for brain tumors. Overall design: Nf1 optic glioma associated microglia from mice were flow sorted. Upregulated genes of glioma associated microglia were verified and further examined.
Project description:Gliomas are aggressive primary brain tumors, presenting surgery limitations due to their highly infiltrative potential. The expression of Angiotensin II (Ang II) receptors in human astrocytomas was previously associated with a poor prognosis. Accordingly, this study was undertaken to reveal the molecular mechanisms underlying Ang II actions in gliomas through the transcriptomic analysis of glioma cells exposed to Ang II. C6 glioma cells were treated with Ang II and specific antagonists of AT1 and AT2. Total RNA was isolated at three and six hours intervals and submitted to oligonucleotide microarray protocol. The differentially expressed genes were obtained by paired t-tests with p<0.05 and interpreted using Venn diagrams, functional enrichment and protein interaction network analyses. Validation of microarray results was carried out through qPCR experiments of selected genes. We found a high number of significant genes with low fold changes in gene expression at the time intervals studied. These genes were regulated in a time dependent-manner, with most gene expression changes being exclusive to one of the time intervals evaluated. Our results indicated that blocking AT1 or AT2 changed the expression of genes involved in regulation of transcription, cell cycle, cell proliferation, differentiation, apoptosis, cell adhesion, cell migration, regulation of actin cytoskeleton, protein transport and protein ubiquitination. Additionally, the signaling pathways over-represented by the significant genes were ErbB, mTOR, MAPK, neurotrophin, insulin and Wnt. Finally, interactome analyses revealed hub genes associated with cell proliferation, survival, migration, transport, structural support, neurotrophin pathway, MAPK signaling and Wnt signaling. Taken together, our findings implicate Ang II-transcriptional regulation in glioma progression by means of the modulation of genes participating in protumoral functions. This transcriptome pattern is observed upon Ang II activation of either AT1 or AT2 receptors, thereby highlighting the relevance of both receptor subtypes in glioma progression. Interactome analyses disclosed hub genes regulated by Ang II which may present higher control over their networks. These genes participate in biological functions that could enhance the degree of malignancy in gliomas and thus should be further explored. C6 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 100 Units/ml penicillin and 100 µg/ml streptomycin. Cells were seeded in cell culture dishes and incubated at 37°C/ 5% CO2 until becoming confluent. Then, these cells were pre-treated (30 minutes) with either AT1 receptor antagonist (Losartan: 10-5M) or AT2 receptor antagonist (PD123319: 10-5M) followed by Ang II treatment (10-7 M) according to the treatment scheme: Group 1 – control; Group 2 – cells only treated with Ang II; Group 3 – cells pre-treated (30 minutes) with Losartan and then treated with Ang II; Group 4 – cells pre-treated (30 minutes) with PD123319 and then treated with Ang II. To identify which genes were significantly differentially expressed, paired t-tests (p<0.05) were performed in the following comparisons: Ang II x Control (3h); Ang II x Control (6h); Ang II +Los x Ang II (3h); Ang II +Los x Ang II (6h); Ang II +PD123319 x Ang II (3h); Ang II +PD123319 x Ang II (6h).
Project description:How malignant gliomas arise in a mature brain remains a mystery, hindering the development of preventive and therapeutic interventions. We previously showed that oligodendrocyte precursor cells (OPCs) can be transformed into glioma when mutations are introduced perinatally. However, adult OPCs rarely proliferate compared to their perinatal counterparts. Whether these relatively quiescent cells have the potential to transform is unknown, which is a critical question considering the late onset of human glioma. Additionally, the events taking place between initial mutation and a fully developed tumor mass (pre-malignant phase) are particularly poorly understood in glioma. Here we used a temporally controllable Cre transgene to delete p53 and NF1 specifically in adult OPCs, and demonstrated that these cells consistently give rise to malignant gliomas. To investigate the transforming process of quiescent adult OPCs, we then tracked these cells throughout the pre-malignant phase, which revealed a dynamic multi-step transformation, starting with rapid but transient hyper-proliferative reactivation, followed by a long period of dormancy, then final malignant transformation. Using pharmacological approaches, we discovered that mTOR signaling is critical for both the initial OPC reactivation step and late stage tumor cell proliferation, and thus might be a potential target for both glioma prevention and treatment. In summary, our results firmly establish the transforming potential of adult OPCs, and reveal an actionable multi-phasic reactivation process that turns slowly dividing OPCs into malignant gliomas. 44K Mouse Development Oligo Microarrays from Agilent Technologies were used for microarray analysis. For each experiment, total RNA was fluorescently labeled and hybridized directly against a common reference sample generated from the RNA pool of four WT P17 mouse brain neocortex.
Project description:Comparison of polysomal profiles of murine adult olig2 cortical progenitors, murine tumor olig2 cells derived from hPDGF-B-driven glioblastomas, and murine olig2 proliferative recruited glioma cells contributing to the tumor mass but not derived from the cell of origin An unbiased way to define a cell population lies in identification of its gene expression signature and subsequent comparison to gene expression signatures of known normal or cancer cells to define its position on the axis of tumorigenesis. To quantify similarities and differences in the expression profiles of recruited cells and tumor cells using microarray analysis, we used the bacTRAP technology that allows immunoprecipitation of polysomes from specific cell types in vivo (Heiman et al 2008, Doyle et al 2008). We compared polysomal expression profiles of hPDGFb-driven Ntv-a Ink4a/Arf+/- olig2 mouse glioma cells representing a histologically defined glioma population, recruited olig2 cells derived from Ink4a/Arf+/- olig2 RP-eGFP mice transplanted with non-fluorescent glioma cells, and normal adult cortical olig2 mouse progenitors, using Affymetrix 430 mouse 2.0 chips and Genespring GX10 software (Doyle et al 2008, Heiman et al 2008). Translational profiles of recruited olig2 cells and glioma olig2 cells were similar to each other and clearly distinct from the normal olig2 progenitors, majority of the differences accounted for by the “statistically significantly changed” set of mRNAs (ANOVA, p<0.05), and sample clustering reduced upon removal of the ANOVA-tested mRNA set. Unsupervised hierarchical mRNA clustering likewise indicated that recruited olig2 cells clustered more closely to glioma olig2 cells. The caveat is that olig2 expression in the adult normal brain of bacTRAP olig2 RP-eGFP mice labels both progenitors and mature oligodendrocytes, while recruited and tumor olig2 cells are almost entirely progenitors; therefore, some of the differences seen in the comparative analysis of normal, tumor and recruited olig2 cells may in part result from the shift in relative abundance of different progenitor and mature olig2 cell populations. Normal olig2 progenitors were collected from three replicates of pooled cortices of three olig2 RP-eGFP bacTRAP reporter mice. Mouse gliomas induced by injection of a non-fluorescent RCAS-hPDGFb (Shih et al 2004) or by transplantation of muring glioma cells derived from Ntv-a gliomas induced by the non-fluorescent RCAS-hPDGFb were collected from Ntv-a olig2 RP-eGFP bacTRAP reporter mice and immunoprecipitated; each tumor was processed as a separate sample. Briefly, mouse tissues were collected into ice-cold cyclohexamide-containing buffer, homogenized, cells were lysed in NP-40 and DHPC-containing buffer, centrifuged at 20,000g for 15 min, supernatant incubated with anti-eGFP-conjugated protein G beads 30 min at 4C, washed and RNA collected using Trizol reagent as per manufacturer’s instructions (Doyle et al., 2008; Heiman et al., 2008). RNA was purified, concentrated using Qiagen RNeasy kit, its quality confirmed by Agilent Bioanalyzer, 15ng per sample amplified using Affymetrix Two-Cycle Amplification kit and hybridized to Affymetrix 430 mouse 2.0 chips as described in Doyle et al., 2008; Heiman et al., 2008. Data was analyzed using Genespring GX10 software, as described in the text and in Doyle et al., 2008; Heiman et al., 2008. Briefly, samples were normalized using GCRMA, filtered to remove probe sets with low intensity, and analyzed using ANOVA (p<0.05) with SNK post-hoc and Benjamini and Hochberg FDR multiple testing correction. Probe quality, Pearson’s correlation, hierarchical clustering, heatmaps, gene profile plots were generated using Genespring GX10 software. Pathway analysis was performed using Ingenuity as per manufacturer’s instructions. Gliomas are thought to form by clonal expansion from a single cell-of-origin, and progression-associated mutations to occur in its progeny cells. Glioma progression is associated with elevated growth factor signaling and loss of function of tumor suppressors Ink4a, Arf and Pten. Yet, gliomas are cellularly heterogeneous; they recruit and trap normal cells during infiltration. We performed lineage tracing in a retrovirally mediated, molecularly and histologically accurate mouse model of hPDGFb-driven gliomagenesis. We were able to distinguish cells in the tumor that were derived from the cell-of-origin from those that were not. Phenotypic, tumorogenic and expression analyses were performed on both populations of these cells. Here we show that during progression of hPDGFb-induced murine gliomas, tumor suppressor loss can expand the recruited cell population not derived from the cell-of-origin within glioma microenvironment to dominate regions of the tumor, with essentially no contribution from the progeny of glioma cell-of-origin. Moreover, the recruited cells can give rise to gliomas upon transplantation and passaging, acquire polysomal expression profiles and genetic aberrations typically present in glioma cells rather than normal progenitors, aid progeny cells in glioma initiation upon transplantation, and become independent of PDGFR signaling. These results indicate that non-cell-of-origin derived cells within glioma environment in the mouse can be corrupted to become bona fide tumor, and deviate from the generally established view of gliomagenesis.
Project description:Diffuse gliomas represent the most prevalent class of primary brain tumor. Despite significant recent advances in the understanding of glioblastoma (WHO IV), its most malignant subtype, lower-grade (WHO II and III) glioma variants remain comparatively understudied, especially in light of their notably variable clinical behavior. To examine the foundations of this heterogeneity, we performed multidimensional molecular profiling, including global transcriptional analysis, on 101 lower-grade diffuse astrocytic gliomas collected at our own institution, and validated our findings using publically available gene expression and copy number data from large independent patient cohorts. We found that IDH mutational status delineated molecularly and clinically distinct glioma subsets, with IDH mutant (IDH mt) tumors exhibiting TP53 mutations, PDGFRA overexpression, and prolonged survival, and IDH wild-type (IDH wt) tumors exhibiting EGFR amplification, PTEN loss, and unfavorable disease outcome. Furthermore, global expression profiling revealed three robust molecular subclasses within lower-grade diffuse astrocytic gliomas, two of which were predominantly IDH mt and one almost entirely IDH wt. IDH mt subclasses were distinguished from each other on the basis of TP53 mutations, DNA copy number abnormalities, and links to distinct stages of neurogenesis in the subventricular zone (SVZ). This latter finding implicates discrete pools of neuroglial progenitors as cells of origin for the different subclasses of IDH mt tumors. In summary, we have elucidated molecularly distinct subclasses of lower-grade diffuse astrocytic glioma that dictate clinical behavior and demonstrate fundamental associations with both IDH mutational status and neuroglial developmental stage. 80 tumor samples, one normal tissue sample (brain)