Skip to main content
Neuro-Oncology logoLink to Neuro-Oncology
. 2022 Apr 25;24(11):1911–1924. doi: 10.1093/neuonc/noac110

MEOX2 homeobox gene promotes growth of malignant gliomas

Anna Schönrock 1,2,#, Elisa Heinzelmann 3,4,5,#, Bianca Steffl 6,#, Engin Demirdizen 7,#, Ashwin Narayanan 8, Damir Krunic 9, Marion Bähr 10, Jong-Whi Park 11,2, Claudia Schmidt 12, Koray Özduman 13, M Necmettin Pamir 14, Wolfgang Wick 15,16, Felix Bestvater 17, Dieter Weichenhan 18, Christoph Plass 19, Julian Taranda 20, Moritz Mall 21,22,23, Şevin Turcan 24,
PMCID: PMC9629421  PMID: 35468210

Abstract

Background

Glioblastoma (GBM) is an aggressive tumor that frequently exhibits gain of chromosome 7, loss of chromosome 10, and aberrantly activated receptor tyrosine kinase signaling pathways. Previously, we identified Mesenchyme Homeobox 2 (MEOX2), a gene located on chromosome 7, as an upregulated transcription factor in GBM. Overexpressed transcription factors can be involved in driving GBM. Here, we aimed to address the role of MEOX2 in GBM.

Methods

Patient-derived GBM tumorspheres were used to constitutively knockdown or overexpress MEOX2 and subjected to in vitro assays including western blot to assess ERK phosphorylation. Cerebral organoid models were used to investigate the role of MEOX2 in growth initiation. Intracranial mouse implantation models were used to assess the tumorigenic potential of MEOX2. RNA-sequencing, ACT-seq, and CUT&Tag were used to identify MEOX2 target genes.

Results

MEOX2 enhanced ERK signaling through a feed-forward mechanism. We identified Ser155 as a putative ERK-dependent phosphorylation site upstream of the homeobox-domain of MEOX2. S155A substitution had a major effect on MEOX2 protein levels and altered its subnuclear localization. MEOX2 overexpression cooperated with p53 and PTEN loss in cerebral organoid models of human malignant gliomas to induce cell proliferation. Using high-throughput genomics, we identified putative transcriptional target genes of MEOX2 in patient-derived GBM tumorsphere models and a fresh frozen GBM tumor.

Conclusions

We identified MEOX2 as an oncogenic transcription regulator in GBM. MEOX2 increases proliferation in cerebral organoid models of GBM and feeds into ERK signaling that represents a core signaling pathway in GBM.

Keywords: cerebral organoids, ERK signaling, glioblastoma, homeobox, MEOX2


Key Points.

  1. MEOX2 is aberrantly upregulated in glioblastoma.

  2. MEOX2 leads to the activation of oncogenic molecular programs.

  3. Together with PTEN/p53 loss, MEOX2 induces proliferation in cerebral organoid models of GBM.

Importance of the Study.

In this study, we show that MEOX2, a transcription factor with no known role or expression in the CNS, is significantly upregulated by GBM. In addition, MEOX2 cooperates with the loss of tumor suppressors PTEN and p53 to promote tumor growth. These results suggest that MEOX2 is a candidate oncogene on chromosome 7 that may be co-opted during tumor initiation. As our knowledge of aberrant transcription factor networks in GBM increases and innovative approaches to target transcription factors emerge, targeting GBM-associated transcription factors such as MEOX2 or their associated transcriptional networks may eventually lead to a clinical benefit.

Glioblastoma multiforme (GBM) is a highly proliferative and invasive brain tumor. According to the 2021 WHO classification of CNS tumors,1 GBM refers to Glioblastoma, IDH wild-type and common genetic alterations include gain of chromosome 7, loss of chromosome 10, alterations in TP53, epidermal growth factor receptor (EGFR), and platelet-derived growth factor receptor (PDGFR), and mutations in mouse double minute homolog 2 (MDM2) and phosphatase and tensin homolog (PTEN) gene.2 These alterations lead to activation of multiple signaling pathways including, p53, receptor tyrosine kinase (RTK)/Ras/phosphoinositide 3-kinase (PI3K), and retinoblastoma (Rb) signaling pathways.3 These molecular alterations promote GBM growth, survival, and facilitate escape from cell-cycle checkpoints, senescence, and apoptosis.

In addition to genomic alterations, aberrantly overexpressed transcription factors have the potential to drive GBM.4,5 In GBM, we previously identified MEOX2 as a highly expressed transcription factor,6 and a recent study has linked MEOX2 expression to poor prognosis.7MEOX2 is localized to chromosome 7, which is gained early in gliomagenesis.8 Several genes on chromosome 7, such as platelet-derived growth factor (PDGFA), homeobox A5 (HOXA5), HOXA9, and HOX10 have been shown to drive tumor aggressiveness and radio-resistance.8–12

MEOX2 is a homeobox-containing transcription factor that plays an essential role in developing tissues.13,14 It activates p16 and p21 through DNA-dependent and independent mechanisms, and increased MEOX2 expression leads to cell cycle arrest and endothelial cell senescence.15,16 In cancer, the role of MEOX2 is poorly defined and likely context-dependent.17–20

Here, we hypothesized that MEOX2 could provide a selective advantage in GBM. We show that MEOX2 overexpression can enhance tumor growth in vivo and leads to a significant growth advantage in human cerebral organoid models of GBM, highlighting the potential role of MEOX2 as an early driver of gliomagenesis. Mechanistically, we find that gene regulatory activity of MEOX2 is likely controlled by its phosphorylation level, which in turn leads to increased ERK signaling. Our results suggest that MEOX2 is likely one of the key oncogenes on chromosome 7 that is co-opted by GBM to drive gliomagenesis.

Methods

Additional details about materials and methods can be found in the Supplementary Methods.

Patient Sample

The fresh-frozen GBM sample was obtained from a patient following surgical resection at the Department of Neurosurgery at the Acıbadem Mehmet Ali Aydınlar University. Use of patient material was approved by the Institutional Review Board at the Medical Faculty of Acıbadem Mehmet Ali Aydınlar University. Informed consent was obtained from the patient included in the study.

Cell Culture

Patient-derived GBM lines were maintained in Neurocult Basal Medium with proliferation supplements, 20 ng/mL EGF, 20 ng/mL basic-FGF, and 2 μg/mL Heparin (StemCell Technologies). L0125 and L0512 GBM tumorspheres were kindly provided by Dr. Rossella Galli (San Raffaele Scientific Institute, Milan, Italy). Immortalized human astrocytes were a gift of Dr. Russell O. Pieper (UCSF).

Viral Transductions

TS667 and TS600 lines were infected with lentiCRISPRv2 encoding either nontargeting control or sgRNA targeting MEOX2 and selected with 0.5 μg/mL of puromycin. MEOX2 was stably overexpressed in L0125 and L0512 cell lines using pLVX-puro-MEOX2-FLAG or pLVX-puro-MEOX2-S155A-FLAG plasmid. Additional luciferase expression was transduced using pLenti-PGK-V5-LUC Neo plasmid and selected with 400 μg/mL of G418.

Site-Directed Mutagenesis

Ser155 of MEOX2 was mutated using QuikChange Site-Directed Mutagenesis Kit (Agilent). The mutagenic primers were CCGCCAGGCACTGGCACCTGCGGAGGC, which converted Ser155 to alanine (S155A), and CCGCCAGGCACTGGATCCTGCGGAGGC, which converted Ser155 to aspartate (S155D). Mutation was confirmed by Sanger sequencing.

Trametinib Treatment

HEK cells were transfected 24 h after seeding. Medium was changed to trametinib (100 nM) containing medium 24 h after transfections. Treated cells were harvested after 24 h of trametinib treatment.

SDS PAGE and Western Blot

Cells were lysed using NP40 lysis buffer containing 1% Halt Protease and Phosphatase Inhibitor Cocktail (ThermoFisher). For EGF stimulation experiments, TS667, TS600, and TS543 cells were seeded without EGF/bFGF supplementation, cultured for 48 h; afterwards medium was changed to neural stem cell media containing 50 ng/mL EGF. Cells were then collected at different timepoints (0 min, 10 min, 30 min, 2 h, 4 h, 6 h, 8 h) to observe pathway initiation by EGF treatment.

Orthotopic Transplantation

All mouse experiments were approved by the Institutional Animal Care and Use Committee at DKFZ. Female athymic nude mice at age of 8 weeks (n=7 per group) were used. 100,000 cells were intracranially injected 3 mm deep into the striatum at a speed of 1 μL/min. The following coordinates from bregma were used: anterio-posterior = 0; medio-lateral= +2.5mm; dorso-ventral= –3.0 mm.

Cerebral Organoid Model

Human induced pluripotent stem cells (IPSCs) were treated with TripleLE dissociation reagent to obtain single cells and plated (9000 cells/well) into an ultra-low-binding 96-well plate in mTeSR™1 medium (Stemcell Technologies) containing ROCK inhibitor (1:200). After three days, EBs, now induced to cerebral organoid (COs), were pooled into a 6 cm dish containing fresh cortical induction medium. The next day, COs were nucleofected to introduce TP53, PTEN, and neurofibromin (NF1) using CRISPR-mediated deletion. Each nucleofection reaction contained a stably integrating PiggyBac plasmid, expressing mNeonGreen with or without additional MEOX2 expression. The nucleofected COs were embedded into Geltrex and transferred to a low-attachment 6 cm dish containing cortical induction medium. The next day, COs were transferred to the orbital shaker. From the following day, medium was replaced by differentiation medium with medium change every 3–4 days. After four weeks, mNeonGreen expression in the organoids were imaged using a fluorescence microscope (Leica). For each organoid, three images in different Z-dimensions were acquired. ImageJ was used to measure mNeonGreen intensity of each organoid by calculating the corrected total cell fluorescence (CTCF). This method determines the corrected total fluorescence by subtracting out background signal, which is useful for comparing the fluorescence intensity between different organoids. Briefly, all three images of one organoid were stacked and maximum intensity projection was performed. Within the stacked organoid image, all fluorescent areas were outlined and integrated density and mean grey value were measured. For background subtraction, three small areas within the same image were selected that have no fluorescence and the mean fluorescence of background readings was calculated. Finally, the CTCF value was calculated using the following formula: CTCF = Integrated Density – (Area of Selected Cell x Mean Fluorescence of Background readings). The CTCF value for each organoid was normalized to the whole area of the corresponding organoid.

Immunofluorescence

A total of 5000 cells were seeded, washed, and fixed with paraformaldehyde. The fixed cells were incubated with primary antibody at room temperature, followed by secondary antibody incubations for 1 h at room temperature. Coverslips were mounted on slides using mounting media containing DAPI (VECTASHIELD).

RNA-Sequencing and Data Analysis

Total RNA was extracted using the QIAGEN RNeasy RNA isolation kit. Sequencing was performed by the DKFZ genomics core facility. The aligned bam files and feature counts were generated by the Omics IT and data management core facility at DKFZ. Normalization of raw counts and differential expression analysis were performed using DESeq2 R package.21 An absolute log2 fold-change of 0.5 and adjusted P-value of .05 was used to identify the differentially expressed genes.

Antibody-Guided Chromatin Tagmentation Sequencing (ACT-seq) and Cleavage Under Targets and Tagmentation (CUT&Tag) Sequencing

ACT-seq was largely performed according to Carter et al.22 as described in detail recently23 on triplicates of NTC and KD2 cells. For CUT&Tag experiments, nuclei isolation was performed according to the published protocol24 using a minimum of 105 nuclei. Nuclei were resuspended in 200 μl antibody buffer containing 4 μg MEOX2 antibody or IgG for control and incubated at 4ºC overnight. The Hyperactive In-Situ ChIP Library Prep Kit for Illumina (pG-Tn5) (Vazyme) was used for CUT&Tag. Bulk library was prepared using Kapa and QC was performed using D1000 Tapestation.

ACT-seq and CUT&Tag Analysis

The fastq files were trimmed using trimmomatic and aligned to hg19 with bowtie2 using parameters --end-to-end --very-sensitive --no-mixed --no-discordant. Mitochondrial and blacklisted regions were removed from the aligned bam files using samtools. Duplicates were removed using Picard tools (MarkDuplicates). Low quality reads (quality score < 2) were removed using samtools. Peaks were called using MACS version 2.1.2. Differentially bound sites were identified using DiffBind (R statistical package).

Immunoprecipitation and Mass Spectrometry

In total, 10 μg of MEOX2 antibody or IgG were added to each sample. After incubation and washing steps, proteins were removed from the beads using 100 μL of Low pH Elution Buffer. Tubes were incubated for 10 min at RT and supernatant was collected and neutralized using 15 μL of Neutralization Buffer per tube. Proteomics was performed at the Mass Spectrometry Core Facility at the German Cancer Center (DKFZ).

Antibodies

The following antibodies were used for all the experiments: MEOX2 (HPA053793, Sigma), FLAG M2 (F1804, Sigma-Aldrich), Phospho MAPK (#9101, Cell Signaling), MAPK (#9107, Cell Signaling), p21 (#2947, Cell Signaling), GAPDH (#2118, Cell Signaling), Actin (#4967, Cell Signaling), H3 (#9715, Cell Signaling), Rabbit IgG (PP64, Merck), Alexa Fluor 488 (A-21206, ThermoFisher), and Alexa Fluor 594 (A-21207, ThermoFisher).

Statistical Analysis

Two-sided t-test with Welsh’s correction, one-way ANOVA or two-way ANOVA were performed. Significance is indicated using following legend: ≥0.1234 (n.s.), ≤0. 0332 (*), ≤0.0021 (**), ≤0.0002 (***), ≤0.0001 (****).

Results and Discussion

MEOX2 is Highly Expressed in GBM

Analysis of combined molecular profiling datasets for lower-grade gliomas and GBM from TCGA revealed significantly higher MEOX2 expression in GBM (Figure 1A; Supplementary Figure S1A–C). MEOX2 expression in GBM was restricted to malignant cells as shown by the analysis of single-cell RNA-seq data by Neftel et al.25 (Figure 1B). In comparison, MEOX2 expression in the normal brain is low or undetectable (Supplementary Figure S1D). Nuclear MEOX2 staining in tumor cells was also conclusively shown by Tachon et al.7 in primary GBM patient samples.

Fig. 1.

Fig. 1

MEOX2 is expressed in GBM. (A) MEOX2 expression from The Cancer Genome Atlas (TCGA) database of IDH mutant and wild-type lower grade gliomas and GBM tumors. Data downloaded from GlioViz.44 Histopathological classification is based on the 2007 WHO classification of CNS tumors.45 (B) MEOX2 expression in published GBM single-cell RNA-seq data25 assigned to macrophages (n = 536), oligodendrocytes (n = 210), T cells (n = 80) and malignant tumor cells (n = 4916) (C) Pathway enrichment analysis of differentially upregulated genes in GBMs with high compared to low MEOX2 expression in the TCGA dataset. (D) p-ERK levels in TS667 nontargeting control (NTC) or MEOX2 KD (KD1, KD2). Cells were either kept under steady-state (ss) condition or starved (-) for 48 h and stimulated with 50 ng/ mL EGF for 10 or 30 min. (E) Quantification of (D) showing p-ERK relative to total ERK (n = 3). (F) p-ERK levels in TS543 shRNA control (scrambled) or MEOX2 KD (MEOX2 sh1, MEOX2 sh2). Cells were either kept under steady-state condition (EGF+) or starved for 48 h (0 min) and stimulated with 50 ng/ mL EGF for 10 or 30 min. (G) Quantification of (F) showing p-ERK relative to total ERK (n = 4). Mean ± SD. Two-way ANOVA with Tukey’s post-hoc test.

The transcriptional effects of MEOX2 in GBM are not known, we, therefore, aimed to elucidate the pathways that may be associated with MEOX2. We identified the top 50 genes positively correlated with MEOX2 expression in TCGA GBM dataset (Supplementary Table S1). Gene programs involving ERK1/2 cascade (eg, SPRY2, EGFR, PDGFA, P2RY1, PLA2G5, SPRY1, FGFR3, SPRY4) and fatty acid synthesis (ELOVL2, ACSL3, PLA2G5) correlated significantly with MEOX2 expression (Supplementary Table S2). Fatty acid metabolism is significantly altered in cancers, including gliomas.26,27 Furthermore, ELOVL2, an enzyme in polyunsaturated fatty acid synthesis, has been shown to maintain cell membrane organization and EGFR signaling in GBM.28 To determine the pathways associated with MEOX2 in GBMs, we stratified TCGA GBM tumors into two cohorts based on the first (low) and fourth quartile (high) of MEOX2 expression and identified the significantly deregulated pathways (Supplementary Table S3). Pathways including RTK activity, extracellular matrix, GBM signaling, and regulation of epithelial-mesenchymal transition (EMT) were significantly enriched among the upregulated genes (Figure 1C). In addition to the extreme groups approach, dividing the GBM samples into high and low groups based on median MEOX2 expression revealed enrichment of similar pathways (Supplementary Figure S2). Based on these findings and the observation that MEOX2 is upregulated in GBMs, we hypothesized that MEOX2 may function as an oncogene in GBM, potentially by enhancing signaling through RTKs.

MEOX2 Induces Phosphorylation of ERK

Given that MEOX2 expression correlates with oncogenes such as EGFR and hallmark oncogenic pathways such as RTK signaling and EMT in GBM (Figure 1C), we wondered whether MEOX2 may co-operate with signaling pathways activated by RTKs. We established MEOX2 knockdowns in GBM tumorsphere lines TS600 and TS667 using either CRISPR/Cas9 and two single-guide RNAs (KD1, KD2) (Supplementary Figure S3) or shRNA-mediated knockdown in TS543 line (sh1 and sh2). MEOX2 knockdown significantly reduced phospho-ERK (p-ERK) levels in the TS667 (Figure 1D,E) and TS543 (Figure 1F,G) lines, whereas p-ERK levels were not altered in the TS600 line (Supplementary Figure S4A,B). These results suggest that MEOX2 potentially regulates ERK phosphorylation in GBM. However, the regulation may depend on baseline ERK activity, because MEOX2 had no effect on p-ERK levels in the EGFR-amplified TS600 line with higher ERK phosphorylation at steady state.

MEOX2 Transcriptional Activity is Regulated by Phosphorylation

As the activity of transcription factors can be modulated by phosphorylation, we wondered whether MEOX2 harbored phosphorylation sites. We identified ERK-dependent phosphorylation sites upstream of the homeobox domain of MEOX2 in GBM tumorspheres by mass spectrometry following immunoprecipitation of endogenous or overexpressed MEOX2 (Figure 2A). We identified Ser155 which harbors a consensus MAPK motif (S/T)P and has been reported in a large-scale phospho-proteomics analysis of breast cancer.29 Therefore, we considered whether this may be a regulatory phosphorylation site in MEOX2. To determine whether Ser155 could lead to altered MEOX2 activity, the serine residue was changed to either alanine (S155A) or the phosphomimetic aspartate (S155D). Compared to the empty vector, transient overexpression of MEOX2 and MEOX2S155A increased ERK1/2 phosphorylation; however, the increase with MEOX2S155A was less pronounced (Figure 2B,C). Importantly, p21, a known transcriptional target of MEOX2, was only slightly induced upon MEOX2S155A overexpression, such that p21 levels were comparable to p21 levels of MEK inhibitor trametinib treated MEOX2-overexpressing cells. (Supplementary Figure S4C,D). In the S155D-expressing cells, the levels of p-ERK, p21, and MEOX2 were comparable to MEOX2 overexpression (Figure 2E–G; Supplementary Figure S4D). We did not observe an effect of S155A substitution on protein stability using cycloheximide chase assay (Supplementary Figure S5A–C). Nevertheless, MEOX2 expression was significantly decreased in this mutant, which could explain the decrease in ERK phosphorylation (Figure 2D).

Fig. 2.

Fig. 2

Phosphorylation of S155 affects the transcriptional activity of MEOX2. (A) The amino acid sequence of MEOX2. Red boxed regions include the phosphorylated S155 residue. S-P motifs are underlined and marked with an asterisk (*). The homeobox domain is highlighted in yellow. (B) Western blot for MEOX2, phospho-ERK (p-ERK1/2), and p21 in HEK293TN cells transiently transfected with empty vector, MEOX2 or MEOX2S155A for 48 h and treated with 100 nM trametinib (+) or control DMSO (-) for 24 h. (C) Quantification of (B) showing p-ERK relative to total ERK levels. EV, empty vector; WT, MEOX2; S155A, MEOX2S155A. Mean ± SD, n = 4, one-way ANOVA with Dunnett’s correction. (D) Quantification of (B) showing FLAG (detecting MEOX2-FLAG) relative to β-actin in trametinib treated and control cells. Mean ± SD, n = 4, two-way ANOVA with Tukey’s multiple comparisons test. (E) Western blot showing FLAG, p-ERK1/2, ERK1/2 and p21 in HEK293TN cells transiently transfected with empty vector, MEOX2, MEOX2S155A or MEOX2S155D. (F) Quantification of (E) showing p-ERK relative to total ERK levels. (G) Quantification of (E) showing FLAG (detecting MEOX2-FLAG, MEOX2S155A-FLAG or, MEOX2S155D-FLAG) normalized to β-actin. Mean ± SD, n = 3, one-way ANOVA with Dunnett’s correction. (H) Localization of MEOX2 and MEOX2S155A in HEK293TN cells 48 h after transient transfection. Immunofluorescence using MEOX2 antibody (green) and DAPI (blue). Insets are zoomed images. Scale bar = 25 μm. (I) Quantification of MEOX2 localization in HEK293TN cells 48 h after transient transfection with MEOX2 (n = 30 cells) versus MEOX2S155 (n = 36 cells). Localized signal in nuclear membrane (rim) divided by nuclear signal (nucleus) is shown. Mean ± SD, two-sided Welsh’s test (See online version for color figure).

We next investigated whether the nuclear distribution of the mutant MEOX2 proteins may be altered. Using maximum intensity projection of z stack images, we analyzed the fluorescence signal distribution around the nuclear periphery compared to the whole nucleus (Figure 2H, Supplementary Figure S5D). Indeed, MEOX2S155A showed altered nuclear localization with a significantly increased distribution in the nuclear envelope and nuclear lamina area compared with the whole nucleus (Figure 2H,I).

We then aimed to observe the localization kinetics of MEOX2WT, MEOX2S155A, and MEOX2S155D constructs by immunofluorescence staining 24, 48, and 72 hours after transfection (Supplementary Figure S5E). After 24 hours, MEOX2WT localizes homogenously in the nucleus without a specific accumulation at the nuclear rim. However, after 24 and 48 hours, MEOX2S155A localization was not as specific as MEOX2WT. After 72 hours, localization of all three constructs was comparable (Supplementary Figure S5E). Nuclear periphery has been associated with gene silencing in several previous studies30–32 and nonactivated transcription factors have been shown to be sequestered in the inner nuclear envelope.33 These results indicate that S155 may play a role in subnuclear localization kinetics of MEOX2.

MEOX2 can Promote a Growth Phenotype

To determine whether MEOX2 expression regulates tumor growth in vivo, we identified two GBM tumorsphere lines (L0125 and L0512) with low endogenous MEOX2 levels (Supplementary Figure S6A). These lines were engineered to overexpress MEOX2 and labeled with luciferase for in vivo bioluminescence (BLI) tracking. BLI measurements showed that MEOX2 overexpression in L0125 induced a significantly faster growth phenotype than in mice implanted with empty vector expressing cells, whereas no difference in growth kinetics was observed in the MEOX2 overexpressing L0512 line (Figure 3A,B; Supplementary Figure S6B–D). In addition, we implanted TS600 and TS667 cells with MEOX2 KD; however due to the high proliferative capacity of these cells in vivo (data not shown), we did not observe a difference in tumor growth between MEOX2 KD and control cells.

Fig. 3.

Fig. 3

MEOX2 increases in vivo tumor growth in a context-dependent manner. (A) Serial bioluminescence (BLI) imaging of L0125 transduced with control or MEOX2 (MEOX2OE) (7 mice in each group). BLI of each individual mouse is plotted. Mean ± SEM, two-sided Welsh’s test. (B) Example images of BLI in (A) are shown for mice implanted with MEOX2OE (top) or control (bottom) cells. (C) Heatmap of differentially expressed genes (DEGs) in immortalized human astrocytes transiently transfected with MEOX2 compared to empty vector controls. (D) Enriched pathways in DEGs shown in (C). Colors of the bars indicate predicted activation states of the pathways identified by the Ingenuity Pathway Analysis (IPA). Green, activated; orange, inhibited; grey, unknown; B-H, Benjamini-Hochberg (See online version for color figure).

The heterogeneous in vivo growth response of GBM lines to MEOX2 overexpression could be due to the faster basal growth kinetics of L0512, TS600, and TS667 in vivo, suggesting that MEOX2 may increase tumor growth in GBM lines with slow basal in vivo growth dynamics such as L0125. Although we did not observe an effect of MEOX2 knockdown on tumor growth in the TS600 and TS667 KD lines, further studies are needed to clearly elucidate whether MEOX2 loss may suppress tumor growth and improve overall survival in preclinical models. In addition, the use of patient-derived tumorspheres from different GBM subtypes to study the in vivo phenotype associated with MEOX2 will address whether MEOX2 is involved in tumor growth in a specific GBM context.

Based on our in vivo data, we wanted to test whether MEOX2 may be involved in tumor initiation and if this role may be mediated via target gene expression that could be usurped by constitutive RTK signaling in GBM. To determine which genes may be deregulated by MEOX2 in nontransformed cells, we transiently overexpressed MEOX2 in immortalized human astrocytes (IHA) and performed RNA-seq. We identified 136 upregulated and 12 downregulated genes (Figure 3C). Pathway analysis showed enrichment of several networks, including STAT activation and HIF1α signaling (Figure 3D). Upregulated genes included NGFR and members of NGF-stimulated transcription (VGF, EGR3, ARC) and hypoxia pathways (EFNA1, CDKN1C, PDGFB, PKP1) (Figure 3C; Supplementary Table S4). NGFR is highly expressed in GBM and inhibits the transcriptional activity of p53 to exert its oncogenic function.34AQP1 and AQP3, members of the aquaporin (AQP) family, are upregulated upon MEOX2 overexpression. AQPs have been implicated in tumor cell growth and migration.35

Next, we aimed to determine whether MEOX2 might play a role in glioma initiation. To mimic tumor initiation in vitro, we used a cerebral organoid model to study whether overexpression of MEOX2 synergizes with loss of canonical GBM tumor suppressors (PTEN, p53, and NF1) Organoids were electroporated with CRISPR-Cas9 to induce loss of indicated tumor suppressors together with stable overexpression of mNeonGreen with or without MEOX2 co-expression using a PiggyBac transposon system (Figure 4A; Supplementary Figure S7A–C). This allows to monitor clonal outgrowth of genetically engineered fluorescently labeled cells in the context of a normal cerebral forebrain organoid. Remarkably, we found that MEOX2 synergized with p53 and PTEN loss to significantly increase clonal expansion of affected cells to the same level as the PTEN, p53, and NF1 triple knockout positive control (Figure 4A,B; Supplementary Figure S7C,D). Compared to scrambled control, MEOX2 overexpression (adjusted P-value = .23), as well as p53 and PTEN loss alone (adjusted P-value = .08), did not lead to a significant increase in growth (Figure 4A,B).

Fig. 4.

Fig. 4

MEOX2 leads to increased proliferation in human cerebral organoid models of GBM. (A) Images of cerebral organoids following nucleofection with scrambled control, NF1-/-PTEN-/-p53-/-, PTEN-/-p53-/-, MEOX2OEPTEN-/-p53-/-or MEOX2OE plasmids. Black color indicates mNeonGreen positive cells. Growth measured after 2 days (top row) and one month after nucleofection (bottom row). OE, overexpression. (B) Quantification of mNeonGreen fluorescence intensity in cerebral organoids shown in (A). Control (n = 32), NF1-/-PTEN-/-p53-/- (n = 26), PTEN-/-p53-/- (n = 25), MEOX2OE (n = 27), MEOX2OEPTEN-/-p53-/- (n = 30). Only significant comparisons are shown (*P-value < .05). One-way ANOVA, Tukey’s multiple comparisons test. (C) PCA plot showing the separation of the various organoid models (n = 2 per group). (D) Pathway analysis showing the commonly altered pathways among the upregulated genes in the organoid models compared to the scrambled control (See online version for color figure).

To determine transcriptional alterations upon MEOX2 overexpression, we performed RNA-seq from the organoid models (Supplementary Table S5). Principal component analysis (PCA) indicated that MEOX2 was sufficient to shift the transcriptomic profile distinct from control organoids (Figure 4C). Among the differentially upregulated genes, several pathways including ECM organization, collagen formation, protein phosphorylation were enriched among the upregulated genes (Figure 4D; Supplementary Figure S8). Several pathways involving neurotransmitter receptors and RTK protein phosphatases were downregulated (Supplementary Figure S9). Pathway enrichment results suggest that MEOX2 perturbs similar oncogenic pathways when compared to tumorigenic models with loss of PTEN/p53 or NF1/PTEN/p53. In addition, as MEOX2 cooperated with p53 and PTEN loss to drive proliferation, we wondered which genes may be expressed at higher levels in PTEN-/-p53-/-MEOX2OE organoids. When we overlapped the up- or downregulated genes (Supplementary Figure S10A), 119 upregulated genes were unique to PTEN-/-p53-/-MEOX2OE organoids and included several glioma-associated genes such as DLK1, SLC2A3, EDN1, NOTCH1 (Supplementary Figure S10B). These results suggest that MEOX2 in conjunction with additional mutations could act as an early driver of malignant transformation in gliomas.

MEOX2 Alters Molecular Pathways Involved in Tumorigenesis

To better understand the oncogenic role of MEOX2 at the molecular level, we performed transcriptomic analysis of two MEOX2 KD lines (TS600 and TS667) and two MEOX2 overexpression lines (L0125 and L0512) by RNA-seq. First, we analyzed the molecular changes in the MEOX2 KD lines compared to nontargeting controls (NTC) and identified the differentially expressed genes (Figure 5A; Supplementary Table S6). Notably, several genes, including ALK, EGFR, NOTCH3, and NRCAM were differentially expressed in both TS667 and TS600 lines (Figure 5B,C). EGFR, a hyperactivated GBM oncogene, and the associated signaling axis activate transcription factor networks that relay oncogenic signals.36 In addition, Notch signaling is deregulated in malignant gliomas,37 and NOTCH3 drives cell motility and mesenchymal gene programs in neuroblastoma.38 Furthermore, VGF is transcriptionally induced by MEOX2, as shown by RNA-seq data from IHA and TS667. VGF has been shown to trigger EMT and confer resistance to tyrosine kinase inhibitors.39

Fig. 5.

Fig. 5

MEOX2 mediates activation of oncogenic pathways. (A) Venn diagrams showing overlap between up- or down-regulated differentially expressed genes (DEGs) in KD1 and KD2 TS667 (top) or TS600 (bottom). KD, knockdown. (B) Normalized counts of ALK, EGFR, NOTCH3 and NRCAM in TS667 cells. Dark gray dots indicate differential expression (adjusted p-value < 0.05) compared to NTC. (C) Normalized counts of ALK, EGFR, NOTCH3 and NRCAM in TS600 cells. Dark gray dots indicate differential expression compared to NTC. NTC, non-targeting control. (D) Significantly enriched pathways in TS667 KD1, TS600 KD2 and TS667 KD2. (E) Venn diagram showing overlap between up-regulated (left) and down-regulated (right) DEGs in L0125 and L0512 MEOX2 overexpressing compared to empty vector expressing control lines. (F) Pathway enrichment among up- or down-regulated DEGs in L0125 and L0512 MEOX2 overexpressing compared to empty vector expressing control lines.

Several pathways and upstream regulators were coherently altered upon MEOX2 loss. These included pathways such as hepatic fibrosis signaling, HOTAIR regulatory pathway, ILK, and GP6 signaling. The upstream regulators such as TGFB1 and AGT, were predicted to be less active (Supplementary Table S7). Several pathways including extracellular matrix (ECM) organization and collagen biosynthesis were significantly enriched among downregulated genes upon MEOX2 loss (Figure 5D).

We next identified the differentially expressed genes in MEOX2 overexpressing L0125 and L0512 GBM tumorsphere lines (Figure 5E; Supplementary Table S8). The upregulated genes were enriched for several pathways, including extracellular matrix organization, integrin signaling pathway, and EMT (Figure 5F). Overall, our results indicate that similar pathways (eg, ECM remodeling, collagen biosynthesis) are altered in the tumorsphere and organoid models upon MEOX2 modulation, Notably, several genes including AQP1, AQP3, NGFR are upregulated in MEOX2 overexpressing IHA and organoid models, and EGFR is upregulated in both tumorsphere lines. Taken together, these results suggest that MEOX2 may modulate ECM and collagens as part of its malignant gene expression program.

Identification of Direct MEOX2 Target Genes

To determine the genome-wide binding patterns of MEOX2, we used antibody-guided chromatin tagmentation sequencing (ACT-seq) with KD2 as the control and performed cleavage under targets and tagmentation sequencing (CUT&Tag) in a primary fresh frozen GBM tumor. Heatmaps of the correlations between ACT-seq biological replicates per cell line are shown in Supplementary Figure S11. In total, 1519 peaks in TS667 and 1855 peaks in TS600 lines were identified compared to their respective KD2 and IgG controls (Figure 6A; Supplementary Figure S12A; Table S9). Compared to IgG control, 2188 peaks were identified in the primary GBM (Figure 6B; Supplementary Table S9). MEOX2 showed a distal binding occupancy, with most peaks located within intergenic and intronic regions (Figure 6C). This suggests that MEOX2 may bind to enhancer regions which can be located at distal sites from the transcription start sites.40 Peaks from both tumorspheres and the GBM tumor harbored significant motif enrichment for putative MEOX1/2 sites (Figure 6D; Supplementary Table S10). MEOX2 motif was highly ranked among the MEOX2 associated peaks in the TS667 and GBM compared to the TS600 line. Given our results that, in the TS600 line, p-ERK levels are unchanged upon MEOX2 loss and the overall number of differentially expressed genes are less compared to TS667 upon MEOX2 knockdown indicates that MEOX2 binding to its target genes may be weaker in TS600.

Fig. 6.

Fig. 6

Identification and characterization of MEOX2 bound genomic regions.(A) Heatmap of normalized reads of genomic regions differentially bound by MEOX2 in TS600. MEOX2 peaks are ranked by intensity and shown relative to IgG negative controls in TS600 NTC and TS600 KD2. Each experimental group includes 3 biological replicates. (B) Profiles of normalized reads of MEOX2-bound peaks in MEOX2 (dark gray) and IgG (light gray) CUT&Tag in a fresh frozen GBM sample. (C) Bar plot indicating the genomic feature distributions of the MEOX2-bound peaks in TS667, TS600, and GBM samples. (D) Predicted sequence motifs and target genes for MEOX2-bound DNA sequences in TS600. PWMs, position weight matrices. (E) Pathway enrichment of MEOX2 peaks in the primary GBM tumor. (F) Venn diagram showing overlap of annotated genes identified from the differentially bound peaks in TS600, TS667 and primary GBM tumor. (G) Gene track view of normalized bigwig reads at the promoters of FABP7 (left) and LARS2-AS1 (right) of TS600, TS667, and GBM samples. The black lines indicate the called peaks corresponding to each sample.

To determine the biological functions of MEOX2 bound regions, we subjected the peaks to pathway enrichment analysis. This analysis identified several significantly enriched pathways, including PI3K/AKT signaling, MAPK signaling, cell-cell communication, and diseases of signal transduction (Figure 6E; Supplementary Figure S12B, C). A total of 138 genes were identified as bound by MEOX2 across all samples, including the ETS family (ETV1, ETV5, ETS1) and MAPK attenuators SPRY2 and DUSP10 (Figure 6F). These results suggest that MEOX2 may potentially regulate the MAPK/ERK signaling by corralling and fine-tuning ERK1/2 activity to prevent hyperactive ERK-associated toxicities. This might enable tumor initiation and growth in GBMs driven by RTK alterations. In addition, we overlapped the peaks in the tumorspheres and GBM tumor, which identified 44 overlapping peak-associated regions in all three samples (Supplementary Table S11). Among the overlapping genomic regions were several oncogenes such as FABP7 and long noncoding RNAs such as BHLHE40-AS1, LARS2-AS1, and DLGAP1-AS1 (Figure 6G). MEOX2 recruitment to the promoters of FABP7 and LARS2-AS1 was validated by PCR (Supplementary Figure S12D). Intriguingly, FABP7 is downregulated upon MEOX2 loss, suggesting MEOX2 as a transcriptional activator of FABP7 (Supplementary Table S6). Nuclear FABP7 is associated with infiltrative gliomas and poor prognosis in EGFR-overexpressing GBM.41 Notably several genes including EGFR, ALK, NRCAM are bound and activated by MEOX2 (Supplementary Figure S13).

Taken together, we identified MEOX2 as an oncogenic mesodermal transcription factor on chromosome 7 that activates several oncogenic pathways such as MAPK signaling that is central to GBM biology. We demonstrate that MEOX2 overexpression cooperates with loss of tumor suppressors (eg, PTEN and p53) to drive significant growth in human cerebral organoid models of GBM, highlighting its role in tumor initiation. Given that MEOX2 is undetectable in the normal brain but co-opted and upregulated by GBM underscores the complexity of this disease. Although aberrant TFs play a major role in cancers,42 from a therapeutic perspective, there are still major challenges associated with targeting TFs. Over the last years, several promising strategies have emerged including targeted protein degradation approaches and inhibition of transcriptional machinery that cooperates with TFs.43 As we advance our understanding of cancer-associated TFs and the transcriptional networks hijacked by corresponding tumors, we are likely to identify their associated therapeutically relevant vulnerabilities.

Accession Numbers

All data have been deposited in NCBI’s Gene Expression Omnibus under accession number GSE181146.

Supplementary Material

noac110_suppl_Supplementary_Table_S1
noac110_suppl_Supplementary_Table_S2
noac110_suppl_Supplementary_Table_S3
noac110_suppl_Supplementary_Table_S4
noac110_suppl_Supplementary_Table_S5
noac110_suppl_Supplementary_Table_S6
noac110_suppl_Supplementary_Table_S7
noac110_suppl_Supplementary_Table_S8
noac110_suppl_Supplementary_Table_S9
noac110_suppl_Supplementary_Table_S10
noac110_suppl_Supplementary_Table_S11
noac110_suppl_Supplementary_Figures
noac110_suppl_Supplementary_Data

Acknowledgments

We thank the members of the Turcan lab for helpful discussions. We thank the Genomics and Proteomics Core Facility (GPCF) at the DKFZ for providing next-generation sequencing (NGS) services and proteomics services and analysis. We thank the Omics IT and Data Management Core Facility (ODCF) at the DKFZ for data management and technical support. We thank the DKFZ Single-cell Open Lab (scOpenLab) for the support and experimental assistance.

Conflict of interest statement. The authors declare no conflicts of interest.

Contributor Information

Anna Schönrock, Clinical Cooperation Unit Neurooncology, German Consortium for Translational Cancer Research (DKTK), German Cancer Research Center (DKFZ), Heidelberg, Germany; Neurology Clinic and National Center for Tumor Diseases, University Hospital Heidelberg , Heidelberg, Germany.

Elisa Heinzelmann, Cell Fate Engineering and Disease Modeling Group, German Cancer Research Center (DKFZ) and DKFZ-ZMBH Alliance, Heidelberg, Germany; HITBR Hector Institute for Translational Brain Research gGmbH, Heidelberg, Germany; Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany.

Bianca Steffl, Clinical Cooperation Unit Neurooncology, German Consortium for Translational Cancer Research (DKTK), German Cancer Research Center (DKFZ), Heidelberg, Germany.

Engin Demirdizen, Neurology Clinic and National Center for Tumor Diseases, University Hospital Heidelberg , Heidelberg, Germany.

Ashwin Narayanan, Neurology Clinic and National Center for Tumor Diseases, University Hospital Heidelberg , Heidelberg, Germany.

Damir Krunic, Core Facility Unit Light Microscopy, German Cancer Research Center (DKFZ), Heidelberg, Germany.

Marion Bähr, Division of Cancer Epigenomics, German Cancer Research Center (DKFZ), Heidelberg, Germany.

Jong-Whi Park, Neurology Clinic and National Center for Tumor Diseases, University Hospital Heidelberg , Heidelberg, Germany.

Claudia Schmidt, Core Facility Unit Light Microscopy, German Cancer Research Center (DKFZ), Heidelberg, Germany.

Koray Özduman, Department of Neurosurgery, School of Medicine, Acibadem Mehmet Ali Aydinlar University, Istanbul, Turkey.

M Necmettin Pamir, Department of Neurosurgery, School of Medicine, Acibadem Mehmet Ali Aydinlar University, Istanbul, Turkey.

Wolfgang Wick, Clinical Cooperation Unit Neurooncology, German Consortium for Translational Cancer Research (DKTK), German Cancer Research Center (DKFZ), Heidelberg, Germany; Neurology Clinic and National Center for Tumor Diseases, University Hospital Heidelberg , Heidelberg, Germany.

Felix Bestvater, Core Facility Unit Light Microscopy, German Cancer Research Center (DKFZ), Heidelberg, Germany.

Dieter Weichenhan, Division of Cancer Epigenomics, German Cancer Research Center (DKFZ), Heidelberg, Germany.

Christoph Plass, Division of Cancer Epigenomics, German Cancer Research Center (DKFZ), Heidelberg, Germany.

Julian Taranda, Neurology Clinic and National Center for Tumor Diseases, University Hospital Heidelberg , Heidelberg, Germany.

Moritz Mall, Cell Fate Engineering and Disease Modeling Group, German Cancer Research Center (DKFZ) and DKFZ-ZMBH Alliance, Heidelberg, Germany; HITBR Hector Institute for Translational Brain Research gGmbH, Heidelberg, Germany; Central Institute of Mental Health, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany.

Şevin Turcan, Neurology Clinic and National Center for Tumor Diseases, University Hospital Heidelberg , Heidelberg, Germany.

Funding

This work was supported by the German Cancer Aid, Max Eder Program grant number 70111964 (S.T.). J.T. is supported by a Deutsche Forschungsgemeinschaft (DFG) Mercator Fellowship (DFG grant number TU 585/1-1). M.M and E.H. are supported by the Hector Stiftung II gGmbH.

Authorship Statement.

A.S., E.H., and S.T. designed and directed the study. A.S., E.H., B.S., E.D., A.N. performed the experiments. A.S., E.H., E.D., B.S., D.W., M.M, S.T. analyzed the data. A.S., E.H., E.D., D.W., F.B, J.T., M.M., S.T. interpreted the data. J-W.P., M.B., C.S. provided technical assistance. D.K. provided a custom-built macro plugin for image analysis. E.H. performed the organoid experiments. M.M. supervised the organoid experiments. K.O. and M.N.P. provided the patient sample. M.B, D.W., C.P. provided expertise and supervised the ACT-seq experiments. F.B, W.W., C.P., J.T. and M.M. provided conceptual advice. A.S. and S.T. wrote the paper. All authors contributed to the writing and/or editing of the manuscript.

References

  • 1. Louis DN, Perry A, Wesseling P, et al. . The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neuro Oncol 2021;23(8):1231–1251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Brennan CW, Verhaak RG, McKenna A, et al. . The somatic genomic landscape of glioblastoma. Cell 2013;155(2):462–477. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Cancer Genome Atlas Research Network. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 2008;455(7216):1061–1068. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Suva ML, Rheinbay E, Gillespie SM, et al. . Reconstructing and reprogramming the tumor-propagating potential of glioblastoma stem-like cells. Cell 2014;157(3):580–594. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Singh DK, Kollipara RK, Vemireddy V, et al. . Oncogenes activate an autonomous transcriptional regulatory circuit that drives glioblastoma. Cell Rep 2017;18(4):961–976. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Turcan S, Makarov V, Taranda J, et al. . Mutant-IDH1-dependent chromatin state reprogramming, reversibility, and persistence. Nat Genet. 2018;50(1):62–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Tachon G, Masliantsev K, Rivet P, et al. . Prognostic significance of MEOX2 in gliomas. Mod Pathol. 2019;32(6):774–786. [DOI] [PubMed] [Google Scholar]
  • 8. Ozawa T, Riester M, Cheng YK, et al. . Most human non-GCIMP glioblastoma subtypes evolve from a common proneural-like precursor glioma. Cancer Cell 2014;26(2):288–300. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Cimino PJ, Kim Y, Wu HJ, et al. . Increased HOXA5 expression provides a selective advantage for gain of whole chromosome 7 in IDH wild-type glioblastoma. Genes Dev. 2018;32(7-8):512–523. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Murat A, Migliavacca E, Gorlia T, et al. . Stem cell-related “self-renewal” signature and high epidermal growth factor receptor expression associated with resistance to concomitant chemoradiotherapy in glioblastoma. J Clin Oncol. 2008;26(18):3015–3024. [DOI] [PubMed] [Google Scholar]
  • 11. Costa BM, Smith JS, Chen Y, et al. . Reversing HOXA9 oncogene activation by PI3K inhibition: epigenetic mechanism and prognostic significance in human glioblastoma. Cancer Res. 2010;70(2):453–462. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Gallo M, Ho J, Coutinho FJ, et al. . A tumorigenic MLL-homeobox network in human glioblastoma stem cells. Cancer Res. 2013;73(1):417–427. [DOI] [PubMed] [Google Scholar]
  • 13. Mankoo BS, Collins NS, Ashby P, et al. . Mox2 is a component of the genetic hierarchy controlling limb muscle development. Nature 1999;400(6739):69–73. [DOI] [PubMed] [Google Scholar]
  • 14. LePage DF, Altomare DA, Testa JR, Walsh K. Molecular cloning and localization of the human GAX gene to 7p21. Genomics 1994;24(3):535–540. [DOI] [PubMed] [Google Scholar]
  • 15. Smith RC, Branellec D, Gorski DH, et al. . p21CIP1-mediated inhibition of cell proliferation by overexpression of the gax homeodomain gene. Genes Dev. 1997;11(13):1674–1689. [DOI] [PubMed] [Google Scholar]
  • 16. Douville JM, Cheung DY, Herbert KL, Moffatt T, Wigle JT. Mechanisms of MEOX1 and MEOX2 regulation of the cyclin dependent kinase inhibitors p21 and p16 in vascular endothelial cells. PLoS One. 2011;6(12):e29099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Ohshima J, Haruta M, Arai Y, et al. . Two candidate tumor suppressor genes, MEOX2 and SOSTDC1, identified in a 7p21 homozygous deletion region in a Wilms tumor. Genes Chromosomes Cancer 2009;48(12):1037–1050. [DOI] [PubMed] [Google Scholar]
  • 18. Cao G, Huang B, Liu Z, et al. . Intronic miR-301 feedback regulates its host gene, ska2, in A549 cells by targeting MEOX2 to affect ERK/CREB pathways. Biochem Biophys Res Commun. 2010;396(4):978–982. [DOI] [PubMed] [Google Scholar]
  • 19. Peralta-Arrieta I, Trejo-Villegas OA, Armas-Lopez L, et al. . Failure to EGFR-TKI-based therapy and tumoural progression are promoted by MEOX2/GLI1-mediated epigenetic regulation of EGFR in the human lung cancer. Eur J Cancer. 2022;160:189–205. [DOI] [PubMed] [Google Scholar]
  • 20. Tian L, Tao ZZ, Ye HP, et al. . Over-expression of MEOX2 promotes apoptosis through inhibiting the PI3K/Akt pathway in laryngeal cancer cells. Neoplasma 2018;65(5):745–752. [DOI] [PubMed] [Google Scholar]
  • 21. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Carter B, Ku WL, Kang JY, et al. . Mapping histone modifications in low cell number and single cells using antibody-guided chromatin tagmentation (ACT-seq). Nat Commun. 2019;10(1):3747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Liu CS, Toth R, Bakr A, et al. . Epigenetic modulation of radiation-induced diacylglycerol kinase alpha expression prevents pro-fibrotic fibroblast response. Cancers (Basel) 2021;13(10):2455. doi: 10.3390/cancers13102455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Narayanan A, Blanco-Carmona E, Demirdizen E, et al. . Nuclei isolation from fresh frozen brain tumors for single-nucleus RNA-seq and ATAC-seq. J Vis Exp 2020;(162):e61542. doi: 10.3791/61542. [DOI] [PubMed] [Google Scholar]
  • 25. Neftel C, Laffy J, Filbin MG, et al. . An integrative model of cellular states, plasticity, and genetics for glioblastoma. Cell 2019;178(4):835–849.e21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Tao BB, He H, Shi XH, et al. . Up-regulation of USP2a and FASN in gliomas correlates strongly with glioma grade. J Clin Neurosci. 2013;20(5):717–720. [DOI] [PubMed] [Google Scholar]
  • 27. Srivastava NK, Pradhan S, Gowda GA, Kumar R. In vitro, high-resolution 1H and 31P NMR based analysis of the lipid components in the tissue, serum, and CSF of the patients with primary brain tumors: one possible diagnostic view. NMR Biomed. 2010;23(2):113–122. [DOI] [PubMed] [Google Scholar]
  • 28. Gimple RC, Kidwell RL, Kim LJY, et al. . Glioma stem cell-specific superenhancer promotes polyunsaturated fatty-acid synthesis to support EGFR signaling. Cancer Discov 2019;9(9):1248–1267. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Mertins P, Mani DR, Ruggles KV, et al. . Proteogenomics connects somatic mutations to signalling in breast cancer. Nature 2016;534(7605):55–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Lochs SJA, Kefalopoulou S, Kind J. Lamina associated domains and gene regulation in development and cancer. Cells 2019;8(3):271. doi: 10.3390/cells8030271. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Shaklai S, Amariglio N, Rechavi G, Simon AJ. Gene silencing at the nuclear periphery. FEBS J. 2007;274(6):1383–1392. [DOI] [PubMed] [Google Scholar]
  • 32. Ulianov SV, Doronin SA, Khrameeva EE, et al. . Nuclear lamina integrity is required for proper spatial organization of chromatin in Drosophila. Nat Commun. 2019;10(1):1176. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Heessen S, Fornerod M. The inner nuclear envelope as a transcription factor resting place. EMBO Rep. 2007;8(10):914–919. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Zhou X, Hao Q, Liao P, et al. . Nerve growth factor receptor negates the tumor suppressor p53 as a feedback regulator. Elife 2016;5:e15099. doi: 10.7554/eLife.15099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Verkman AS, Hara-Chikuma M, Papadopoulos MC. Aquaporins—new players in cancer biology. J Mol Med (Berl). 2008;86(5):523–529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. An Z, Aksoy O, Zheng T, Fan QW, Weiss WA. Epidermal growth factor receptor and EGFRvIII in glioblastoma: signaling pathways and targeted therapies. Oncogene 2018;37(12):1561–1575. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Bazzoni R, Bentivegna A. Role of notch signaling pathway in glioblastoma pathogenesis. Cancers (Basel) 2019;11(3):292. doi: 10.3390/cancers11030292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. van Nes J, Chan A, van Groningen T, et al. . A NOTCH3 transcriptional module induces cell motility in neuroblastoma. Clin Cancer Res. 2013;19(13):3485–3494. [DOI] [PubMed] [Google Scholar]
  • 39. Hwang W, Chiu YF, Kuo MH, et al. . Expression of neuroendocrine factor VGF in lung cancer cells confers resistance to EGFR kinase inhibitors and triggers epithelial-to-mesenchymal transition. Cancer Res. 2017;77(11):3013–3026. [DOI] [PubMed] [Google Scholar]
  • 40. Spitz F, Furlong EE. Transcription factors: from enhancer binding to developmental control. Nat Rev Genet. 2012;13(9):613–626. [DOI] [PubMed] [Google Scholar]
  • 41. Liang Y, Bollen AW, Aldape KD, Gupta N. Nuclear FABP7 immunoreactivity is preferentially expressed in infiltrative glioma and is associated with poor prognosis in EGFR-overexpressing glioblastoma. BMC Cancer 2006;6:97. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Bushweller JH. Targeting transcription factors in cancer - from undruggable to reality. Nat Rev Cancer. 2019;19(11):611–624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Henley MJ, Koehler AN. Advances in targeting “undruggable” transcription factors with small molecules. Nat Rev Drug Discov. 2021;20(9):669–688. [DOI] [PubMed] [Google Scholar]
  • 44. Bowman RL, Wang Q, Carro A, Verhaak RG, Squatrito M. GlioVis data portal for visualization and analysis of brain tumor expression datasets. Neuro Oncol 2017;19(1):139–141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Louis DN, Ohgaki H, Wiestler OD, et al. . The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol. 2007;114(2):97–109. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

noac110_suppl_Supplementary_Table_S1
noac110_suppl_Supplementary_Table_S2
noac110_suppl_Supplementary_Table_S3
noac110_suppl_Supplementary_Table_S4
noac110_suppl_Supplementary_Table_S5
noac110_suppl_Supplementary_Table_S6
noac110_suppl_Supplementary_Table_S7
noac110_suppl_Supplementary_Table_S8
noac110_suppl_Supplementary_Table_S9
noac110_suppl_Supplementary_Table_S10
noac110_suppl_Supplementary_Table_S11
noac110_suppl_Supplementary_Figures
noac110_suppl_Supplementary_Data

Articles from Neuro-Oncology are provided here courtesy of Society for Neuro-Oncology and Oxford University Press

RESOURCES