Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Nov 28;15(1):10363.
doi: 10.1038/s41467-024-54750-9.

Transcription factors ASCL1 and OLIG2 drive glioblastoma initiation and co-regulate tumor cell types and migration

Affiliations

Transcription factors ASCL1 and OLIG2 drive glioblastoma initiation and co-regulate tumor cell types and migration

Bianca L Myers et al. Nat Commun. .

Abstract

Glioblastomas (GBMs) are highly aggressive, infiltrative, and heterogeneous brain tumors driven by complex genetic alterations. The basic-helix-loop-helix (bHLH) transcription factors ASCL1 and OLIG2 are dynamically co-expressed in GBMs; however, their combinatorial roles in regulating the plasticity and heterogeneity of GBM cells are unclear. Here, we show that induction of somatic mutations in subventricular zone (SVZ) progenitor cells leads to the dysregulation of ASCL1 and OLIG2, which then function redundantly and are required for brain tumor formation in a mouse model of GBM. Subsequently, the binding of ASCL1 and OLIG2 to each other's loci and to downstream target genes then determines the cell types and degree of migration of tumor cells. Single-cell RNA sequencing (scRNA-seq) reveals that a high level of ASCL1 is key in specifying highly migratory neural stem cell (NSC)/astrocyte-like tumor cell types, which are marked by upregulation of ribosomal protein, oxidative phosphorylation, cancer metastasis, and therapeutic resistance genes.

PubMed Disclaimer

Conflict of interest statement

Competing interests: The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. ASCL1 and OLIG2 physically interact and overlap in binding in genome of PDOX-GBMs.
a Venn diagram of ChIP-seq binding peaks for ASCL1 or OLIG2 from two different PDOX-GBMs (R548, R738). b Heatmap of ASCL1 and OLIG2 shared binding sites. c DNA sequence motifs that are enriched within ASCL1&OLIG2 or OLIG2 alone binding peaks. Percentage represents the frequency of indicated DNA motif found within 150-bp peak summits compared to (percentage) frequency of that motif in background genomic sequence along with p-value. Motif enrichment is calculated using cumulative binomial distributions. d Overlap of genes associated with ASCL1 or OLIG2 binding peaks intersecting with genes positively or negatively correlated with ASCL1 and OLIG2 expression in RNA-seq of 164 GBM samples from the TCGA. e Gene ontology analysis of 841 positively correlated genes showing enrichment of molecular functions in the nucleus and cell cycle. f Gene ontology analysis of 211 negatively correlated genes showing enrichment of molecular functions in immune response and inflammation. Significance of enrichment was determined using Fisher’s Exact test. g ASCL1 and OLIG2 binding peaks at the other’s loci and known NOTCH (DLL3, NOTCH1, HES5), NSC (INSM1), and bHLH E-protein (TCF3, TCF4) targets. h Dynamic co-localization of ASCL1 and OLIG2 in PDOX-GBM. Percentage co-localization of ASCL1 and OLIG2 was previously reported. Scale bar: 25 μm. i Co-IP assay of PDOX-GBM using anti-ASCL1 antibody. Immunoblot (IB) showing presence of OLIG2 (~40 kDa, red arrowhead) in IP lane. Note that the 55 kDa band in ASCL1 IB panel is IgG. This result was observed in two independent experiments.
Fig. 2
Fig. 2. ASCL1 and OLIG2 are required for tumor formation but inversely regulate different aspects of tumor migration in GBM mouse model.
a Schematic showing induction of tdTOM+ brain tumors via electroporation of indicated Cre + CRISPR plasmids into neural progenitor cells in the right lateral ventricle of R26RT/T reporter mice at birth for longitudinal analyses. (bq, s, t) Representative images of tdTOM+ tumors at P30, P60, or terminal stage in control (be), Ascl1-CKO (fi), Olig2-CKO (jm), and double CKO (nq, s, t) mice (number of tumors imaged: n = 4/genotype for P30 & P60 and n = 6/genotype for terminal tumors). Arrows indicate midline and arrowheads mark the distance of migration of tdTOM+ tumor cells on the contralateral corpus callosum (CC). Asterisks demonstrate region imaged for (e, i, m, q, and t). ASCL1 and OLIG2 are highly co-expressed in control tdTOM+ tumor cells, but absent in the single or double CKO tdTOM+ tumors. Scale bars: 1 mm for whole brain sections; 25 μm for (e, i, m, q, t), and 12.5 μm for all insets. r Kaplan-Meier survival curves of each group of tumor mice showing statistical significance (Mantel-Cox test) between control versus experimental groups.
Fig. 3
Fig. 3. ASCL1 overexpression (OE) promotes tumor cell migration.
a Schematic induction of Ascl1-OE GFP+ tumor model. Cre-mediated recombination following electroporation results in sustained expression of tetracycline transactivator (tTA), which binds to TetO-promoter to drive expression of Ascl1 and GFP. be Representative images of P30, P60, and terminal-stage tumors highlighting extensive migration and co-expression of ASCL1 and OLIG2 in GFP+ tumor cells. Asterisk indicates region imaged for (e). Arrow indicates midline and arrowhead marks the distance of migration of GFP+ tumor cells on the contralateral CC. Scale bars: 1 mm for (bd); 25 μm for (e); and 12.5 μm for all insets. f, g Scatter plot of immunofluorescent intensity of ASCL1 or OLIG2 within individual DAPI+ nuclei of tumor cells of genotypes indicated (n represents the number of mice). Statistical significance is determined by comparing the mean immunofluorescent intensity of ~60 tumor cells/mouse/genotype using unpaired t-tests with Welch’s correction. h, i Quantification of the density of DAPI+ tumor cells in tumor bulk, and distance of migration of reporter+ tumor cells on contralateral CC normalized to the total length of the contralateral CC. Data shown as mean ± SEM. Statistical significance is determined by comparing the means of tumor types using unpaired t-tests with Welch’s correction. (h: control n = 5 mice, Ascl1-CKO n = 5 mice, Olig2-CKO n = 6 mice, Ascl1-OE n = 6 mice; i: n = 6 mice/genotype). j Co-IP assay of Ascl1-OE and Ascl1-CKO tumors using anti-ASCL1 antibody. Immunoblot (IB) showing presence of OLIG2 (~40 kDa) in IP lane of Ascl1-OE tumor but not in Ascl1-CKO tumor. Note that ASCL1 (~32 kDa) is detected in both Input and IP of Ascl1-OE tumor but not in Ascl1-CKO negative control tumor, demonstrating specificity of anti-ASCL1 antibody. This result was observed in two independent experiments. Source data are provided as a Source Data File.
Fig. 4
Fig. 4. ASCL1 and OLIG2 inversely regulate glioma tumor cell types.
Representative immunofluorescent images and H&E staining of terminal control (ad), Ascl1-CKO (eh), Olig2-CKO (il), dCKO (mp), and Ascl1-OE (qt) tumors. High magnification showing differences in co-localization, or lack thereof, of GFAP (white arrow) or SOX10 (white arrowhead) in reporter+ tumor cells. Dotted line delineates DAPI+ cell nuclei that overlap with tumor reporter and/or cell type markers. Similar staining was seen in n = 3/genotype. Black arrows and arrowheads in H&E panels indicate the presence of GFAP+ pink fibers and “fried-egg” cell shapes characteristics of astrocytoma and oligodendroglioma, respectively. Scale bars: 12.5 μm for (c, g, k, o, s) and 50 μm for all other panels.
Fig. 5
Fig. 5. ASCL1 overexpression promotes migration of newly transformed tumor cells.
aj Immunofluorescence of SOX10 and GFAP in P4 brains of control (ae) and Ascl1-OE (fj) tumor mice. tdTOM+ cells of control brain are mostly restricted to the SVZ but GFP+ cells of Ascl1-OE brain have migrated extensively into the striatum (Str), corpus callosum CC, and cortex (Ctx) (arrows in (a vs f)). Note that neither tdTOM nor GFP co-localize with SOX10 or GFAP (arrows in (ce) and (hj)), indicating that they are unspecified migrating NPC-like tumor cells. Results were observed for n = 3 mice/genotype. Scale bar: 500 μm for (a, f); 100 μm for (b, g); 50 μm for (ce, hj).
Fig. 6
Fig. 6. Single-cell RNA-seq reveals high degrees of transcriptional diversity for control and Ascl1-OE tumor cells.
a Workflow of FAC-sorted tdTOM+ or GFP+ tumor cells for scRNA-seq. b, c Number of cells per tumor and genes per cell sequenced for control (n = 3) and Ascl1-OE (n = 3) tumors. Box plot extends from the 25th to the 75th percentile of each sample’s distribution with the center line denoting the median number of genes per cell. Whiskers extend to 1.5 of the interquartile range. Dots are cells outside of the range. df UMAP visualization of unsupervised clustering of control and Ascl1-OE tumor cells yielded 16 different cell clusters. gl Gene expression confirming increased levels of Ascl1-ires-GFP, known target genes Dll3, Notch1, and Hes5, as well as binding partners Tcf4 and Tcf12 in Ascl1-OE tumor cells compared to controls. m, n GBM subtypes (PN proneural, MS mesenchymal, CL classical) and cell cycle phase analyses of control and Ascl1-OE tumor cells.
Fig. 7
Fig. 7. Ascl1 overexpression promotes NSC/astroglial-like cells and suppresses OPC/oligodendroglial-like cells in mouse GBMs.
af UMAPs demonstrating the proportion and distribution of 7 assigned cell types in control (ac) and Ascl1-OE (df) tumors based on cell type-specific gene signatures. g, h Differential gene expression confirms downregulation of oligodendrocyte lineage-specific genes (g) and upregulation of NSC/astrocyte-specific genes (h) in Ascl1-OE tumor cells. Bar graphs are mean Log2 Fold Change ± SEM for indicated genes by comparing cell type x cell type between control and Ascl1-OE tumors. Open circles within each bar graph represent number of cell types with indicated genes significantly altered (adjusted p-value < 0.05). Note that MOL genes are mostly restricted to MOL and thus only downregulated in that cell type. i Heatmap of averaged transcriptome for each assigned cell type into unionized RNA-seq triplicates (columns) for control and Ascl1-OE tumors showing specificity of signature genes (rows) to their respective cell types (black rectangles). Note that all cell types of control tumors expressed some level of OPC signature genes, which were the most drastically downregulated in Ascl1-OE tumors followed by NFOL and MOL signature genes (blue arrows), while NSC and astrocyte signature genes were highly upregulated across all cell types in Ascl1-OE tumors (red arrows).
Fig. 8
Fig. 8. ASCL1 and OLIG2 shared target cell type-specific genes are differentially expressed in control and Ascl1-OE tumors.
ac ChIP-seq tracks demonstrating shared binding of ASCL1 and OLIG2 at cis-regulatory sites of OPC/oligodendroglial (a), NSC/astroglial (b), and microglial lineage genes (c) and UMAP showing differential expression in control and Ascl1-OE cell clusters. All genes illustrated are shared targets of ASCL1 and OLIG2.
Fig. 9
Fig. 9. Genes highly upregulated in Ascl1-OE tumors are important for NSC maintenance, cancer metastasis and invasion, and therapeutic resistance.
a Differentially expressed genes (DEGs) significantly up- or downregulated by comparing cell type x cell type between control and Ascl1-OE tumors. DEGs include both ASCL1 target and non-target genes. b Gene ontology analysis of upregulated DEGs showing enrichment of genes important for protein synthesis and mitochondrial function. c, d Heatmap of Unionized Cell-Type RNA-seq (columns) demonstrating upregulation of ribosomal (c) and mitochondrial (d) genes (rows) in Ascl1-OE tumors. e UMAPs showing specific upregulation of cytochrome C oxidase subunit genes. f, g Top 25 genes upregulated (f) or downregulated (g) in ≥ 5 cell types, with delineation of ASCL1 targets and known functions in cancer or GBMs. Bar graphs are mean Log2 Fold Change ± SEM for indicated genes by comparing cell type x cell type between control and Ascl1-OE tumors. Open circles within each bar graph represent number of cell types with indicated genes significantly altered (adjusted p-value < 0.05). UMAP gene expression of 3 of the topmost upregulated (h) and downregulated (i) genes. Proportions of tumor cells expressing these genes are indicated for control and Ascl1-OE tumors.
Fig. 10
Fig. 10. Model of ASCL1 & OLIG2 Function in Gliogenesis and Gliomagenesis.
a Schematic of the role of ASCL1 and OLIG2 in glial cell fate specification in the dorsal forebrain. Developmentally, ASCL1 expression leads to generation of intermediate NPCs from radial glia and transcriptional activation of Olig2. Depending on sustained levels of OLIG2, NPCs are specified into glial precursor cells (APCs or OPCs). Markers of astrocyte and oligodendrocyte lineages are indicated. bf Schematic summary of tumor induction from radial glia and the role of ASCL1 and OLIG2 in specifying glioma cell types in control (b), Ascl1-CKO (c), Olig2-CKO (d), Ascl1-OE (e), and Ascl1;Olig2-dCKO (f) tumors. High levels of ASCL1 specifies NPC-like and AS-like tumor cells, which are highly migratory, whereas high levels of OLIG2 specifies OPC-like tumor cells. Tumor induction is mostly compromised in the absence of both ASCL1 and OLIG2. Created in BioRender. Myers, B. (2024) BioRender.com/p53n344.

Update of

References

    1. Ostrom, Q. T. et al. CBTRUS statistical report: primary brain and other central nervous system tumors diagnosed in the United States in 2015-2019. Neuro Oncol.24, v1–v95 (2022). - DOI - PMC - PubMed
    1. Bjorland, L. S. et al. Butterfly glioblastoma: clinical characteristics, treatment strategies and outcomes in a population-based cohort. Neurooncol. Adv.4, vdac102 (2022). - PMC - PubMed
    1. Youngblood, M. W., Stupp, R. & Sonabend, A. M. Role of resection in glioblastoma management. Neurosurg. Clin. N. Am.32, 9–22 (2021). - DOI - PubMed
    1. Bao, S. et al. Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature444, 756–760 (2006). - DOI - PubMed
    1. Beier, D., Schulz, J. B. & Beier, C. P. Chemoresistance of glioblastoma cancer stem cells-much more complex than expected. Mol. Cancer10, 128 (2011). - DOI - PMC - PubMed

Publication types

MeSH terms

Substances

Associated data