Figure 4: DAXX inhibition affects H3.3 genomic enrichment, gene expression and oncogenesis in PTEN-null/GBM cells.

(a) Genomic distribution of H3.3 peaks in DAXX knockdown GBM cells. (b) Volcano plot of H3.3 differential binding (DiffBind) peaks between shControl (shLuc) and shDaxx PTEN-null/GBM samples by ChIP-seq (P<0.05&log2(Fold Change)>1). (c) Gene ontology analysis of H3.3 differential binding genes in DAXX knockdown PTEN-deficient GBM-spheres. (d) Volcano plot of differentially expressed (DiffExp) gene between shLuc and shDaxx PTEN-null/GBM samples by RNA-seq (P<0.05&log2(Fold Change)>1). (e) Heatmap of Top 100 DiffExp genes highlighting tumour suppressors genes and oncogenes according to the Cancer Gene Census in DAXX-kd compared with shControl (shLuc) cells (P-value<0.05, log2(Fold change)>1). (f) Scatter plot of overlay genes between H3.3 DiffBind and DiffExp genes from ChIP- and RNA-seq, respectively. Red square, upregulated genes with H3.3 enrichment. Green square, downregulated genes with H3.3 enrichment (P<0.05&log2(Fold Change)>1). Top 100 DiffExp genes are labelled with green dots in ChIP- and RNA-seq plots and scatter plot. (g,h) Quantification of the total number of spheres (g) and sphere size (h) in GBM-PDX lines transduced with shLuc or shDaxx (n=3 biological samples with six replicates each, NS: no significant differences, *P<0.05, **P<0.001 and ***P<0.0001, one-way ANOVA). (i) Plots of sphere-forming frequencies using PTEN-NULL (GSC11 and GSC23) and PTEN-WT (GBM6 and TS576) GBM-PDX neurospheres, after stable knockdown of DAXX or shControl. The assay was performed by in vitro limiting dilution using a 0.95 confidence interval.