Epigenetic activity of phosphorylated PHF8 counteracts differentiation block via induction of cell-intrinsic immune response in AML. (A-C) Bar charts represent colony numbers in a serial replating assay and typical morphology of cells from fourth round of replating; primary mHSPCs were cotransduced with the indicated AML oncogenes or oncofusions as well as with wt PHF8 (PHF8), PHF8-F279S (F279S), or empty vector (control). Error bars indicate standard deviation (SD) of 3 independent experiments. Significance was tested using the Tukey multiple comparison test (2-way analysis of variance [ANOVA], ∗P < .04; ∗∗∗∗P < .000, nonsignificant [ns]). (D) Bar charts depict colony numbers of human AML cells transduced with PHF8 vs F279S vs Control. Error bars indicate SD of at least 4 independent experiments. Significance was tested using the Tukey multiple comparison test (2-way ANOVA, ∗∗∗∗P < .0001; ∗∗∗P < .0004; ∗∗P < .006; ∗P < .04). (E) Bar charts represent early apoptosis as percentage of Annexin-V–positive cells from flow cytometry analysis. Error bars indicate SD of 4 independent experiments. Sidak multiple comparison test (2-way ANOVA, ∗∗∗∗P < .0001 each cell group). (F) Kaplan-Meier curves for disease-free survival (DFS) of NSG mice (Charles River Laboratories) that received transplantation with human THP1 cells transduced with indicated constructs (n = 10 mice per group). Six-to-8-week-old NSG mice were injected with 1 × 10⁵ cells, after 5 days antibiotic selection, via tail vein injection. Before transplantation, mice were conditioned with sublethal radiation (2.5 Gy). The log-rank (Mantel-Cox) test was used to compare survival curves. (G) Pathways enrichment analysis (Fisher exact test) of control, PHF8, and F279S cells (n = 3 biological replicates per group). Significantly regulated proteins were compared with the unchanged proteome in the data set based on the gene ontology (GO), KEGG, and keyword terms. (H) Bar charts represent colony numbers of primary mHSPCs cotransduced with MLL-fusions or CDX2 and sgRNAs for specific CRISPR-Cas9–mediated knockout of PHF8 (sgRNA 1, sgRNA 2) or nontargeting control (control). Error bars indicate SDs of 3 independent experiments. Significance was tested using Tukey multiple comparison test (2-way ANOVA, ns). (I) Bar charts represent colony numbers of human AML cells transduced with lenti-CRISPRv2 nontargeting (control) or lenti-CRISPRv2 guide targeting endogenous PHF8 (sgRNA 1 or sgRNA 3). Error bars indicate SDs of 4 independent experiments. Significance was tested using Tukey multiple comparison test (2-way ANOVA, ns); (lower panel) immunoblot analyses showing PHF8 levels. (J) Profile plot showing significant increases of phosphorylation in human THP1 AML cells at defined phosphosites of endogenous PHF8 upon 12 hours of ATRA treatment compared with controls. Each data point is the averaged median of biological quadruplicate of z scored log 2 phosphosite intensities; significance was tested using multiple sample t test (1-way ANOVA, permutation-based false discovery rate [FDR] <0.05). (K) Violin plots depict numbers of colonies of human AML cells treated with 500 nM OKA (OKA 20 minutes incubation) or/and ATRA 10^–8 M. Significance was tested using the Tukey multiple comparison test (2-way ANOVA, ∗P < .05; ∗∗P < .0085, ns). (L) Immunoblot of endogenous total PHF8 protein levels. (M) Western blot analyses (upper panel) of PHF8 levels before and after knocking out in AML cells; (lower panel) typical INT-stained colony pictures. KEGG, Kyoto Encyclopedia of Genes and Genomes; KG-1, Koeffler and Golde 1; MLL, mixed lineage leukemia; sgRNA, single guide RNA.