Fig 2: JMJD3 inhibition promotes in vivo oligodendrocyte-lineage commitment post SCI. a Representative immunoblots of JMJD3, Olig2 and H3K27me3 in JMJD3 inhibition and control group spinal cords. b Quantification of JMJD3, Olig2 and H3K27me3 relative expression levels in inhibition and control group naive spinal cords, n = 6. c Representative IHC image of Olig2 positive cells distribution in JMJD3 inhibition and control group spinal cords, central canal zone (CC). d Quantification of Olig2 positive cells in naive spinal cord sections in GSK-J4 and control groups, n = 6. e Representative immunoblots of Olig2 and GFAP relative expression levels at 1 day, 3 days, 7 days and 14 days post SCI. f Quantification of Olig2 and GFAP relative expression levels on these time points post SCI, n = 6. g Representative immunoblots of Olig2 in JMJD3 inhibition and control group spinal cords tissue around SCI epicenter. h Quantification of Olig2 relative expression in GSK-J4 and control group spinal cords tissue around SCI epicenter, n = 6. i Representative IHC image of Olig2 positive cells distribution post SCI in JMJD3 inhibition and control group (Axial sections around spinal cord epicenter). Red*: Olig2 positive cells, Red + : tube-like structures. j Quantification of Olig2 positive cells and tube-like structures in GSK-J4 and control group spinal cords tissue around SCI epicenter, n = 6. k The oligodendrocyte-lineage associated markers CD133/CD140a in spinal cord lesion post SCI were analyzed by FACS. This is the Representative diagram of FACS assays. l Quantification of CD133/CD140a double positive quadrant percentage in GSK-J4 and control group spinal cord lesion post SCI, n = 3. *: The differences between groups are statistically significant, p < 0.001. Error bars throughout the figure represent the SD (standard deviation) from each independent replicates

Fig 3: MJD3 is induced through activation of the NFkB pathway in a Notch1-dependent mode in T-ALL and binds to NOTCH1 targetsa, Levels of p65 (Rela) protein in control T cells and T-ALL tumor. A representative sample of three mice is shown. b, Schematic representation of the Jmjd3 locus showing the Rela binding site (left panel) and ChIP analysis for p65 (Rela) binding on Jmjd3 locus in mouse control T and T-ALL tumor cells as well as T-ALL cells upon treatment with gSI which affects Notch1 levels (middle panel). NOTCH1 binding to this area upon ?SI treatment in T-ALL cells is also shown (right panel). c, mRNA analysis for JMJD3 and HES1 levels upon ?SI treatment in CUTLL1. The average of three independent studies is shown. d, e, Expression levels of JMJD3 transcript (d) and protein (e) upon treatment of human T-ALL lines (DND41 and CEM) with NEMO binding domain (NBD) inhibitor of the NFkB pathway. f, Jmjd3 levels in T-ALL cells upon inhibition of the NFkB pathway using a dominant negative form of IkBalpha (DN-IkBa). g, h, ChIP for Notch1 (g) and H3K27me3 (h) on Hes1 promoter upon ?SI treatment in murine T-ALL cells. In d and f–h the average of three studies is shown. In e a representative example of three studies is shown. i, Genes correlated with selected human genes (JMJD3, NFkB1, etc) were tested for enrichment in lossof-H3K27me3 genes during the transition to T-ALL in the mouse model. j, Overlap of JMJD3 peaks with peaks of important activating (H3K4me3 and H3K4me1) and repressive (H3K27me3) epigenetic marks as well as members of the NOTCH1 complex. The percentage of TSSs containing JMJD3 peaks is used as a conservative control, alternative to much lower genome-wide JMJD3 occupancy. k, Genome-wide distribution of JMJD3 peaks in human T-ALL.

Fig 4: GSKJ4 inhibitor induces apoptosis and cell cycle arrest of T-ALL-but not myeloid leukemia or physiological LSK-cellsa, Effect of GSKJ4 (at 2µM concentration) on a panel of T-ALL and myeloid lines. The average of three representative studies is shown. b–d, Effects on cell growth (b), apoptosis (c) and cell cycle (d) in three primary T-ALL lines. The average of three representative studies is shown. e, f, Measurement of apoptosis (e, n=3) and cell cycle effects (f, representative study of three experiments) in CUTLL1 cells 72 hours post treatment with the inhibitor. g, h, Apoptosis assays using Annexin V staining of CEM (g) after a period of 72 hours of treatment and measuring caspase 7/9 activity upon treatment of CUTLL1 (h) T-ALL cells with GSKJ5 and GSKJ4 over a period of 24 hours. i, Timecourse studies of Annexin V (upper panel) and cell cycle (lower panel) analysis of CUTLL1 cells over a period of 72h during GSKJ4 treatment according to the scheme on top of the figure. j, Expression of wild-type and catalytic mutant of JMJD3 in T-ALL (CEM) cells. k, Cell growth analysis of T-ALL cells overexpressing wild type and catalytically mutant JMJD3 upon GSJ4 treatment over a period of 72h. Average representation of three independent experiments is shown. l, Cell growth of LSK upon treatment with control (2uM) and different concentrations of GSKJ4 inhibitor. m, Annexin V staining of THP1 cells after a period of 72 hours of GSKJ4 and GSKJ5 (control) treatment at 2mM concentration. Average representation of three independent experiments is shown. n, mRNA levels are shown for three classical NOTCH1 targets (HEY1, NRARP and NOTCH1) over a period of 72h during GSKJ4 treatment. Average representation of three independent experiments is shown.

Fig 5: Silencing KDM6B inhibited lung metastasis of OS in nude mice models. (A) General view of the posterior limb with tumors. (B) Representative images at different timepoints showed the progression of orthotopic tumor and lung metastasis in control and KDM6B-KD groups as examined by in vivo bioluminescent imaging. (C-D) Bioluminescence intensity quantification of orthotopic tumors and lung metastasis (n = 5 per group). (E) Lung tissues in control and two KDM6B-KD groups were excised at day 23 and fixed with Bouin's solution. (F) Representative images of lung tissues subjected to H&E staining in control and two KDM6B-KD groups. (G) Lung metastasis nodules were counted manually. *P < 0.05.