Fig. 6: Rescue the DNMT3A hotspot mutation-induced CpG hypo-methylation and cytokine-independent growth in TF-1 cells.

a Principal component analysis (PCA) using the Infinium methylationEPIC array profiles of TF-1 cells with ectopic expression of the indicated DNMT3A, in comparison to vector control (mock). b Venn diagram using the differentially methylated CpGs (DMCs) exhibiting hypomethylation in TF-1 cells with a DNMT3A hotspot mutant (top: R882C; bottom: R882H) and those re-methylated by an additionally introduced mutation (R676K or M674T/R676K; △beta value greater than 10% and q value less than 0.01 using paired t test). c Heatmap showing the methylation level (beta values) at 20,888 DMCs that were hypomethylated upon transduction of a DNMT3A hotspot mutant when compared to mock. d DNA methylation levels (beta value in the y-axis) of DMCs at EFL4, FOSL2, KDM2A, and HOXB3 in TF-1 cells with ectopic expression of the indicated DNMT3A. e–g Effect of the polymerization-attenuating mutation, either R676K or M674T/R676K, on the GM-CSF (cytokine)-independent proliferation of TF-1 cells, which were transduced with WT DNMT3A (e) or a hotspot mutant (R882C (f) or R882H (g)). The total cell numbers at day 20 post-removal of GM-CSF were used for paired t test (n = 3 biological replicates). Data are mean ± SD. Statistical analysis used a two-tailed paired t test. n.s, not significant (p > 0.05). **p = 0.0013 (R882C vs vector), **p = 0.0056 (R882C vs R882C/R676K), **p = 0.0021 (R882C vs R882C/M674T/R676K), **p = 0.0045 (R882H vs vector), **p = 0.0066 (R882H vs R882H/R676K), **p = 0.0077 (R882H vs R882C/M674T/R676K), and *p = 0.0276 (R882H/R676K vs R882H/M674T/R676K).