PMC

MF-initiating cells are present in the BM and spleen of JAK2^V617F/Ezh2^Δ/Δ mice. (a) Levels of TPO protein in the serum of WT (n = 6), Ezh2^Δ/Δ (n = 4), JAK2^V617F (n = 4), and JAK2^V617F/Ezh2^Δ/Δ (n = 4) mice 1–2 mo after the deletion of Ezh2. (b) qRT-PCR analysis of Il6 and Il1β in MEPs from WT (n = 6), Ezh2^Δ/Δ (n = 8), JAK2^V617F (n = 8), and JAK2^V617F/Ezh2^Δ/Δ (n = 8) mice 1–2 mo after the deletion of Ezh2. (c and d) Chimerism of CD45.2⁺ donor cells in Mac1⁺ myeloid cells, B220⁺ B cells, and CD4⁺/CD8⁺ T cells in the PB (c) and BM LSK cells (d) of CD45.1⁺ recipients (n = 5 each) at 3 mo after transplantation of BM and spleen LSKs isolated from WT and JAK2^V617F/Ezh2^Δ/Δ mice. (a–d) Bars and asterisks show the mean ± SEM and *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by the Student’s t test; two independent experiments.

The loss of Ezh2 promotes the development of EMH. (a) Spleen weight of WT (n = 5), Ezh2^Δ/Δ (n = 5), JAK2^V617F (n = 5), and JAK2^V617F/Ezh2^Δ/Δ (n = 9) mice 4 wk after the deletion of Ezh2. (b) Proportions of Gr-1⁺ and/or Mac-1⁺ myeloid, B220⁺ B cells, CD4⁺ or CD8⁺ T cells, and Ter119⁺ erythroid cells among CD45⁺ hematopoietic cells in the spleen. (c) Histology of the spleen from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice observed by hematoxylin-eosin staining (top) and silver staining (bottom). Bars, 50 µm. (d) Proportions of LSKs and MkPs (Lin⁻Sca-1⁻c-Kit⁺CD150⁺CD41⁺) in the spleen of WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice (n = 5–9) 4 wk after the deletion of Ezh2. (a, b, and d) Bars and asterisks show the mean ± SEM and *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by the Student’s t test; two independent experiments.

Megakaryocyte-restricted deletion of Ezh2 contributes to the progression of JAK2^V617F-induced fibrosis. (a) qRT-PCR analysis of Ezh2 in LSKs, MkPs, and GMPs from WT and JAK2^V617F/Ezh2^MkΔ/Δ cells at 2–3 mo after transplantation. (b) CBC of WT (n = 5), Ezh2^MkΔ/Δ (n = 4), JAK2^V617F (n = 6), and JAK2^V617F/Ezh2^MkΔ/Δ (n = 10) mice 5 mo after transplantation. (c) Spleen weight of WT (n = 5), Ezh2^MkΔ/Δ (n = 4), JAK2^V617F (n = 4), and JAK2^V617F/Ezh2^MkΔ/Δ (n = 5) mice 5 mo after transplantation. (d) Proportions of LSK cells and MkPs in the spleen of WT and JAK2^V617F/Ezh2^MkΔ/Δ mice (n = 3–5) 5 mo after transplantation. (e) Histology of the BM and spleen from WT, Ezh2^MkΔ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^MkΔ/Δ mice observed by hematoxylin-eosin staining (middle) and silver staining (top and bottom). Bars, 100 µm. (f) BM fibrosis grading (grades 0–3; ) of WT, Ezh2^MkΔ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^MkΔ/Δ (n = 3–5) mice 5 mo after transplantation. (a–d) Bars and asterisks show the mean ± SEM and *, P < 0.05; and **, P < 0.01 by the Student’s t test. (f) Asterisk shows *, P < 0.05 by the Mann–Whitney U test; two independent experiments.

The loss of Ezh2 promotes the development of JAK2^V617F-induced MF. (a) BM cell counts (two femurs and two tibias) of WT (n = 5), Ezh2^Δ/Δ (n = 5), JAK2^V617F (n = 5), and JAK2^V617F/Ezh2^Δ/Δ (n = 9) mice 4 wk after the deletion of Ezh2. (b) Proportions of Gr-1⁺Mac-1⁺ neutrophils, Mac-1⁺ monocytes, B220⁺ B cells, and CD4⁺/CD8⁺ T cells among CD45⁺ hematopoietic cells in the BM 4 wk after the deletion of Ezh2. (c) Histology of the BM from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice observed by hematoxylin-eosin staining (top) and silver staining (bottom). (d) Histology of the BM from WT and JAK2^V617F/Ezh2^Δ/Δ mice observed by hematoxylin-eosin staining (an arrow indicates a megakaryocyte with emperipolesis). (e) Proportions of CD150⁺CD34⁻LSK LT-HSCs in the BM of WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice (n = 3 each) 4 wk after the deletion of Ezh2. (f and g) Proportions of LSKs (f) and GMPs (Lin⁻Sca-1⁻c-Kit⁺CD34⁺FcγR⁺; g) in the BM of WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice (n = 3 each). (h) Proportions of CD71⁺Ter119⁺ double-positive erythroblasts in the BM of WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice (n = 5 each) 4 wk after the deletion of Ezh2. (i) Proportions of Annexin V⁺ apoptotic cells in CD71⁺Ter119⁺ erythroblasts in the BM of WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice (n = 3 each). (a, b, and e–i) Bars and asterisks show the mean ± SEM and *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by the Student’s t test; two independent experiments. Bars: (c) 50 µm; (d) 10 µm.

The loss of Ezh2 severely compromises hematopoiesis in the presence of the JAK2^V617F mutant. (a) Experimental scheme of our model mouse using JAK2^V617F transgenic and Ezh2 conditional KO BM cells. (b) qRT-PCR analysis of Ezh2 in LSKs and MEPs from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice 4 wk after the Cre-mediated deletion of Ezh2. (c) Verification of elimination of Ezh2 protein and levels of H3K27me3 in LK cells detected by Western blotting. (d) CBC of WT (n = 5), Ezh2^Δ/Δ (n = 5), JAK2^V617F (n = 5), and JAK2^V617F/Ezh2^Δ/Δ (n = 10) mice 4 wk after the deletion of Ezh2 and moribund JAK2^V617F/Ezh2^Δ/Δ mice (n = 10). (e) Proportions of myeloid (Gr-1⁺ and/or Mac-1⁺), B220⁺ B cells, and CD4⁺ or CD8⁺ T cells among CD45.2⁺ donor–derived hematopoietic cells in the PB (JAK2^V617F/Ezh2^Δ/Δ n = 10, others n = 5). (f) Dysplastic red blood cells in JAK2^V617F/Ezh2^Δ/Δ mice observed by May-Grünewald-Giemsa staining. Bars, 10 µm. (g) Kaplan-Meier survival curves of WT (n = 6), Ezh2^Δ/Δ (n = 6), JAK2^V617F (n = 8), JAK2^V617F/Ezh2^Δ/+ (n = 9), and JAK2^V617F/Ezh2^Δ/Δ (n = 10) mice; three independent experiments were performed. ***, P < 0.0001 by the log-rank test. (h) CBC of WT (n = 5), Ezh2^Δ/Δ (n = 5), JAK2^V617F (n = 6), and JAK2^V617F/Ezh2^Δ/+ (n = 9) mice 4 mo after the deletion of Ezh2. (a, d, e, and h) Bars and asterisks show the mean ± SEM and *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by the Student’s t test; two independent experiments. (b and c) Data are shown as mean ± SD; two independent experiments.

Brd4 inhibition abrogates the MF-initiating capacity of JAK2^V617F/Ezh2^Δ/Δ cells. (a) Experimental scheme of the JQ1 treatment of chimeric mice reconstituted with WT and mutant cells. (b) Fold changes in chimerism of CD45.2⁺ mutant cells in Mac1⁺ myeloid cells in the PB of recipients (n = 5 each) 8 wk after JQ1 administration compared with that of pre-JQ1 administration (red bars show the mean). (c) Experimental scheme of the JQ1 treatment of WT mice and JAK2^V617F/Ezh2^Δ/Δ mice. (d) Hemoglobin levels of WT (open circles) and JAK2^V617F/Ezh2^Δ/Δ (closed circles) mice pretreatment and 7 d after the completion of the JQ1 treatment. (e) Histology and fibrosis grading (grade 0–3) of the BM from DMSO-treated (n = 7) and JQ1-treated (n = 6) JAK2^V617F/Ezh2^Δ/Δ mice observed by silver staining. Bars, 50 µm. (f) Spleen weight of DMSO-treated (n = 5) or JQ1-treated (n = 4) JAK2^V617F/Ezh2^Δ/Δ mice 4 wk after the end of the treatment. (g) Proportions of LSKs and MkPs in the spleen of DMSO-treated and JQ1-treated JAK2^V617F/Ezh2^Δ/Δ mice 4 wk after the completion of the treatment. (h and i) Chimerism of CD45.2⁺ cells in the PB (h) and BM LSKs, spleen LSKs, and spleen MkPs (i) of CD45.1⁺ recipients (n = 5 each) at 4 mo after transplantation of 10⁶ spleen cells isolated from DMSO-treated or JQ1-treated WT and JAK2^V617F/Ezh2^Δ/Δ mice. (d and f–i) Bars and asterisks show the mean ± SEM and *, P < 0.05 by the Student’s t test (d and f–h) or *, P < 0.05 and **, P < 0.01 by the Mann–Whitney U test (i); two independent experiments.

JAK2^V617F and the loss of Ezh2 cooperatively alter transcriptional programs of hematopoiesis. (a) Venn diagrams showing overlaps of up- and down-regulated genes (left and right, respectively) between Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ LSKs and MEPs isolated 4 wk after the deletion of Ezh2 relative to their WT counterparts. (b) Hierarchical clustering based on total gene expression in LSKs and MEPs isolated from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice (linkage scores are indicated on the right). (c) A principal component (PC) analysis based on total gene expression in LSKs (open circles) and MEPs (closed circles) isolated from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice. (d) GSEA plots for Stat5 up-regulated genes defined in human HSPCs comparing mouse LSKs isolated from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice. (e) Levels of the mean fluorescence intensity (MFI) of phospho-Stat5 (p-Y694) in LSKs from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ (n = 3 each) mice after serum starvation (“(−)”) and stimulation of IL-3 (“IL-3”). Bars show the mean ± SEM; three independent experiments. (f) GSEA plots for canonical PRC2 targets defined in LSK HSPCs comparing LSKs isolated from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice. (g) GSEA for gene expression signatures of HSCs and MkPs comparing LSKs isolated from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ mice (the NES, p-value, and FDR q-value [top and bottom rows, respectively] relative to WT LSKs are shown in each cell). (h) GSEA plots for the HSC signature in LSKs isolated from JAK2^V617F/Ezh2^Δ/Δ compared with JAK2^V617F mice. (a–d and f–h) Experiments used cells from two to four mice individual mice per genotype. (d and f–h) The normalized enrichment score (NES), nominal p-value, and false discovery rate (FDR) q-value are indicated.

Alterations in H3K27me3 upon the loss of Ezh2 results in the activation of specific oncogenes. (a) Fold enrichment (ChIP/input) values of H3K27me3 (TSS ± 2.0 kb of RefSeq genes) in WT and JAK2^V617F/Ezh2^Δ/Δ LSKs 4 wk after the deletion of Ezh2. (b) A scatter plot showing the relationship of fold enrichment (ChIP/input) values of H3K27me3 (TSS ± 2.0 kb of RefSeq genes) between WT and JAK2^V617F/Ezh2^Δ/Δ LSKs 4 wk after the deletion of Ezh2. The light diagonal line represents the borderlines for twofold changes in H3K27me3 levels. (c) Fold enrichment (ChIP/input) values of H3K27me3 at TSS ± 2.0 kb of MkP signature genes in WT and JAK2^V617F/Ezh2^Δ/Δ LSKs. (d) A scatter plot showing the relationship of expression of RefSeq genes in JAK2^V617F/Ezh2^Δ/Δ LSKs relative to WT LSKs and H3K27me3 levels of RefSeq genes in JAK2^V617F/Ezh2^Δ/Δ LSKs relative to WT LSKs and potential oncogenes activated in JAK2^V617F/Ezh2^Δ/Δ LSKs (see also Table S2). (e) ChIP-seq view of H3K27me3 levels at the promoter region of Hmga2 in WT and JAK2^V617F/Ezh2^Δ/Δ LSKs 4 wk after the deletion of Ezh2. (f) qRT-PCR analysis of Hmga2 in LSKs and MEPs from WT, Ezh2^Δ/Δ, JAK2^V617F, and JAK2^V617F/Ezh2^Δ/Δ (n = 3–4) mice 1–2 mo after the deletion of Ezh2. (g) The total cell counts of WT (black lines) and JAK2^V617F (red lines) HSCs transduced with either a control (straight lines) or an Hmga2 (broken lines) retrovirus were monitored for 8 d. (h) Representative pictures of Hmga2-transduced or control vector–transduced WT and JAK2^V617F HSCs on day 8 of the culture observed by May-Grünewald-Giemsa staining. Bars: 10 µm (high magnitude); 100 µm (low magnitude). (i) Proportions of megakaryocytes in Hmga2-transduced or control vector–transduced WT and JAK2^V617F HSCs on day 8 of the culture. (a and c) Boxes and whiskers show the mean and minimum to maximum, and asterisks show ***, P < 0.001 by the Student’s t test; experiments used cells from two to four individual mice per genotype. (f, g, and i) Bars and asterisks show the mean ± SEM and *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by the Student’s t test; two independent experiments.

Brd4 inhibition restores de-repression of PRC2 target genes. (a) Fold enrichment (ChIP/input) values of H3K27ac at the enhancer and the promoter regions in JAK2^V617F/Ezh2^Δ/Δ LK cells 2 wk after the completion of the JQ1 treatment. (b) A scatter plot showing the relationship of fold enrichment (ChIP/input) values of H3K27ac (TSS ± 2.0 kb of RefSeq genes) between DMSO- and JQ1-treated JAK2^V617F/Ezh2^Δ/Δ LK cells (the light diagonal line represents the borderlines for twofold changes in H3K27ac levels). (c) GSEA plot for the canonical PRC2 targets defined in comparing LSKs in JQ1-treated with DMSO-treated JAK2^V617F/Ezh2^Δ/Δ mice. (d) Venn diagrams showing overlaps between PRC2 target genes and JQ1 target genes, which were defined in b and are shown in Table S1. (e) Fold enrichment (ChIP/input) values of H3K27ac at the promoter regions of the canonical PRC2 target genes (defined in ) in JAK2^V617F/Ezh2^Δ/Δ LK cells 2 wk after the completion of the JQ1 treatment. (f) ChIP-seq view of H3K27ac levels at the promoter region of Hmga2 in JQ1-treated or nontreated JAK2^V617F/Ezh2^Δ/Δ LK cells. (g) qRT-PCR analysis of the expression of Hmga2 in LSKs and MEPs isolated from DMSO-treated or JQ1-treated WT mice (n = 3) and JAK2^V617F/Ezh2^Δ/Δ mice (n = 3) 2 wk after the completion of the treatment. (h) GSEA plot for the MkPs signature genes described in comparing LSK cells from JQ1-treated and DMSO-treated JAK2^V617F/Ezh2^Δ/Δ mice. (i) GO biological process gene sets that were enriched more in JQ-treated LSK cells than in DMSO-treated LSK cells as determined by GSEA (gene sizes and FDR q-values are shown in the top and bottom x axes). (a and e) Boxes and whiskers show the mean and minimum to maximum. The asterisks show ***, P < 0.001 by the Student’s t test. (c, h, and i) The normalized enrichment score (NES), nominal p-value, and false discovery rate (FDR) q-value are indicated. (a–f, h, and i) Experiments used cells from two individual mice per group. (g) Bars and asterisks show the mean ± SEM and *, P < 0.05 by the Student’s t test; two independent experiments.