p300 suppresses the transition of myelodysplastic syndromes to acute myeloid leukemia

Myelodysplastic syndromes (MDS) are hematopoietic stem and progenitor cell (HSPC) malignancies characterized by ineffective hematopoiesis and an increased risk of leukemia transformation. Epigenetic regulators are recurrently mutated in MDS, directly implicating epigenetic dysregulation in MDS pathogenesis. Here, we identified a tumor suppressor role of the acetyltransferase p300 in clinically relevant MDS models driven by mutations in the epigenetic regulators TET2, ASXL1, and SRSF2. The loss of p300 enhanced the proliferation and self-renewal capacity of Tet2-deficient HSPCs, resulting in an increased HSPC pool and leukemogenicity in primary and transplantation mouse models. Mechanistically, the loss of p300 in Tet2-deficient HSPCs altered enhancer accessibility and the expression of genes associated with differentiation, proliferation, and leukemia development. Particularly, p300 loss led to an increased expression of Myb, and the depletion of Myb attenuated the proliferation of HSPCs and improved the survival of leukemia-bearing mice. Additionally, we show that chemical inhibition of p300 acetyltransferase activity phenocopied Ep300 deletion in Tet2-deficient HSPCs, whereas activation of p300 activity with a small molecule impaired the self-renewal and leukemogenicity of Tet2-deficient cells. This suggests a potential therapeutic application of p300 activators in the treatment of MDS with TET2 inactivating mutations.


Introduction
Normal hematopoiesis is a process that involves the coordination of stem and progenitor cell self-renewal and differentiation. Hematopoietic cell fate is determined by the patterns of gene expression, which are directed by the chromatin structure and the presence or absence of transcription factors (TFs) and cofactors. The regulation of hematopoiesis requires specific and highly ordered epigenetic programming, including dynamic changes in DNA methylation, histone modifications, and chromatin remodeling. Recently, systematic genome profiling screens have identified recurring epigenetic dysregulation in hematopoietic malignancies (1).
Myelodysplastic syndromes (MDS) are among the most common hematological malignancies, with an increased incidence during aging. MDS are stem cell diseases featured by ineffective hematopoiesis with peripheral blood cytopenias, dysplastic and aberrant hematopoietic cell differentiation, and increased leukemia transformation potential (2). Somatic and germline mutations in MDS patients affect genes encoding DNA methylation regulators (TET2 and DNMT3A), chromatin modifiers (ASXL1 and EZH2), splicing factors (SRSF2, U2AF1, and SF3B1), TFs (EVI1, RUNX1, and GATA2), and signaling pathway intermediates (3). Among these mutations, TET2 inactivating mutations are found in approximately 20% of MDS patients but are also seen in patients with acute myeloid leukemia (AML) and myeloproliferative neoplasms (4-7). TET2 converts 5-methylcytosine to 5-hydroxymethylcytosine (5-hmC), a key step Myelodysplastic syndromes (MDS) are hematopoietic stem and progenitor cell (HSPC) malignancies characterized by ineffective hematopoiesis and an increased risk of leukemia transformation. Epigenetic regulators are recurrently mutated in MDS, directly implicating epigenetic dysregulation in MDS pathogenesis. Here, we identified a tumor suppressor role of the acetyltransferase p300 in clinically relevant MDS models driven by mutations in the epigenetic regulators TET2, ASXL1, and SRSF2. The loss of p300 enhanced the proliferation and self-renewal capacity of Tet2-deficient HSPCs, resulting in an increased HSPC pool and leukemogenicity in primary and transplantation mouse models. Mechanistically, the loss of p300 in Tet2-deficient HSPCs altered enhancer accessibility and the expression of genes associated with differentiation, proliferation, and leukemia development. Particularly, p300 loss led to an increased expression of Myb, and the depletion of Myb attenuated the proliferation of HSPCs and improved the survival of leukemia-bearing mice. Additionally, we show that chemical inhibition of p300 acetyltransferase activity phenocopied Ep300 deletion in Tet2-deficient HSPCs, whereas activation of p300 activity with a small molecule impaired the self-renewal and leukemogenicity of Tet2-deficient cells. This suggests a potential therapeutic application of p300 activators in the treatment of MDS with TET2 inactivating mutations.
JCI Insight 2021;6(19):e138478 https://doi.org/10.1172/jci.insight.138478 in the dynamic regulation of DNA methylation. The loss of Tet2 in hematopoietic cells leads to a global loss of 5-hmC and hematopoietic defects, including enhanced hematopoietic stem cell (HSC) self-renewal, myeloid cell expansion, and an increased propensity to develop MDS or acute leukemia (8)(9)(10)(11)(12)(13). p300, encoded by the Ep300 gene, is a lysine acetyltransferase (KAT) that plays pivotal roles in cellular proliferation and signal transduction by acting as a transcriptional cofactor and chromatin modifier (14). In normal hematopoietic cells, p300 regulates HSC self-renewal and differentiation (15) and the loss of p300 impairs B and T cell differentiation (16,17). In malignant hematopoiesis, it is currently unclear whether p300 plays a suppressive or an oncogenic role. For example, in AML1-ETO-driven leukemia, p300 acetylates AML1-ETO to promote leukemia and inhibition of p300 KAT activity impairs leukemia cell growth (18,19). However, Ep300 mutations are found in patients with myeloid neoplasia and the low expression level of p300 is a poor prognostic marker (20). Our previous work also identified a tumor suppressor role of p300 in myeloid transformation based on a Nup98-Hoxd13-transgenic mouse model of MDS (21,22). Comutations of EP300 and TET2 are found in patients with myeloid neoplasms (23). Moreover, p300 directly acetylates TET2 and these enzymes cooperate to regulate HSC enhancer function and facilitate transcription (24,25). Yet, the direct molecular and biological effects of p300 deletion or inhibition on HSC homeostasis in the absence of TET2 are still unknown.
Here, we demonstrate that p300 is a general epigenetic suppressor of MDS progression using 4 clinically relevant, genetically engineered MDS mouse models: Tet2 +/-, Tet2 -/-, Asxl1 +/and Srsf2 P59H . In all models, Ep300 deletion shortened the survival of the mice and accelerated the development of MDS/AML. We focused primarily on the influence of p300 loss in Tet2-deficient hematopoiesis and found that the loss of p300 promoted the self-renewal and proliferation of Tet2-deficient hematopoietic stem and progenitor cells (HSPCs), resulting in an increased HSPC pool in the bone marrow and accelerated leukemia transformation. Mechanistically, the loss of p300 in Tet2-deficient HSPCs affected the interplay between enhancers and TF networks altering the expression of genes related to differentiation, signal transduction, immune response, cell proliferation, and oncogenicity. In particular, we found that the loss of p300 increased the expression of the proto-oncogene Myb in multiple MDS models, identifying a common mechanism underlying the accelerated disease progression seen after p300 loss. Accordingly, the depletion of Myb reversed the hyperproliferative phenotype of HSPCs lacking both p300 and TET2. We also show that manipulating p300 catalytic activity by small molecules exacerbated or mitigated Tet2-deficient hematologic malignancies.

Results
The loss of p300 in Tet2-deficient mouse models shortened survival and accelerated the development of MDS and progression to AML. To investigate the role of p300 in MDS progression driven by Tet2 deficiency, we crossed Tet2-deficient (Tet2 -/and Tet2 +/-) mice with hematopoietic-specific conditional knockout Ep300 mice and generated primary and transplantation mouse models. The deletion of Ep300 was verified by quantitative RT-PCR in the bone marrow cells 2 weeks after poly(I:C) administration (Supplemental Figure 1, A and B; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.138478DS1).
The loss of p300 in Tet2-deficient mice greatly accelerated disease development and significantly shortened their survival ( Figure 1A). The moribund Ep300Δ/ΔTet2 -/and Ep300Δ/ΔTet2 +/mice showed an increased WBC count ( Figure 1B) and splenomegaly and granulocytic sarcomas ( Figure 1C and Supplemental Figure  1C). Increased blasts in the peripheral blood and bone marrow and splenic infiltration were also observed in Ep300Δ/ΔTet2 -/and Ep300Δ/ΔTet2 +/mice ( Figure 1D), suggesting the onset of leukemia. The flow cytometry data confirmed that the loss of p300 increased the frequencies of c-Kit + cells and lineage -Sca1 + c-Kit + (LSK) cells in the bone marrow (Figure 1, E-G); we also found an increased percentage of myeloid cells (Mac1 + ; Figure 1H) and a decreased percentage of B cells (B220 + ; Figure 1I). We found no difference in the RBC count, hemoglobin, platelet count, or the percentage of T cells (CD3 + ) in the bone marrow of the various moribund mice (Supplemental Figure 1, D-G). The results from the transplantation models further confirmed the shortened survival and accelerated leukemia development ( Figure 1J). In contrast, the deletion of p300 in WT mice did not alter hematopoiesis significantly (Figure 1 and Supplemental Figure 1). Taken together, our in vivo results indicate that the loss of p300 in the context of monoallelic or biallelic loss of Tet2 accelerated disease progression and the onset of leukemia.
The loss of p300 promoted the proliferation and self-renewal of Tet2-deficient HSPCs. To further examine the initial consequences of p300 loss on hematopoiesis in Tet2-deficient mice, we analyzed HSPC numbers and function 2 weeks after Ep300 deletion. As shown in Figure 2, A and B, and Supplemental Figure 2A  frequency and absolute number of HSPCs were significantly increased after Ep300 deletion in both Tet2 -/and Tet2 +/mice, but not in the WT mice, suggesting that the early loss of p300 perturbed HSPC function in the context of Tet2 deficiency. Further analysis revealed that p300 loss in Tet2-deficient mice decreased the frequency of long-term HSCs (LT-HSCs) and increased the frequencies of multipotent hematopoietic progenitors (MPPs) and common myeloid progenitors within the HSPC population ( Figure 2, C-E, and Supplemental Figure 2, B and C). This suggests that the loss of p300 in Tet2-deficient mice impaired HSPC differentiation and enhanced skewing toward myeloid progenitors.
Next, we performed serial replating assays to investigate the self-renewal and proliferation of Tet2-deficient HSPC after p300 loss. As shown in Figure 2F, the loss of p300 enhanced the replating capacity of Tet2 +/-HSPCs, suggesting an enhanced self-renewal. Moreover, the loss of p300 increased the number of colonies generated by Tet2-null HSPCs, indicating an enhanced proliferation as well (Figure 2, G and H). In vivo Ki67 and BrdU assays showed increased frequencies of Ki67 + cells and BrdU + cells in Ep300Δ/ΔTet2 -/-HSPCs compared with Tet2 -/-HSPCs, further confirming the hyperproliferative phenotype of Ep300Δ/ΔTet2 -/-HSPCs ( Figure 2I). Meanwhile, the deletion of p300 did not affect the self-renewal of normal HSPCs (Supplemental Figure 2D). Taken together, the in vivo and in vitro results demonstrate that the loss of p300 enhanced the self-renewal and proliferation of Tet2-deficient HSPCs, increasing the bone marrow HSPC pool and accelerating leukemia development of the doubly deficient mice.
The loss of p300 reprogrammed the epigenetic landscape and transcription network to enhance the proliferation and leukemogenicity of Tet2-null HSPCs. Because TET2 and p300 played cooperative roles in maintaining the epigenome and enhancer landscape of HSPCs (26,27), and our phenotypic data indicate that the loss of p300 in Tet2-deficient mice accelerated MDS progression, we next examined the TET2-independent function of p300 in HSPCs by comparing the epigenetic and gene expression profile of Tet2 -/-HSPCs, before and after p300 loss.
Because TET2 catalyzes the conversion of 5 mC to 5 hmC and promotes DNA demethylation, to determine whether further loss of p300 affects DNA hydroxymethylation status, we performed DNA hydroxymethylation profiling (5-hMeDIP-Seq). As expected, Tet2 -/-HSPCs showed a global reduction in 5-hmC-enriched regions, compared with WT HSPCs. The deletion of Ep300 in WT HSPCs resulted in a modest loss of 5-hmC (Supplemental Figure 3A), possibly reflecting the previously reported effects of p300-mediated acetylation on TET2 stability and activity (24). However, the deletion of Ep300 in Tet2-null HSPCs induced minimal changes on global DNA hydroxymethylation levels (Supplemental Figure 3A).
Because p300 acetylates histone H3 at K27, a mark associated with active transcription and enhancer activity, we examined how p300 loss affects the distribution of H3K27ac and the enhancer landscape, in Tet2-null lineage -(Lin -) bone marrow cells by performing ChIP-Seq analyses (Supplemental Figure 3B). In agreement with its role as transcriptional coactivator, the deletion of Ep300 resulted in a genome-wide reduction of H3K27ac signal (Supplemental Figure 3B). The loss of p300 in Tet2-null cells led to a reduction in H3K27ac signal at 8950 genomic sites (Supplemental Figure 3C). Genes annotated to H3K27ac lost peaks were significantly enriched in pathways involved in differentiation, signal transduction, chronic myeloid leukemia, and TNF signaling ( Figure 3A and Supplemental Table 1). We then identified active enhancers by overlapping H3K27ac and H3K4me1 ChIP-Seq peaks and analyzed the enhancer landscape of Tet2-null cells after p300 loss. Notably, p300 loss dramatically reprogrammed the enhancer landscape in Tet2-null cells, with a loss of 5204 active enhancers and a gain of 1732 active enhancers (Supplemental Figure 3D and Supplemental Table 2). Motif analyses of the altered enhancers showed enrichment for binding sites of key hematopoietic regulators including Ets, Runx, Gata, Irf8, NF-κB, and Myb family members ( Figure 3B and Supplemental Figure 3D). These data reveal that p300 controlled the activity of a specific subset of enhancers in Tet2-deficient hematopoietic cells, enhancers that control the expression of genes involved in differentiation and leukemic transformation.
Because the regulation of gene expression by enhancers can be reflected in chromatin accessibility, we performed ATAC-Seq assays using HSPCs from Ep300Δ/ΔTet2 -/and Tet2 -/mice (Supplemental Figure  3E). We found that p300 loss in Tet2-null HSPCs triggered alterations in chromatin accessibility, mostly at the intronic and intergenic regions (potential enhancer regions), and most often associated with a gain in ATAC-Seq signal (Supplemental Figure 3F). These additional epigenetic alterations may have occurred due to secondary or indirect effect of p300 loss in Tet2 -/cells. Of note, increases in chromatin accessibility were reported in Tet2 -/hematopoietic progenitors and AML cells, which were also suggested to be induced indirectly or secondarily by TET2 loss (25). Interestingly, changes (gains and losses) in ATAC-Seq signals   in Ep300Δ/ΔTet2 -/-HSPCs (compared with Tet2 -/-HSPCs) annotated to genes related to UV response, signal transduction, inflammatory response, and hypoxia ( Figure 3C and Supplemental Table 3). To better identify potential TF binding patterns affected by p300 loss in the context of Tet2 -/cells, we analyzed our ATAC-Seq data using a recently developed computational method, diffTF (25,28). The loss of p300 in Tet2null HSPCs increased the binding accessibility of Spi1, Gata, Tal1, Irf, and Cebp TF families and decreased the binding accessibility of NF-κB and Myb family members ( Figure 3D), indicating a major impact on transcription networks that control HSPC function.
To determine whether the alterations in the enhancer landscape induced by p300 loss in Tet2 -/-HSPCs were associated with changes in gene expression, we performed RNA-Seq and Gene Set Enrichment Analysis (GSEA) using HSPCs isolated from WT, Ep300Δ/Δ, Tet2 -/and Ep300Δ/ΔTet2 -/mice (Supplemental Figure 4, A-C, and Supplemental Table 4). The loss of p300 or Tet2 in normal HSPCs altered the expression of genes related to signaling pathways and proliferation, which is consistent with previous reports (29,30). In Tet2-null HSPCs, p300 loss led 939 genes downregulated and 438 genes upregulated (Supplemental Figure 4B). GSEA analyses of the differentially expressed (DE) genes in Tet2-null HSPCs following Ep300 deletion showed downregulation of genes that control hematopoietic differentiation (such as B cell and T cell differentiation), signal transduction (such as WNT, Notch and KRAS signaling), as well as inflammatory response (including IL2-STAT5 signaling and TNF-a signaling), and apoptosis ( Figure 3E and Supplemental Figure 4, C and D). Notably, we observed a positive enrichment in gene pathways governing DNA repair and proliferation, such as E2F targets and Myc targets ( Figure 3E). These results indicate that p300 loss triggered major alterations in the epigenomic landscape of Tet2 -/-HSPCs, resulting in altered gene expression patterns that affect their proliferation and differentiation, which is consistent with the hyperproliferative and enhanced self-renewal capacity of Ep300Δ/ΔTet2 -/-HSPCs ( Figure 2).
To assess the association between the altered enhancer landscape and differential gene expression after p300 loss in Tet2-null hematopoiesis, we overlapped the DE genes and the genes that annotated to the altered enhancers. As shown in Figure 3F, out of 939 downregulated genes after p300 loss, 284 genes showed a loss of nearby active enhancers (i.e., loss of H3K27ac enrichment at intronic and intergenic regions marked by H3K4me1). These 284 genes are enriched for pathways controlling immune response, hypoxia, apoptosis, signal transduction, and differentiation ( Figure 3G), and included important regulators of hematopoiesis such as Dntt and Notch1 (Figure 3, H and I; refs. 31, 32). These changes likely reflect a direct effect of p300 on the enhancer accessibility and transcription output of Tet2 -/cells.
The elevated Myb expression induced by p300 loss in Tet2-null HSPCs influenced the hyperproliferative phenotype. In addition to the downregulated genes, associated with a loss of nearby putative enhancers after p300 depletion in Tet2-null HSPCs, we observed a small portion of genes that gained expression and gained active enhancers (i.e., enrichment of H3K27ac at intronic and intergenic regions marked by H3K4me1) in Ep300Δ/ΔTet2 -/-HSPCs compared with Tet2 -/-HSPCs ( Figure 4A). Among these genes, we identified the Myb locus and the Hoxb gene cluster (Figure 4, A and B, and Supplemental Figure 5, A and B), which have both been directly associated with enhanced proliferation and oncogenicity of HSCs (33,34). The ChIP-X Enrichment Analysis (ChEA) of the DE genes between Ep300Δ/ΔTet2 -/and Tet2 -/-HSPCs also revealed a significant enrichment for Myb target genes ( Figure 4C). Thus, we hypothesized that the elevated expression of Myb triggered by p300 loss in Tet2-null HSPCs may have driven the changes in gene expression and subsequent alterations in HSPC function, including malignant transformation. To evaluate this hypothesis, we identified Myb target genes by performing ChIP-Seq analyses of Myb in WT Lincells and compared them with the gene expression profile (Supplemental Table 5). Notably, 350 DE genes in Ep300Δ/ΔTet2 -/-HSPCs compared with Tet2 -/-HSPCs were identified as targets of Myb ( Figure 4D); of these 350 genes, 303 of them were downregulated and enriched in pathways related to signal transduction, immune response, and apoptosis ( Figure 4E), verifying that changes in Myb expression and activity may have accounted for some of the key transcriptional alterations seen in Tet2 -/-HSPCs after p300 loss. Myb is known to promote the survival and proliferation of HSCs and leukemia stem cells (34)(35)(36); thus, we performed in vitro and in vivo assays, using Myb-depleted cells, to determine the impact of depleting Myb on the proliferation of HSPCs lacking p300 and TET2 (Supplemental Figure 5C). The depletion of Myb significantly attenuated the in vitro colony-forming capacity of Ep300Δ/ΔTet2 -/-HSPCs, but not Tet2 -/-HSPCs ( Figure 4F and Supplemental Figure 5D). We also depleted Myb in the leukemia cells from Ep300Δ/ΔTet2 -/mice and performed transplantation assays. As shown in Figure 4G, mice receiving Myb-depleted cells showed prolonged survival, compared with the mice receiving scrambled shRNA-infected cells. These results suggest that the increased Myb expression caused by p300 loss contributed to the proliferation and oncogenicity of Tet2-null HSPCs. The loss of p300 accelerated disease progression and shortened survival in multiple MDS models. Although p300 loss in normal HSPCs has mild phenotypic consequences, its deletion in HSPCs lacking Tet2 greatly alters HSPC behavior, triggering the development of AML. This suggests that p300 activity blocks oncogenic transformation of MDS/AML. We then examined the effects of p300 loss in other mouse models of MDS, driven by Asxl1 haploinsufficiency or an Srsf2 mutation (Srsf2 P95H ; ref. 3; Figure 5 and Supplemental Figure  5E). As shown in Figure 5, p300 loss shortened the survival of recipient mice, due to accelerated disease development, in both the Asxl1 +/and Srsf2 P95H bone marrow transplantation models ( Figure 5, A and C). The loss of p300 increased Myb expression in both Asxl1 +/and Srsf2 P95H HSPCs ( Figure 5, B and D), highlighting the consistent upregulation of Myb in the effect of p300 loss on MDS models driven by an array of mutations, which represent up to 70% of the genetic abnormalities found in MDS patients.
Augmenting p300 KAT activity attenuated the enhanced self-renewal and proliferative capacity of Tet2-deficient HSPCs. Given the tumor suppressor role of p300 in multiple MDS models, we next examined whether chemical modulation of p300 KAT activity using an inhibitor (A-485; refs. 37,38) and an activator (I-CBP112; ref. 39) would similarly affect the self-renewal and proliferation of Tet2-deficient HSPCs (Figure 6A). First, we demonstrated that A-485 effectively inhibited, whereas I-CBP112 activated, the KAT activity of p300 toward histone H3K18 and H3K27 in the in vitro assays ( Figure 6B). We then treated Tet2-deficient HSPCs with these compounds and performed serial replating assays. We found that A-485 enhanced the proliferation of Tet2 -/-HSPCs and the replating capacity of Tet2 +/-HSPCs ( Figure 6, C-E), phenocopying Ep300 deletion in Tet2-deficient HSPCs. Conversely, I-CBP112 impaired the self-renewal and growth of both Tet2 -/and Tet2 +/-HSPCs ( Figure 6, C-E). Neither A-485 nor I-CBP112 influenced the self-renewal of normal HSPCs (Supplemental Figure 5F). Because increased Myb expression appears to be a p300-dependent event, we examined the expression of Myb in Tet2 -/-LK (Linand c-Kit + ) cells  after treatment in vitro with the p300 KAT modulators. We found increased Myb expression after A-485 treatment and decreased Myb expression after I-CBP112 treatment ( Figure 6F), indicating that the regulation of Myb expression by p300 was dependent on its acetyltransferase activity.
We next evaluated the effect of I-CBP112 on the growth of Tet2-null leukemia cells in vivo. As shown in Figure 6G, mice receiving DMSO-treated Tet2-null leukemia cells developed aggressive AML with severe anemia (4 weeks after transplantation), whereas recipients of I-CBP112-treated Tet2-null leukemia cells showed improved survival with less severe anemia. Altogether, our data indicate that the KAT activity of p300 was required to limit the self-renewal, proliferation, and leukemogenicity of Tet2-null HSPCs, in part by controlling the expression of Myb. Small molecules that enhance p300 KAT activity could have been an effective therapeutic strategy for those hematopoietic malignancies that contain decreased TET2 activity.

Discussion
We have uncovered a general role of p300 on the behavior of premalignant (Tet2-deficient, ASXL1 +/-, and SRSF2 P95H ) HSPCs. In Tet2-deficient mice, the loss of p300 promoted the proliferation of HSPCs with myeloid skewing, resulting in an increased HSPC pool in the bone marrow and accelerated disease progression in vivo. Mechanistically, the lack of p300 activity disrupted the epigenetic landscape of Tet2-null HSPCs, as shown by the cumulative alterations in enhancer activity and chromatin accessibility observed in Ep300Δ/ΔTet2 -/cells compared with Tet2 -/cells, highlighting an important function of p300 at the enhancers that control HSPC malignant transformation. Because p300 acetylates H3K27ac, which marks active enhancers, Ep300 deletion triggered the loss of a large group of enhancers containing motifs for key hematopoietic regulators, such as NF-κB, Gata, Runx, Ets, and Myb family members ( Figure 3B and Supplemental Figure 3D). These results aligned with our chromatin accessibility data, with motifs for the same hematopoietic TFs showing altered accessibility after p300 deletion ( Figure 3D). As a consequence of this reprogrammed epigenome, the gene expression profile of Tet2-null HSPCs after p300 loss showed downregulation of pathways involved in B cell and T cell differentiation and signal transduction. For example, enhancer marks within the Notch1 locus were decreased in Tet2-null cells by Ep300 deletion and gene expression was also decreased. Silenced Notch signaling pathway has been directly linked to MLL-AF9-driven acute leukemogenesis (40). Our results provide another link between downregulated Notch1 signaling and cancer cells with diminished TET2 function and suggest a therapeutic potential of Notch receptor agonists in this setting. Although p300 loss in Tet2-null HSPCs resulted in a large group of lost enhancers, we still observed a small group of gained enhancers in Ep300Δ/ΔTet2 -/-HSPCs, some of which were associated with increased expression of genes, such as the Hoxb gene cluster and Myb, which are related to HSPC stemness and proliferation (33,34).
The impact of p300 on MDS hematopoiesis is multifactorial, but our mechanistic and biological data implicate Myb as a critical player in Tet2-null hematopoiesis. First, p300 loss reduced the accessibility of Myb to chromatin and the expression of a subset of Myb target genes. Second, p300 loss increased Myb expression and the accessibility of the Myb enhancer. Third, we found that the depletion of Myb attenuated the hyperproliferative phenotype of Ep300Δ/ΔTet2 -/-HSPCs. Consistent with our data, it has been reported that blocking the interaction between Myb and p300 increases HSC numbers, and impairs lymphopoiesis and erythropoiesis, in mice (41). Thus, p300 deletion may affect the function and protein-protein interactions of Myb, contributing to defects in differentiation and signaling pathways and the enhanced proliferation and leukemogenicity that characterizes Tet2-null HSPCs. Although further studies are needed to decipher the precise mechanisms whereby p300-dependent regulation of Myb controls MDS pathogenesis, our study further implicates Myb as a promising target for therapy in MDS or AML patients (42).
Many studies have characterized p300 KAT activity and function in various cell types, including the identification of the p300 acetylome (43) and the potential use of p300 KAT modulators (inhibitors and activators) in solid tumor and leukemia treatment (44). We found that a p300 KAT inhibitor promoted the proliferation of Tet2-null HSPCs, whereas a KAT activator impaired their self-renewal and proliferative capacity.
In summary, we found that the loss of p300 or its chemical inhibition induced (a) a profound reprogramming of the epigenome and transcriptome of premalignant HSPCs, leading to impaired differentiation and signal transduction; (b) a unique gene expression signature in Tet2-null HSPCs, related to enhanced proliferation and oncogenicity; (c) an increase in Myb expression, which directly amplifies malignant transformation of MDS HSPCs; (d) an accelerated disease progression in multiple MDS models. Lastly, we demonstrate the therapeutic potential of enhancing p300 catalytic activity to treat MDS/AML patients with deficient TET2 function, and potentially individuals with Tet2-mutant clonal hematopoiesis.

Transplantation of Ep300Δ/ΔTet2 -/leukemia cells after Myb knockdown. Leukemia cells obtained from
Ep300Δ/ΔTet2 -/mice were cultured in X-VIVO medium with IL-3 (10 ng/mL), IL-6 (10 ng/mL), and SCF (100 ng/mL) overnight and infected with lentiviruses, which express shMyb or scrambled shRNA in a modified PLKO-RFP-expressing vector using a spinfection protocol. Three days after infection, RFP + cells were sorted and 250,000 cells/mouse were transplanted into lethally irradiated recipients with 500,000 normal bone marrow cells. The mice were monitored and euthanized when moribund.
Ex vivo treatment of I-CBP112 in Tet2 -/leukemia cells. Leukemia cells from the bone marrow of Tet2 -/mice were cultured in vitro and treated with DMSO and I-CBP112 for 24 hours. Recipient mice (C57BL/6. SJL mice) were lethally irradiated and transplanted with 2 × 10 6 treated Tet2 -/leukemia cells with 500,000 helper cells from WT mice (C57BL/6). Peripheral blood was collected at indicated time point.
ChIP ChIP-Seq and analysis. Reads were trimmed for adapters using Cutadapt (v1.15 -nextseq-trim=20 -m 18). Fastq files were aligned to mouse mm10 using BWA (v0.7.13 aln -q 5 -l 32 -k 2) and deduplicated using Picard Mark Duplicates (v2.10.6). For H3K27ac, H3K4me3, H3K4me1, and Myb, peaks were determined by overlapping called narrow peaks by macs2 (v2.1.1.20160309) from pseudo-replicates with q < 0.05 with H3 used as background for histone marks and IgG used as background for Myb. Reads were quality filtered (Mapq > 30) and the mm10 blacklist (46). Shift and extension sizes were determined using phatompeakqualtools (v1.1). ChIPseeker (v1.14.0) and R (v3.4.1) were used for genomic annotation with promoters defined as a 3 kb window around the TSS (Ensembl). Bigwig signal files were generated by Macs2 as fold change over background. Peak overlaps were determined by merging peaks list from a given comparison and intersecting individual peak files to the combined list by at least 1 bp. Putative enhancers were determined by intersecting H3K27ac and H3K4me1 peak regions outside of promoter regions.
RNA-Seq and analysis. Linbone marrow cells were enriched using the Direct Lineage Cell Depletion kit with LS columns (Miltenyi Biotec). Enriched Lincells were stained with the mouse bone marrow biotin lineage panel, Streptavidin APC-cy7, APC-c-Kit, and PEcy7-Sca-1, and after staining all HSPCs were sorted. Total RNA was extracted from 1 × 10 5 to 2 × 10 5 sorted HSPCs using AllPrep DNA/RNA Plus micro kit (QIAGEN). The quality of RNA was evaluated on an Agilent Bioanalyzer (RIN>9). Ribosomal RNA was removed using the NEBNext rRNA Depletion kit (New England Biolabs E6310) following the manufacturer's instructions, and RNA-Seq libraries were prepared using the Universal Plus mRNA-Seq kit (NuGEN Technologies). RNA-Seq libraries were sequenced on an Illumina HiSeq 2500 platform (pairedend, 100-bp reads) at the Sylvester Oncogenomics Shared Resource, to obtain more than 40 million pairedend reads per sample. Sequencing and analysis were performed of 3 independent biological replicates per genotyping group. Reads were aligned to the mm10 transcriptome and ERCC spike-in indexes using RSEM (v1.3.0) and STAR (v2.6.0c) with Gencode annotations. Differentially expressed genes were determined by DESeq2 (v1.18.1, Wald test, adjusted P < 0.05) after gene counts were corrected based on ERCC variances using RUVseq (v1.12.0). GSEA was performed using GSEA (v3.0) with genes preranked by the Wald statistic. Gene Ontology Analysis was performed using the Enrichr API and sorted by adjusted P value. Sample blind variance stabilized and ERCC-corrected log2 transformed gene counts and transcripts per million were used for plots with expression data. Heatmaps, GSEA and GO bar plots, and Venn diagrams were generated using python (v3.6.3), scipy (v1.1.0), matplotlib, and seaborn (v.0.8.1).
JCI Insight 2021;6(19):e138478 https://doi.org/10.1172/jci.insight.138478 5-hMeDIP, sequencing and analysis. Genomic DNA from sorted HSPCs was isolated using the All-Prep DNA/RNA Plus micro kit (QIAGEN). Approximately 200 ng of genomic DNA was sonicated followed by end-repair, A-tailing, and adaptor ligation using standard Illumina protocols. Ten percent of the volume was retained for Input. Immunoprecipitation with anti-5-hmC antibodies (1 μg/mL, Active Motif) was performed at 4°C overnight and immunoprecipitated DNA was recovered using protein G magnetic beads (Invitrogen). After proteinase K digestion, DNA was purified from all input and immunoprecipitated samples and amplified by PCR for library preparation. Libraries were sequenced by an Illumina HiSeq 2500 sequencer by the Cancer Genomic Core Facility at University of Miami. Reads were trimmed using cutadapt (-m25), aligned using bowtie2 (v2.0.5) to the mouse mm10 genome build, quality filtered (mapq > 30), and deduplicated. Peaks were called using macs2 (-g mm --nomodel --extsize 200 -B) and differential peaks were determined using macs2 bdgdiff from the intermediate lambda and pileup bed graphs generated from macs2 call peaks.
Moreover, we applied DiffTF software (https://git.embl.de/grp-zaugg/diffTF), which was developed to identify the quantification of TF activity using ATAC-Seq data (25,28). Briefly, the software identified the binding sites of 422 TFs using the mouse PWMs deposited in the database Hocomoco v10, and PWMScan. For each TF we then extracted the signal around each binding site (± 100 bp around the core motif) and calculated a fold change between 2 conditions. Welch's 2-sample t test was used for significance of the differences in a comparison.
Data availability. 5-hMeDIP-Seq, RNA-Seq, and ChIP-Seq data sets generated in this study have been deposited in the Gene Expression Omnibus database under accession number GSE145878.
Statistics. Kaplan-Meier curves were used for survival analysis. Graphpad Prism 6 was used for flow cytometry assays, complete blood count, colony-forming unit and serial replating assays, and quantitative RT-PCR assays. P values of less than 0.01 and 0.05 were considered significant. All graphs represent mean ± SEM. Survival curves were estimated using the Kaplan-Meier method and compared by the log-rank test. Two-tailed Student's t test, one-way ANOVA tests, and two-way ANOVA tests were applied (indicated in each figure).
Study approval. All animal studies were conducted in compliance with Guide for the Care and Use of Laboratory Animals (National Academies Press, 2011) and were approved by the IACUC of the University of Miami (protocol no. 18-164).

Author contributions
NM, GM, and SDN designed the experiments, and wrote and edited the manuscript. NM and GM conducted and analyzed the experiments, with input from SDN, PAC, LC, FCY, and MX. MTM, DLK, and MEF contributed to the 5-hMeDIP-Seq and conducted the analysis. GM, DLK, QY, and JS contributed to the ChIP-Seq, ATAC-Seq, and RNA-Seq experiments and conducted the analysis. HI and SC contributed to the pathology analysis. CM contributed to the in vitro acetylation assays, generated plasmids, and produced the lentiviral reagents. SD assisted with animal experiments. FL, JS, PJH, YX, and CC assisted with the Myb studies, BEZ and PAC contributed to the A-485 and I-CBP112 studies.