Loss of the de novo DNA methyltransferases Dnmt3a and Dnmt3b in embryonic stem cells obstructs differentiation; however, the role of these enzymes in somatic stem cells is largely unknown. Using conditional ablation, we show that Dnmt3a loss progressively impairs hematopoietic stem cell (HSC) differentiation over serial transplantation, while simultaneously expanding HSC numbers in the bone marrow. Dnmt3a-null HSCs show both increased and decreased methylation at distinct loci, including substantial CpG island hypermethylation. Dnmt3a-null HSCs upregulate HSC multipotency genes and downregulate differentiation factors, and their progeny exhibit global hypomethylation and incomplete repression of HSC-specific genes. These data establish Dnmt3a as a critical participant in the epigenetic silencing of HSC regulatory genes, thereby enabling efficient differentiation.

Chuvash polycythemia is an autosomal recessive disorder that is endemic to the mid-Volga River region. We previously mapped the locus associated with Chuvash polycythemia to chromosome 3p25. The gene associated with von Hippel-Lindau syndrome, VHL, maps to this region, and homozygosity with respect to a C-->T missense mutation in VHL, causing an arginine-to-tryptophan change at amino-acid residue 200 (Arg200Trp), was identified in all individuals affected with Chuvash polycythemia. The protein VHL modulates the ubiquitination and subsequent destruction of hypoxia-inducible factor 1, subunit alpha (HIF1alpha). Our data indicate that the Arg200Trp substitution impairs the interaction of VHL with HIF1alpha, reducing the rate of degradation of HIF1alpha and resulting in increased expression of downstream target genes including EPO (encoding erythropoietin), SLC2A1 (also known as GLUT1, encoding solute carrier family 2 (facilitated glucose transporter), member 1), TF (encoding transferrin), TFRC (encoding transferrin receptor (p90, CD71)) and VEGF (encoding vascular endothelial growth factor).

Fusobacterium species are part of the gut microbiome in humans. Recent studies have identified over-representation of Fusobacterium in colorectal cancer (CRC) tissues but it is not yet clear whether this is pathogenic or simply an epiphenomenon. In this study, we evaluated the relationship between Fusobacterium status and molecular features in CRCs through quantitative real-time PCR in 149 CRC tissues, 89 adjacent normal appearing mucosae and 72 colonic mucosae from cancer-free individuals. Results were correlated with CpG island methylator phenotype (CIMP) status, microsatellite instability (MSI) and mutations in BRAF, KRAS, TP53, CHD7 and CHD8. Whole exome capture sequencing data were also available in 11 cases. Fusobacterium was detectable in 111/149 (74%) CRC tissues and heavily enriched in 9% (14/149) of the cases. As expected, Fusobacterium was also detected in normal appearing mucosae from both cancer and cancer-free individuals but the amount of bacteria was much lower compared to CRC tissues (a mean of 250-fold lower for Pan-fusobacterium). We found the Fusobacterium-high CRC group (FB-high) to be associated with CIMP positivity (p=0.001), TP53 wild type (p=0.015), hMLH1 methylation positivity (p=0.0028), MSI (p=0.018) and CHD7/8 mutation positivity (p=0.002). Among the 11 cases where whole exome sequencing data was available, two that were FB-high cases also had the highest number of somatic mutations (a mean of 736 per case in FB-high vs. 225 per case in all others). Taken together, our findings show that Fusobacterium enrichment is associated with specific molecular subsets of CRCs, offering support for a pathogenic role in CRC for this gut microbiome component

386 Aberrant genomic DNA methylation patterns are widely reported in human cancers but the prognostic value and pathological consequences of these marks remain uncertain. CpG methylation is catalyzed by a family of DNA methyltransferase enzymes comprised of three members – Dnmt1, Dnmt3a and Dnmt3b. Mutations in the de novo DNA methyltransferase enzyme DNMT3A have now been reported in over 20% of adult acute myeloid leukemia (AML) and 10–15% of myelodysplastic syndrome (MDS) patients. However, analysis of promoter methylation and gene expression in these patients has thus far failed to yield any mechanistic insight into the pathology of DNMT3A mutation-driven leukemia. In this study, we have used a conditional knockout mouse model to study the role of Dnmt3a in normal hematopoiesis. Hematopoietic stem cells (HSCs) from Mx1-Cre:Dnmt3afl/fl mice were serially transplanted into lethally irradiated recipient mice to study the effect of loss of Dnmt3a on HSC self-renewal and differentiation. We show that loss of Dnmt3a progressively impedes HSC differentiation over four-rounds of serial transplantation, while simultaneously expanding HSC numbers in the bone marrow. Examination of the bone marrow post-transplant revealed that control HSCs showed a gradual decline in their ability to regenerate the HSC pool at each successive round of transplantation, while in contrast Dnmt3a-KO HSCs show a remarkably robust capacity for amplification, generating 40,000 – 100,000 HSCs per mouse. Quantification of peripheral blood differentiation on a per HSC basis demonstrated in the absence of Dnmt3a, a cell division is more likely to result in a self-renewal rather than differentiation fate (Figure 1). Using semi-global reduced representation bisulfite sequencing (RRBS), we show that Dnmt3a-KO HSCs manifest both increased and decreased methylation at distinct loci, including dramatic CpG island hypermethylation. Global transcriptional analysis by microarray revealed that Dnmt3a-KO HSCs show upregulation of HSC multipotency genes coupled with simultaneous downregulation of early differentiation factors (e.g. Flt3, PU.1, Mef2c), likely inhibiting the initial stages of HSC differentiation. Upregulation of key HSC regulators including Runx1, Gata3 and Nr4a2 was associated with gene-body hypomethylation and activated chromatin marks (H3K4me3) in Dnmt3a-KO HSCs. Finally, we show that Dnmt3a-KO HSCs are unable to methylate and transcriptionally repress these key HSC multipotency genes in response to chemotherapeutic ablation of the hematopoietic system, leading to inefficient differentiation and manifesting hypomethylation and incomplete repression of HSC-specific genes in their limited differentiated progeny. In conclusion, we show that Dnmt3a plays a specific role in permitting HSC differentiation, as in its absence, phenotypically normal but impotent stem cells accumulate and differentiation capacity is progressively lost. This differentiation-deficit phenotype is reminiscent of Dnmt3a/Dnmt3b-null embryonic stem (ES) cells while markedly distinct from that of Dnmt1-KO HSCs which show premature HSC exhaustion and lymphoid-deficient differentiation, demonstrating distinct roles for the different DNA methyltransferase enzymes in HSCs. In light of the recently-identified DNMT3A mutations in AML and MDS patients, these studies are the first biological models linking mutation of Dnmt3a with inhibition of HSC differentiation which may be one of the first pathogenic steps occuring in such patients.Figure 1Dnmt3a-KO HSCs become biased towards self-renewal as opposed to differentiation. At each transplant round, the self-renewal quotient was calculated as the number of donor-derived HSCs recovered at the end of the transplant divided by 250 (the number of HSC initially transplanted). The differentiation quotient was calculated as (the white blood cell count per μl of blood at 16 weeks) X (percentage of donor-cell chimerism)/number of donor HSC at the end of the transplant. Over serial transfer, Dnmt3a-KO HSCs more rapidly lose their differentiation capacity compared to control HSCs, while sustaining robust self-renewal.Figure 1. Dnmt3a-KO HSCs become biased towards self-renewal as opposed to differentiation. At each transplant round, the self-renewal quotient was calculated as the number of donor-derived HSCs recovered at the end of the transplant divided by 250 (the number of HSC initially transplanted). The differentiation quotient was calculated as (the white blood cell count per μl of blood at 16 weeks) X (percentage of donor-cell chimerism)/number of donor HSC at the end of the transplant. Over serial transfer, Dnmt3a-KO HSCs more rapidly lose their differentiation capacity compared to control HSCs, while sustaining robust self-renewal. Disclosures: Issa: Novartis: Honoraria; GSK: Consultancy; SYNDAX: Consultancy; Merck: Research Funding; Eisai: Research Funding; Celgene: Research Funding; Celgene: Honoraria; J&J: Honoraria.