ALBERT

Description: Homeobox a1 (Hoxa1) is one of the most rapidly induced genes in ES cell differentiation and it is the earliest expressed Hox gene in the mouse embryo. In this study, we used genomic approaches to identify Hoxa1-bound regions during early stages of ES cell differentiation into the neuro-ectoderm. Within 2 h of retinoic acid treatment, Hoxa1 is rapidly recruited to target sites that are associated with genes involved in regulation of pluripotency, and these genes display early changes in expression. The pattern of occupancy of Hoxa1 is dynamic and changes over time. At 12 h of differentiation, many sites bound at 2 h are lost and a new cohort of bound regions appears. At both time points the genome-wide mapping reveals that there is significant co-occupancy of Nanog (Nanog homeobox) and Hoxa1 on many common target sites, and these are linked to genes in the pluripotential regulatory network. In addition to shared target genes, Hoxa1 binds to regulatory regions of Nanog, and conversely Nanog binds to a 3′ enhancer of Hoxa1. This finding provides evidence for direct cross-regulatory feedback between Hoxa1 and Nanog through a mechanism of mutual repression. Hoxa1 also binds to regulatory regions of Sox2 (sex-determining region Y box 2), Esrrb (estrogen-related receptor beta), and Myc, which underscores its key input into core components of the pluripotential regulatory network. We propose a model whereby direct inputs of Nanog and Hoxa1 on shared targets and mutual repression between Hoxa1 and the core pluripotency network provides a molecular mechanism that modulates the fine balance between the alternate states of pluripotency and differentiation.

Description: Hematopoietic stem cells (HSCs) sustain lifelong production of multiple blood cell types through a finely-tuned balance between stem cell maintenance and activation to prevent bone marrow exhaustion or overgrowth. The highly conserved Hox family of homeodomain containing transcription factors have been identified as key regulators and contributors in both normal hematopoiesis and leukemogenesis. Most previous work has focused on individual Hox genes; however, it remains largely unknown whether and how multiple Hox genes in a cluster are regulated and function in hematopoiesis. We initiated a study to perform systematic, high-throughput transcriptome analysis in the following 17 cell types from the bone marrow (BM) of C57BL/6J mice: 4 hematopoietic stem and progenitor cells (CD49blo long-term (LT)-HSC, CD49bhi intermediate-term (IT)-HSC, short-term (ST)-HSC, and MPP); and 4 committed progenitors (CLP, CMP, GMP and MEP); and 9 mature lineage cells (B cell, T cell, NK cell, dendritic cell, monocyte, macrophage, granulocyte, megakaryocyte and nucleated erythrocyte). Intriguingly, as part of a unique fingerprint observed in the most primitive CD49blo LT-HSCs, we detected expression from the Hoxb cluster. Further analysis on all the four Hox clusters revealed that most of the genes from the Hoxb cluster, and not from the other Hox clusters, were predominantly expressed in the CD49blo LT-HSCs. This suggests that they might function as a cluster to maintain CD49blo LT-HSCs. A previous study has shown that one cis -regulatory retinoic acid responsive element (RARE), is conserved among vertebrate species and regulates multiple Hoxb gene expression in central nervous system development. Thus, we asked whether RARE is essential for maintenance of primitive CD49blo LT-HSCs by regulation of Hoxb cluster. To test this hypothesis, we utilized a RAREΔ knockout mouse model and assayed for HSC numbers in BM. We observed that homozygous deletion of RARE led to 2-fold reduction in both the frequency and absolute number of CD49blo LT-HSCs. Functionally, we first conducted limiting dilution, competitive repopulating unit (CRU) assays by transplanting 2.5×104, 7.5×104 or 2×105 of BM cells from RAREΔ mutants and their control littermates, together with 2×105 recipient BM cells derived from the Ptprc mutant strain, into lethally irradiated recipient mice. Our data showed a 2.5-fold decrease in functional HSCs in RAREΔ HSCs (1/20,326) compared to control (1/50,839). To further evaluate the long-term effect of RARE on HSCs, we performed serial BM transplantation and observed a 12.9-fold reduction of reconstitution ability after secondary transplantation. These data indicate that deletion of RARE compromised HSC long-term reconstitution capacity. Collectively, our work provides evidence showing that RARE is essential for maintenance of the primitive HSCs by regulation of Hoxb cluster genes. Disclosures No relevant conflicts of interest to declare.

Description: The balance between self-renewal and multilineage differentiation in hematopoietic stem cells (HSCs) is orchestrated by genetic regulatory networks, which are constituted by hundreds of thousands of cis-regulatory elements, such as promoters, enhancers, insulators, etc. Aberrant mutations in these cis-acting modules and their trans-acting factors have been frequently found in hematopoietic malignancies including leukemia. Although next-generation sequencing technologies have identified comprehensive maps of these modules, much remains unknown about their physiological functions and underlying mechanisms in HSC maintenance and leukemogenesis. Our transcriptome analysis in 17 murine hematopoietic cell types, including HSCs, committed progenitors and lineage cells, showed that most of HoxB cluster genes were predominantly enriched in the permanently reconstituting long-term (LT) HSCs. Interestingly, one of the two putative enhancers within HoxB cluster, identified by H3K27ac ChIP-Seq analysis, shared the same sequence as the retinoic acid responsive element DERARE, which was recently reported to regulate multiple HoxB gene expression in the central nervous system. To test whether DERARE is required for normal hematopoiesis, we utilized the DERARE knockout mouse and found that homozygous deletion of DERARE led to 2-fold reduction in both the frequency and absolute number of LT-HSCs. Functionally, limiting dilution, competitive repopulating unit assays showed a 2.5-fold decrease in functional HSCs of DERAREΔ mice compared to wildtype control. We further performed serial transplantation and observed a 4.3-fold reduction of repopulation rate after secondary transplantation of DERAREΔ HSCs, indicating long-term reconstitution capacity was impaired. Mechanistically, RNA-Seq in LT-HSCs from DERAREΔ mice exhibited significantly enriched apoptosis pathway in DERAREΔ compared with that from Wt control. Furthermore, DNA methylation analysis showed gradually gained methylation on DERARE during HSC differentiation, which is negatively correlated with HoxB cluster gene expression, suggesting DERARE might be a methylation-sensitive enhancer to control HoxB genes. Moreover, HoxB gene expression is markedly upregulated in Dnmt3a KO HSCs, indicating that Dnmt3a is responsible for the DERARE methylation in HSCs. Finally, the analysis of clinical data from acute myeloid leukemia 200 patients in the Cancer Genome Atlas project revealed that lowly methylated DERARE was significantly correlated with overexpression of HoxB genes, high cytogenetic risk, and poor prognosis, suggesting abnormal regulation of DERARE contributes to leukemogenesis. Collectively, our study demonstrates the essential roles of the cis-regulatory element DERARE in both maintenance of LT-HSCs and contribution to leukemogenesis through regulation of HoxB cluster genes. Disclosures No relevant conflicts of interest to declare.

Description: A population of reserve HSCs has been reported to be in a deep-quiescent (or dormant) state and function as a back-up population of HSCs to support life-time hematopoiesis (Li and Clevers, 2010; Wilson et al., 2008). Currently, characterization of reserve HSCs is mainly based on cell cycle quiescence, however, the metabolic state in reserve HSCs is yet to be defined. Here, we show that reserve HSCs maintain not only a quiescent state but also an overall low metabolic activity whereas the primed HSCs maintain still a quiescent state but primed at metabolic state. First, we used CD49b(Benveniste et al., 2010) to further separate conventional long-term (LT) HSCs (CD34-Flk2-Lineage-Sca-1+c-Kit+)(Yang et al., 2005) into CD49blo and CD49bhi subpopulations and confirmed their enrichment with previously identified dormant (Scl-H2B-GFP label retaining cells, LRCs), thus we termed CD49blo and CD49bhi subpopulations as candidates of reserve and primed HSCs. We then determined the cell cycle frequencies of reserve, primed, and ST-HSCs (Cd34+Flk2-LSK) respectively as once in 3 months, 3-4 weeks, and 3-4 days. Functionally, Reserve HSCs had on average 3.5-fold higher functional capacity compared with primed HSCs. RNA-seq analysis revealed that reserve HSC predominantly expressed a list of imprinting genes that associate with growth-restriction functions; primed HSCs expressed relatively-high number of genes involving in mitochondria fusion, organization, and function, while ST-HSCs expressed genes reflecting an active cycling state. Conducting metabolic assays, we found that reserve HSCs not only maintain quiescence but also maintain overall low metabolic activity in both glycolysis and mitochondrial capacity. In contrast, primed HSCs are in a quiescent state, but with their metabolic state primed as evidenced by an increased glycolytic activity and mitochondrial potential for subsequent active proliferating state in ST-HSCs. Intriguingly, our data suggests that the functionality of reserve HSCs correlates to metabolic state rather than cell cycle. Disclosures No relevant conflicts of interest to declare.

Description: Balanced regulation is essential for the long-term preservation of stem cells while providing for ongoing tissue maintenance. We and others have previously shown that these dichotomous functions are accomplished through the co-existence of at least two stem cell populations—reserve and primed stem cells. This balance has been shown previously to be regulated by protein-coding genes; however, the potential roles of noncoding RNAs (ncRNAs) and their relationships with protein-coding genes in regulating hematopoietic stem cells (HSCs) remain largely unknown. To systematically identify ncRNAs involved in the murine hematopoiesis, we used RNA sequencing and identified unique, differentially expressed (fingerprint) ncRNAs representing reserve HSCs, primed HSCs, and more active stem/progenitor cells. These were also compared with committed progenitors and all major mature hematopoietic lineages. Intriguingly, all of the fingerprint ncRNAs uniquely expressed in reserve HSCs were derived from the imprinted Dlk1-Gtl2 locus, which spans a 780kb region on the mouse chromosome 12qF1 and is precisely controlled by the Intergenic Germ line-derived Differentially Methylated Region (IG-DMR). The Gtl2 locus contains a large cluster of snoRNAs (23 snoRNAs) and the largest cluster of mammalian miRNAs (57 miRNAs) as part of a single transcript of long length ncRNA downstream of Gtl2. To determine the role of Dlk1-Gtl2 locus in hematopoiesis, we utilized the IG-DMR knockout mouse model and carried out phenotypic and functional assays in E15.0 fetal liver HSCs since the embryos loss of maternal IG-DMR are lethal after E16. We observed that deletion of the maternal IG-DMR (ΔmIG-DMR), but not the paternal one, leads to 2-fold reduction in CD93+ fetal liver HSC number and 4-fold decrease of reconstitution ability after tertiary transplantation. Further, we employed RNA-seq using fetal liver HSCs from wt and ΔmIG-DMR and found that several pathways involved in growth control, mitochondrial function and energy metabolism, such as mTOR, PI3K/Akt and Wnt, are significantly enhanced in ΔmIG-DMR HSCs. We also carried out small RNA-seq in both adult HSCs and fetal liver HSCs and identified 13 HSC-specific miRNAs, which are predominantly expressed in reserve HSCs and predicted to target multiple proteins in PI3K/Akt/mTOR pathway. Mechanistically, maternal IG-DMR deletion leads to down-regulation of Gtl2-derived 13 miRNAs and hyperactivation of PI3K/Akt/mTOR pathway, which further enhances mitochondrial activity and biogenesis, increases oxidative phosphorylation (OXPHOS) mediated ATP production and ROS levels, and eventually causes HSC exhaustion. Moreover, either pharmacological inhibition of the mTOR activity by rapamycin or overexpression of Gtl2-derived miRNAs could partially, if not all, rescues the defective HSC phenotype and bioenergetic activities caused by mIG-DMR deletion. Collectively, our work provides a global landscape of murine hematopoietic lncRNAs and demonstrates that Dlk1-Gtl2 locus is critical in maintaining primitive HSCs with a fundamentally epigenetic regulation of mitochondrial function, energy metabolism via repression of PI3K/Akt/mTOR pathway. Disclosures No relevant conflicts of interest to declare.

Description: Within the bone marrow, three hematopoietic stem cell niches have been identified; the osteoblastic niche (Arai et al., 2004; Calvi et al., 2003; Nilsson et al., 2001; Zhang et al., 2003), the vascular niche (Kiel et al., 2005), and the CAR cell niche (Sugiyama et al., 2006). The adhesion molecule N-cadherin has been found associated with the osteoblastic and CAR cell niches, implicating N-cadherin’s function for hematopoietic stem cell (HSC) anchoring and signaling (Arai et al., 2004; Muguruma et al., 2006; Zhang et al., 2003). However, as of yet, a HSC population expressing N-cadherin has not been fully characterized. Therefore, we examined how N-cadherin expression in HSCs relates to their function and lifecycle. Unexpectedly, we found that doses of 5000 bone marrow cells expressing the highest level of N-cadherin (N-cadherinhi) failed to reconstitute hematopoietic lineages in irradiated recipient mice. An explanation for this engraftment failure came with detailed cell surface phenotyping which revealed that these N-cadherinhi cells were primarily Lineage+ and devoid of the characteristic hematopoietic stem pool, Lineage-Sca+cKit+ (LSK) cells. Instead, we found that Flk2-LSK HSCs express a gradient of N-cadherin which could be described as low (N-cadherinlo) to intermediate (N-cadherinint) levels (Figure1). FACS applications were used to isolate pure populations of these N-cadherinlo and N-cadherinint Flk2-LSK HSCs. Real time RT-PCR (N-cadherin primers crossed the intron between Exon 2 and 3). (Figure 2), microarray studies, and competitive reconstitution transplantation assays revealed that this N-cadherin division of Flk2- LSK HSCs formed two populations with very distinct properties. In transplantation assays the N-cadherinlo population more robustly reconstituted the hematopoietic system. Principle Component Analysis and gene ontology analysis of microarray data revealed that the N-cadherinlo cells expressed genes that may prime them to respond to signals and to mobilize. This data was confirmed with mobilization studies which showed that HSCs mobilized from bone marrow to spleen were predominantly N-cadherinlo. In contrast, the expression profile of N-cadherinint cells suggests they are more ‘reserved’ and this population was also maintained with HSCs spared by 5-fluouracil (5FU) treatment. Our results suggest that differential N-cadherin expression reflects important functional distinctions between HSC subpopulations. N-cadherinlo HSCs, with their robust reconstitution efficiency and properties similar to mobilized HSCs, may have clinical relevance. Figure Figure Figure Figure