Les leucémies aiguës lymphoblastiques de la lignée B (LAL-B) sont les cancers pédiatriques les plus fréquents. Dans ce type de leucémie, l'une des anomalies génétiques les plus fréquentes est la translocation t(12 ;21) aboutissant à la protéine de fusion ETV6-RUNX1. Cette pathologie est décrite comme un modèle à deux « hits ». Le premier, se produit in utero et génère la protéine de fusion. Le second, correspond à l’acquisition d’anomalies génétiques après la naissance. Ces réarrangements génomiques aberrants ont été décrits comme provenant d’une activité anormale de la recombinasse RAG. Notre travail a consisté dans un premier temps à compléter le modèle de leucémogénèse à plusieurs « hits ». En continuant notre étude des LAL B à translocation ETV6-RUNX1, nous nous sommes concentrés sur le rôle de RUNX1, gène dérégulé dans ce type de leucémie.L’ensemble de nos résultats confirme le rôle prépondérant de RUNX1 dans l’hématopoïèse et la leucémogenèse grâce à sa capacité à s’associer à des protéines aux fonctions différentes et grâce à son implication dans la transcription de gènes clé en hématologie. Nos résultats ouvrent donc de nouvelles perspectives dans la compréhension du contrôle de l’activité transcriptionnelle de RUNX1 et dans son rôle dans les hémopathies malignes. / B-cell precursor acute lymphoblastic leukemia (B-ALL) is the most common pediatric cancer. In this type of leukemia, one of the most common genetic abnormalities is the ETV6-RUNX1 rearrangement. This malignancy is described as a two "hits" model. The first event occurs mainly in utero and generates the fusion gene ETV6-RUNX1. The second event consists in the acquisition of additional genetic abnormalities after birth. These aberrant genomic modifications have been described as resulting from abnormal activity of the RAG recombinase. Our work consisted initially in completing the leukemogenesis model. In continuing our study of ETV6-RUNX1 B-ALL, we focused on the role of RUNX1, an upregulated gene in this type of leukemia. All results confirm the predominant role of RUNX1 in hematopoiesis and leukemogenesis thanks to its ability to associate with proteins with different functions and its involvement in the transcription of key genes in hematology. Our results therefore open new perspectives in understanding the control of transcriptional activity of RUNX1 and its role in malignant hematology.

The first transplantable hematopoietic stem cells (HSC) arise in the aorta-gonad mesonephros region (AGM) during early stages of embryo development. Specifically, ventral aspect of embryonic dorsal aorta (DA) contains HSC that upon transplantation into irradiated recipients can reconstitute all lineages of the haematopoietic system [Medvinsky et al. 1993; Muller and Medvinsky, 1994; Medvinsky and Dzierzak, 1996; Cumano et al., 1996; Tavian et al., 1996; Peault and Tavian, 2003; Taoudi and Medvinsky, 2007; Ivanovs et al., 2011, 2014]. The ventral aspect of DA bears so-called intra-aortic cell clusters (IAC), the appearance of which coincides with the emergence of HSC [Babovic and Eaves, 2014; Bhatia, 2007; Boisset et al., 2010, 2011; Bollerot et al., 2005; de Bruijin et al., 2002; Bertrand et al., 2010]. According to recent reports, HSC are a heterogeneous population of cells [Dykstra et al., 2007; Seita and Weissman, 2010; Muller-Sieburg et al., 2012]. It is unclear whether all HSC precursors originate from the same location, for example, DA lining, IAC or sub-aortic tissues; or HSC precursors migrate into DA lining from other parts of the embryo [Tavian et al., 1999; Yoder et al., 1997; Oberlin et al., 2002; Peault and Tavian, 2003; Dzierzak, 2003; Samokhvalov et al., 2007; Medvinsky et al., 2011]. To elucidate ontogeny of early HSC precursors (pro-HSC), two approaches were applied in this PhD project. First, we mapped potential pro-HSC in pre-circulation mouse embryos (embryonic day 6-8.5, E6-E8.5). We defined potential pro-HSC as cells co-expressing the transcription factor Runx1, endothelial markers (VE-Cad or CD31) and/or haematopoietic markers (CD45, CD41) [Oberlin et al., 2002; de Bruijn and Dzierzak, 2012; Liakhovitskaia et al., 2009, 2014]. In E6-E8 mouse embryo, prospective pro-HSC were found to be located in chorionic plate, yolk sac and in allantoic core domain. In early somitic mouse embryo (E8-8.5) cells with pro-HSC phenotype (Runx1+CD31+CD41+) were found to be in cell clusters in forming vessel of confluence and in nascent dorsal aortae lining. Pro-HSC are not directly transplantable [Cumano et al., 1996., 2001; Godin et al., 1993; 1995; Batta et al., 2016; Matsuoka et al., 2001; Nishikawa et al., 1998]. Therefore, cells and tissues containing prospective pro-HSC were initially matured using several in-vitro culture systems. According to our results, E8 mouse embryo pro-HSC are only preserved in explant cultures, but not in co-aggregate cultures with stroma cells. After culture, cells were transplanted into sub-lethally irradiated recipients. Six weeks after transplantation 19 out of 82 transplanted recipients had donor derived blood cells' chimerism at the level of 0.1-0.3%. Forty six percent of these grafts were derived from rostral part of the embryo tissues (head, heart, upper somites). Only one out of 82 recipients had donor cells contribution above 1% (1.2 %). This recipient was engrafted with cells derived from the E8 mouse embryo head and heart region. Recipients having blood chimerism at the range of 0.1-0.3% had mainly lymphoid donor derived cells in their peripheral blood. The only recipient showing the high donor cells contribution (1.2%) had contribution mainly to myeloid lineage. Recorded low levels of blood chimersims are in line with those reported by Rybtsov et al. (2014) for early E9 mouse embryos. Donor derived cells formed clearly distinguishable populations on cytometry plots. This population of cells were absent from control engraftment experiments with carrier cells only. Previously, lymphoid potential was detected in paraaortic spnanchnopleura (P-Sp) of E8.5-9 mouse embryos, but not in E8 mouse embryos (0-5 somites, pre-circulation) and later in yolk sac [Cumano et al., 1996; Nishikawa et al., 1998; Fraser et al., 2002; Yokota et al., 2006]. However, prior works used different criteria to establish recipient reconstitution. Therefore, it is possible that recipients repopulated with E8 derived cells at the level of 0.1% were not considered as repopulated and hence, presence of lymphoid lineage precursors was overlooked in early somitic mouse embryos. The only recipient showing substantial myeloid cells contribution (73% Mac1+Gr1+ cells of donor derived cells) received engrafted cells from an older (6-13 sp) embryo and therefore potentially has yolk sac derived myeloid cells. Yolk sac cell contribution to myeloid lineage, specifically to the brain microglia was reported in prior works [Samokhvalov et al., 2007]. Our data show that early E8 AGM cells do not expand in in vitro conditions. While in AGM, cells from E9 mouse embryo expand in culture [Rybtsov et al., 2014]. We have analysed Runx1 expression pattern and dorsal aorta morphology at the time when E9 HSC precursors acquire ability to expand in in vitro culture. Runx1 expression becomes clearly polarised at the time point (22-26 sp), when paired dorsal aortae fusion is initiated. We envision that intimate connection between DA fusion events and induction of pro-HSC maturation exists. According to prior reports, Bmp, Shh and VEGF signalling regulate DA fusion [Garriock et al., 2010]. Thereofore, to enhance in vitro HSC maturation system, DA fusion triggers (for example, Bmp4) might be added to culture. Since, pro-HSC maturation methods established to date are not efficient to expand and differentiate E8 pro-HSC into potent HSC, another approach had to be implemented to study HSC ontogeny. The second approach we utilized was to trace the origin of HSC in chicken embryo, starting from the very beginning of cell fate specification, i.e. from gastrulation stages. Chick embryo haematopoiesis is similar in both human and mouse: precursors of HSC arise in the embryo proper in AGM, and IAC are formed in DA ventral aspect [Dieterlen-Lièvre, 1975; Dieterlen-Lièvre and Martin, 1981; Dieterlen-Lièvre and Jaffredo, 2009; Jaffredo et al., 2000; Le Douarin and Dieterlen-Lièvre, 2013]. In contrast to mammals, chick embryo develops ex vivo, making direct labelling and cell tracing possible. We aimed to identify cells giving rise to regions of DA that produce IAC. Therefore, segments of primitive streak (PS) were labelled with lipophilic dyes or by substituting segments of host PS with PS sections derived from transgenic (GFP+) stage matched chicken embryos. Our results show that in an 18-25h chicken embryo (Hamburger and Hamilton developmental stage 4-6, HH4-6) cells giving rise to DA ingress through the wide region of PS (35-60% of its length) [Hamburger and Hamilton, 1951]. We identified that the section of DA producing HSC is formed by cells ingressing through PS in region of 40-55% of its length at 18-25h of chick embryo development. Regardless of the embryo development stage (HH4-6), in chimeras grafted at 40-55% of PS length, GFP+ cells contributed to DA and to the IAC. Within GFP+ labelled areas, we observed clusters consisting entirely of GFP+ and clusters having a mixture of GFP+ and GFP- cells. Entirely GFP+ clusters were found in the stretch of DA that had the entire aortic endothelial lining labelled. Clusters formed on the mosaic (GFP+/GFP-) aortic endothelium also had mosaic nature. According to our data, multiple descendants of PS contribute to the same stretch of dorsal aorta. This explains mosaicity of dorsal aorta lining and IAC labelling. Since we encountered clusters with mixture of GFP+ and GFP- cells, we conclude that IAC are not clonal formations. Mosaicity of IAC also does not exclude a scenario when cells migrate in and out of a cluster. Further tracing experiments are required to establish HSC nature of cells within a cluster.

Haematopoietic stem and progenitor cells (HSPCs) are de novo generated within in the ventral aspects of the embryonic dorsal aorta (DA). Cells of this haemogenic endothelium (HE) will eventually undergo an endothelial to haematopoietic transition (EHT) that involves cell budding out of the aortic wall. Despite the detailed description of the cellular events, the exact haemogenic lineage path and the underlying molecular mechanism that establish full haematopoietic competence are still not entirely understood. The transcription factor Runx1 is critical for the emergence of HSPCs and shows expression in the zebrafish HE as early as 24 hpf. To facilitate a detailed analysis of the transient HE population I generated a TgBAC(runx1P2:Citrine) reporter line under the control of the endogenous runx1 promoter on a bacterial artificial chromosome (BAC). Double-transgenic reporter lines for runx1 and the endothelial marker kdrl allowed us to isolate specifically cells of the DA away from the whole endothelial population, which could be further sub-divided into HE and non-haemogenic cells. Genomewide expression analysis within the respective tissues and upon Runx1 loss of function enabled the identification of HE-specific Runx1-regulated genes. Hereby, the gfi1ab gene appeared as the functional homologue of the murine Gfi1. I show that in zebrafish, EHT is orchestrated through a conserved Runx1-Gfi1-Lsd1 axis. The cellular functions of the remaining Runx1 targets imply that maturation into fully functional HSCs depends on epigenetic regulation due to the up-regulation of de novo DNAmethyltransferases, as well as on factors that allow the developing HSCs to respond to extrinsic cues from haematopoietic niches. Lastly, it became evident that the early HE expresses dll4 at similar levels to the rest of the aortic endothelium, indicating a common lineage path. In the absence of RUNX1 the HE remains essentially arterial and persists as an integrated part of the DA.

Runx1 is a central transcription factor in the development of the murine haematopoietic system and in the emergence and specification of its main key component, the haematopoietic stem cell. Previous studies suggested a requirement for Runx1 in a window of time stretching from mesoderm specification (E6.5) to mid-gestation (E11), but these studies did not investigate each primary haematopoietic site separately. During this PhD project, a Runx1 reactivatable knockout mouse model was used to study the impact of the absence of Runx1 from E9.5 to E11 in primary haematopoietic sites on early precursor populations, especially PreHSC Type I and II. At E9.5, the KO conceptus was already developmentally retarded, lacking progenitors and PreHSC Type II but was not devoid of PreHSC Type I, as demonstrated by flow cytometry, thus suggesting a requirement for Runx1 in the transition from PreHSC Type I to PreHSC Type II stage. Using a novel culture system that enables the potent in vitro maturation of precursors of HSCs into fully mature adult-repopulating HSCs, it was found that maturation of PreHSC Type I into HSCs was hindered in KO tissues, despite the expression of Runx1 in OP9 niche compartment, thus pointing towards a cell autonomous requirement for Runx1. In this model, the Runx1 allele was subsequently reactivated to a functional state by tamoxifen-induced Cre-mediated recombination (CreERT2 system). Tamoxifen / Cre toxicity on HSC maturation was evaluated during AGM reaggregate culture to achieve the best balance between the highest recombination levels and the lowest toxicity when Cre was induced in cell suspension prior to reaggregation. It was found that reactivation of Runx1 at E9.5 in primary haematopoietic sites was not sufficient to rescue haematopoietic development, thus suggesting a requirement for Runx1 before E9.5.

Tissue-resident macrophages (MΦ) such as microglia, Kupffer and Langerhans cells derive from Myb-independent yolk sac (YS) progenitors generated before the emergence of hematopoietic stem cells (HSCs). Myb-independent YS-derived resident MΦ self-renew locally, independently of circulating adult monocytes and HSCs. In contrast, adult blood monocytes as well as infiltrating, gut and dermal MΦ derive from Myb-dependent HSCs and are less proliferative. These findings are derived from the mouse, using gene knock-outs and lineage tracing, but their applicability to human development has not been formally demonstrated. Here I use a human pluripotent stem cell (hPSC) differentiation model of hematopoiesis, capable of monocyte/MΦ production over prolonged periods of time, as a tool to investigate human mononuclear phagocyte ontogeny. Using a transcriptomic approach I showed that hiPSC-derived monocytes/MΦ (iPS-Mo/MΦ) produced early in differentiation (first weeks) are more proliferative and less immunologically mature than iPS-Mo/MΦ produced at a later time point. I therefore hypothesised either that iPS-Mo/MΦ only become fully mature after several weeks of differentiation or that there are two developmentally distinct waves of MΦ produced over time. By comparing the transcription profile of iPS-Mo/MΦs to that of primary adult blood monocytes and fetal microglia I then showed that early and late iPS-Mo/MΦs were transcriptionally closer to fetal microglia than to adult blood monocytes. To further investigate if iPS-Mo/MΦs are indeed of the same developmental origin as MYB-independent MΦ such as microglia, I used a CRISPR-Cas9 knock-out strategy to show for the first time, that human iPS-Mo/MΦs develop in a MYB-independent, RUNX1 and SPI1 (PU.1)-dependent fashion. This result makes human iPS-Mo/MΦs developmentally related to, and a good model for, MYB-independent tissue-resident \Macros such as alveolar and kidney MΦs, microglia, Kupffer and Langerhans cells. Interestingly, while MYB was not required for the generation of iPS-Mo/MΦs, its knock-out resulted in an increase in iPS-Mo/MΦ production. To investigate this increase I developed two methods for quantifying the differentiation bottleneck occurring during hiPSC differentiation to iPS-Mo/MΦs. Those techniques highlighted a potential increase in progenitor cell generation in MYB KO cells and thus lay foundation to improve our technical understanding of EB differentiation and will enable enhanced manipulation of the EB model.

No description available.

京都大学 / 新制・論文博士 / 博士(医学) / 乙第13480号 / 論医博第2255号 / 新制||医||1059(附属図書館) / 京都大学大学院医学研究科医学専攻 / (主査)教授 小川 誠司, 教授 羽賀 博典, 教授 伊藤 貴浩 / 学位規則第4条第2項該当 / Doctor of Medical Science / Kyoto University / DFAM

Breast cancer arises from a series of acquired mutations that disrupt normal mammary epithelial homeostasis and create multi-potent cancer stem cells that can differentiate into clinically distinct breast cancer subtypes. Despite improved therapies and advances in early detection, breast cancer remains the leading diagnosed cancer in women. A predominant mechanism initiating invasion and migration for a variety of cancers including breast, is epithelial-to-mesenchymal transition (EMT). EMT— a trans-differentiation process through which mammary epithelial cells acquire a more aggressive mesenchymal phenotype—is a regulated process during early mammary gland development and involves many transcription factors involved in cell lineage commitment, proliferation, and growth. Despite accumulating evidence for a broad understanding of EMT regulation, the mechanism(s) by which mammary epithelial cells maintain their phenotype is unknown. Mitotic gene bookmarking, i.e., transcription factor binding to target genes during mitosis for post mitotic regulation, is a key epigenetic mechanism to convey regulatory information for cell proliferation, growth, and identity through successive cell divisions. Many phenotypic transcription factors, including the hematopoietic Runt Related Transcription Factor 1 (RUNX1/AML1), bookmark target genes during mitosis. Despite growing evidence, a role for mitotic gene bookmarking in maintaining mammary epithelial phenotype has not been investigated. RUNX1 has been recently identified to play key roles in breast cancer development and progression. Importantly, RUNX1 stabilizes the normal breast epithelial phenotype and prevents EMT through repression of EMT-initiating pathways. Findings reported in this thesis demonstrate that RUNX1 mitotically bookmarks both RNA Pol I and II transcribed genes involved in proliferation, growth, and mammary epithelial phenotype maintenance. Inhibition of RUNX1 DNA binding by a specific small molecule inhibitor led to phenotypic changes, apoptosis, differences in global protein synthesis, and differential expression of ribosomal RNA as well as protein coding genes and long non-coding RNA genes involved in cellular phenotype. Together these findings reveal a novel epigenetic regulatory role of RUNX1 in normal-like breast epithelial cells and strongly suggest that mitotic bookmarking of target genes by RUNX1 is required to maintain breast epithelial phenotype. Disruption of RUNX1 bookmarking results in initiation of epithelial to mesenchymal transition, an essential first step in the onset of breast cancer.

Somatosensory neurons are essential for detecting diverse environmental stimuli, thus critical for survival of mammals. In order to achieve sensory modality specificity, many somatosensory subtypes emerge with various receptor and ion channel expression, as well as terminal morphologies. How the somatosensory system achieves such a high variety of neuronal subtypes is unknown. In this thesis, I used a newly discovered subtype, VGLUT3-expressing unmyelinated low-threshold mechanoreceptors (C-LTMRs), as a model to try to answer this question. C-LTMRs have been proposed to play a role in pleasant touch in humans or pain in mice. Previously, our lab has identified the Runt domain transcriptional factor Runx1 to be pivotal for the development of a cohort of sensory neurons such as pain related nociceptors, thermal receptors, as well as itch related pruriceptors. Here I found that Runx1 is also required to establish all known features associated with C-LTMRs. In search of the mechanism of how Runx1 controls C-LTMR development, I found that the zinc finger protein Zfp521 is predominantly expressed in C-LTMRs and its expression is Runx1 dependent. By generating and analyzing Zfp521 conditional knock out animals, I found Zfp521 is required for part of C-LTMR molecular identities and nerve terminal morphologies. Our studies suggest that Runx1 acts through Zfp521-dependent and Zfp521-independent pathways to specify C-LTMR identities. To study C-LTMR functions, we performed a series of behavioral analysis and found the loss of VGLUT3 and mechanosensitivities in C-LTMRs does not markedly affect acute or chronic mechanical pain measured from the hind paws, which argues against the proposed role of VGLUT3 in C-LTMRs in mediating mechanical pain in mice. In the future, we will continue to use our mutant mice to study physiological functions of C-LTMRs.

Runx1/AML1 transcription factor expression in hematopoietic cell lineages is differentially regulated via usage of two distinct promoters. The 5' UTR and a 19 amino acid encoding sequence transcribed from the distal promoter is inserted via alternative splicing into the 5' end of the mRNA transcript, replacing the 5' UTR and a 5 amino acid encoding sequence usually transcribed from the proximal promoter. Expression of proximal Runx1 in 32Dcl.3 cells delays G-CSF induced neutrophil terminal differentiation by increasing viability compared to distal Runx1. We utilized Runx1 Nterminal deletion and point mutants of three evolutionarily conserved residues to describe dual N-terminal isoform motifs that promote two distinct differentiation phenotypes as regulatory elements in hematopoietic cell differentiation. Runx1 isoforms were evaluated in established hematopoietic in vitro and ex vivo differentiation systems. Deletion of amino acids 3’-14’ (Δ3-14) or 3’-19’ (Δ3-19) of the distal Runx1 N-terminus delayed terminal differentiation of the 32Dcl.3 myeloid cell line, indicating a regulatory motif in distal Runx1 abrogates the delay of terminal differentiation induced by proximal Runx1. Deletion of amino acids 3’-8’ (Δ3-8) or mutation of amino acids serine 3’, serine 5’ and phenylalanine 7’ of the distal Runx1 N-terminus reduce Runx1 expression in the 32Dcl.3 cell line. The N-terminus motif, runt domain and nuclear matrix-targeting sequence of Runx1 modulated Ets1 activity on the KIR3DL1 bidirectional promoter element. The transcription factor YY1 promotes both forward and reverse activation of the KIR3DL1 bidirectional promoter element dominantly in the presence of Runx1, and additively with Ets1. Distinct Runx1 proximal and distal N-termini induced phenotypes were observed in myeloid and thymocyte differentiation, but not with the KIR3DL1 luciferase assay system. This work identifies a previously unknown N-terminal regulatory motif that acts with spatio-temporal and gene target specificity to add another level of control over Runx1 activity during hematopoiesis.