Anna Bigas, Constanze Bonifer, Marella de Bruijn, Andrea Ditadi, Charles Durand, Andrew Elefanty, Thierry Jaffredo (organiser), Valérie Kouskoff, Georges Lacaud (organiser), Alexander Medvinsky, Pablo Menendez, Elizabeth Ng, Trista North, James Palis, Roger Patient, Catherine Robin (organiser), David Traver, Laurent Yvernogeau
and Héloïse Dufour (Cercle FSER) & Claude Tran (journalist Educavox) for the 4th MSOP (Scientific Meeting Open to Public).
by Thierry Jaffredo, Catherine Robin and Georges Lacaud
20 – 23 March 2017
The seminar « Hemogenic Endothelium and Hematopoietic Stem Cell Production » took place in March 2017 at the splendid location of the “Domaine des Treilles”. It gathered eighteen developmental and cellular biologists (comprising 7 women and 4 junior participants) from Europe and overseas. The format of this closed meeting fostered informal and fruitful interactions as well as intense and passionate discussions around the cellular and molecular mechanisms and regulation of the hematopoietic system during embryonic development.
Keywords: hemogenic endothelium, hematopoietic stem cells, human embryonic stem cells, hematopoietic niches, animal and cellular models
General Aims and Questions
Hematopoietic Stem Cells (HSCs) are rare cells located in the Bone Marrow (BM) of each individual that maintain blood homeostasis. HSCs are responsible for the life-long curative effects of the transplantation procedures performed to treat patients with various blood related disorders. HSC transplantation has been successfully used therapeutically for over 50 years and accordingly the demand for clinical grade allogeneic HSCs has significantly increased over the past decades. However, the shortage in compatible HSCs available for clinic has become a real issue. A promising objective is to derive large quantities of fully functional HSCs from Embryonic Stem Cells (ESCs) or induced Pluripotent Stem Cells (iPSCs) in vitro without genetic modifications. Taking advantage of our knowledge that HSCs are generated from specialized endothelial cells (referred to as hemogenic endothelial cells or HE cells), recent exciting progresses in the generation of HSCs, from either endothelial cells or HE cells, have been reported and demonstrated to benefit from the support of an adequate vascular niche. However, these results were obtained with the forced expression of transcriptional factors. The generation of bona fide transgene-free HSCs desirable for therapy still remains an unattained goal, which will have profound impact on developing optimal therapies and will weigh heavily on health budgets. Therefore, there is an urgent need to improve our knowledge on HSC specification and, more generally on HSC production as it occurs physiologically in vivo.
Along this line, functional analysis of endothelial cells and lineage-tracing experiments performed in chicken, mice and zebrafish have shown that all HSCs emerge from HE cells through an endothelial to hematopoietic transition (EHT). There is also evidence that different blood cell types are generated from HE cells, depending on the anatomical site of HE and the developmental stage. HE cells within the yolk sac (YS) give rise to hematopoietic cells that exhibit a poor ability to reconstitute irradiated adult recipients upon transplantation whereas the aorta of the aorta-gonad-mesonephros (AGM) region produces long-term repopulating HSCs. AGM hematopoiesis appears under the form of small groups of cells, called hematopoietic clusters, that are attached to the endothelium lining the arteries wall and generally polarized to the ventral side. In the mouse YS, hematopoietic production starts as early as day 7.5 while in the AGM, hematopoiesis starts at day 9.5. AGM produces multipotent progenitors from day 9.5 to day 12, and HSCs from day 10.5 to day 12. A paradigm is that about 700 hematopoietic cluster cells are produced within the aorta while only less than 3 HSCs emerge among these. The cellular and molecular bases for these differences in numbers are unknown even though hematopoietic maturation has been evoked.
Hematopoietic emergence is thought to be partly controlled by the microenvironment, in particular the ventral mesenchyme of the AGM. The absence of mesenchyme clearly impairs Runx1 expression and the emergence of hematopoietic clusters. However, the exact role and composition of the HSC promoting/supporting microenvironment in terms of cell types and regulatory pathways involved is largely unknown. Furthermore, HE cells face different and changing microenvironments in the YS and AGM, which most likely influences the site of origin or time of formation of the different hematopoietic cell types. HE cells constitute less than 2% of the total endothelial cells present in YS and AGM. There is a major hurdle in dissecting EHT and HSC production. Indeed, the use of ESC models has been instrumental in revealing several cellular and molecular features of EHT. Although HE cells were recognized as arterial endothelium, this concept has been recently challenged and the precise cellular origin and nature of HE is still unknown. Moreover, the embryonic stage(s) at which endothelial progenitor cells become committed into HE cells and how this fate change is controlled at the molecular level remain poorly understood. Similarly, the cellular and molecular processes by which EHT proceeds are largely unknown, i.e. how flat endothelial cells become round and acquire a hematopoietic identity. Finally, one of the key questions is whether HE and hematopoietic production are restricted to embryonic life or whether vasculature of other fetal/adult organs harbours HE cells and are able to produce hematopoiesis through EHT.
– Anna Bigas (IMIM Barcelona, Barcelona, Spain) gave a historical overview of Notch function in the EHT process and HSC specification in different animal models. She addressed several key questions to understand HSC generation such as the importance of Notch levels, proliferation and clonality of cells forming hematopoietic clusters in the AGM in the mouse embryo by using imaging technologies and genetic labelling.
– Catherine Robin (Hubrecht Institute, Utrecht, Netherlands) discussed how the journey of HE cells to become HSCs involves cell polarization, cell shape remodelling, cell division, homing and cell attachment that require extensive cytoskeletal rearrangements. Following on her previous demonstration that adult mice knockout for Clasp2 (a microtubule-stabilizing protein) have a strong HSC defect in the BM, she now discussed new data indicating that CLASP2 already plays a critical role in the first steps of HSC production in the aorta of mouse (and zebrafish) embryos.
– Georges Lacaud (CRUK Manchester Institute, Manchester, UK) presented his recent work that established that the two transcriptional targets of RUNX1, GFI1 and GFI1B, are critical for the EHT process as they are driving the down-regulation of the endothelial program and the changes in morphology associated with the EHT. Both GFI1 proteins recruit epigenetic modifiers, including the CoREST complex with HDAC1 and HDAC2, to silence transcription at their target loci. They therefore evaluated the consequences of conditional deletion of Hdac1 and/or Hdac2 alleles on EHT and revealed the critical role of HDAC1 and HDAC2 and identified regulators of the TGFb pathway as their main targets. Finally, he presented some recent work indicating that a strict dosage of RUNX1 activity is critical for EHT. At the onset of the EHT, RUNX1 activity is moderate and negatively regulated by interaction with SOX7, whereas at later stage RUNX1 expression increase and its activity is synergized by interaction with CBFb.
– Charles Durand (UPMC, Paris, France) discussed his work to explore the transcriptional landscape of the AGM hematopoietic microenvironment. They isolated the tissues surrounding the dorsal aorta in mid-gestation mouse embryos using laser micro-dissection. By combining high-throughput technologies with computational analyses, they identified a unique molecular network characterizing the sub-aortic mesenchyme. They are currently analyzing the expression profiling and functional role of some candidate genes with the aim to identify novel regulators of the first emerging HSCs.
– Laurent Yvernogeau (Hubrecht Institute, Utrecht, Netherlands) also presented his search for regulatory signals emitted by the AGM microenvironment that control spatiotemporal HSC emergence. To search for genes that are differentially expressed in specific areas of the aorta where HSC emergence is either promoted or prevented, they used a novel genome-wide RNA tomography approach (tomo-seq). This technique allows evaluating gene expression patterns in the aorta microenvironment by using a genome-wide RNA sequencing approach with spatial information. Tomo-seq was performed on the aorta of zebrafish, chicken and mouse embryos. Their data will be a resource of gene expression patterns around the aorta. Moreover, the comparison of different species should highlight conserved regulatory molecules and pathways involved in the complex process of HSC generation during embryonic development.
– Alexander Medvinsky (CRM, Edinburgh, UK) has developed new cultures systems that recapitulate HSC development in vitro. This approach has allowed his group to establish the developing HSC hierarchy and determine requirements for Runx1 and Notch signalling during HSC maturation. They established that inductive interactions within the AGM region play an important role in HSC development. Ventrally expressed SCF and BMP inhibitors are important for HSC maturation. Although BMP signalling plays a role during early HSC specification, at later HSC formation stages, it is down-regulated, which is evident by functional transplantation studies and lack of pSmad1, 5, 8 in intra-aortic hematopoietic clusters.
– Marella de Bruijn (Oxford University, Oxford, UK) summarized her work on the identification of cell type and developmental stage specific enhancers of Runx1, a critical regulator of EHT, and generation of transgenic enhancer-reporter mouse models to isolate HE. Her functional and expression analysis work established that these cells undergo hematopoietic specification early in development when still part of the endothelial cell layer and identified new candidate players and pathways involved in the birth of blood.
– David Traver (UCSD, San Diego, USA) presented an unpublished story resulting from the identification of a gene uncovered in a forward genetic mutagenesis screen required for HSC emergence. In the absence of supt16h, HSC specification is lost, which is due to its requirement in transcriptional elongation of all components of the Notch signalling pathway. In contrast to previous literature on Supt16h function, no role in chromatin accessibility was uncovered in terms of HSC formation. Current directions are working to understand how this ‘housekeeping’ gene is needed in a relatively specific manner to form HSCs.
– Roger Patient (Oxford University, Oxford, UK) presented evidence that the lateral plate mesoderm (LPM) in zebrafish contains cells co-expressing blood and endothelial markers, possibly representing bipotent hemangioblast precursors, and that the cells that give rise to HSCs are biased towards endothelium in their expression profile, compared to the equivalent cells that give rise to the primitive blood, which are biased towards blood. He also presented the gene regulatory network for the definitive hemangioblast in Xenopus, whose programming begins with the miRNA inhibition of TGFbeta signalling and an input from BMP to enable the cells to respond to VEGFA signalling from the adjacent somites. As the cells migrate to the midline, they receive another VEGFA input from the somites but this time it is the larger isoforms of VEGFA which is involved. Once at the midline, the cells receive further input from TGFbeta and/or BMP. Finally, he presented some novel insights into the roles played by the critical transcription factors, GATA2 and RUNX1.
– Trista North (Harvard University, Boston, USA) spoke about her work examining extrinsic regulation of HSC formation in the developing embryo. In particular, she highlighted on going work from her lab focused on local (niche) and organismal cues such as metabolism and inflammatory signalling, which influence the timing, localization and scale of HSC production. She also highlighted the complex interplay of the primitive myeloid population, and their unique ability to perpetuate inflammatory networks and effector functions, such as MMP activity, to actively direct HSC formation and fate, including self-renewal, migration and differentiation in the vertebrate embryo.
– Andrew Elefanty (Monash University, Melbourne, Australia) discussed his work on human hematopoiesis using in vitro differentiation of human ESCs (hESCs) reporter lines to follow endothelial (SOX17) and hematopoietic (RUNX1C) development. Modelling extra-embryonic YS-like hematopoiesis revealed that the first endothelial–like cells did not express SOX17 (SOX17–CD34+CD43–) were the major source of hematopoietic progeny. Deletion of the hematopoietic transcription factor RUNX1 completely abrogated all blood cell formation during in vitro hESC differentiation, demonstrating that all human blood cell generation is RUNX1-dependent. Conversely deletion of SOX17 or of all three Group F SOX genes did not influence the production of endothelial or hematopoietic progeny.
– Elisabeth Ng (Monash University, Melbourne, Australia) examined the transcriptional profiles of hESC-derived and umbilical cord-derived CD34+ hematopoietic progenitors and found that the hESC-derived cells lacked the HOXA expression observed in repopulation-competent cord blood cells, suggesting that mesoderm was incorrectly patterned during the early stages of hESC differentiation. Using the SOX17 and RUNX1C reporter hESC lines, they showed that a two day treatment with a WNT agonist and ACTIVIN antagonist yielded CD34+ hematopoietic cells with a HOXA expression pattern that more closely resembled that of cord blood. The cultures generated a network of aorta-like SOX17+ vessels, expressing HOXA genes, from which RUNX1C+ blood cells emerged, similar to hematopoiesis in the AGM. Comparing the expression profiles of SB/CHIR treated CD34+ endothelial and hematopoietic cells with comparable cells sorted from human AGM showed similar expression of cell surface receptors, signalling molecules and transcription factors.
– Constanze Bonifer (Birmingham University, Birmingham, UK) presented work from a consortium of researchers that gathered data on the transcriptional regulation of embryonic hematopoiesis using ESC differentiation as model. This study identified and characterized the dynamic gene regulatory network driving hematopoietic specification and differentiation. Using chromatin profiling, the group also identified crucial components of the signalling network involved in this process, identifying Hippo signalling as being a crucial part. They also showed that transcription factors of the AP-1 family, which mediate MAPK signalling, are important for regulating the decision between vascular and hematopoietic fate. An important result from this work was that AP-1 transcription factors interact with TEAD factors mediating Hippo signalling, thus demonstrating how the two signalling pathway converge on the genome and regulate gene expression.
– Andrea Ditadi (SR-TIGET, Milan, Italy) presented a summary of the previous work he performed while being a post-doctoral fellow in Prof Gordon Keller’s lab, where he focused on lineage specification during the establishment of the human hematopoietic system using the human pluripotent stem cell (hPSC) model. Owing to its association with the arterial vasculature, it is assumed that HE represents a subpopulation of arterial vascular endothelium (VE). By careful dissection of lineage specification in hPSC model, he was able to challenge this assumption and show at a single cell level that HE represents a different lineage from VE. Additionally, he demonstrated that following differentiation, HE generates a multipotent hematopoietic progenitor that may represent the pre-HSC stage of human hematopoietic development. In parallel, HE generates hematopoietic progenitors with a more limited potential, which is restricted to erythroid and myeloid lineages. Both this developmental restriction and the hemoglobin composition of the erythroid progenitors generated following HE differentiation are reminiscent of YS-derived erythro-myeloid progenitor (EMP) cells. Andrea Ditadi is currently working to characterize this population and evaluate the potential contribution of these progenitors and bona fide EMPs to the onset of hematological disorders that manifest perinatally.
– Valerie Kouskoff (University of Manchester, Manchester, UK) presented data showing that the transcription factors SOX7 and RUNX1 are found co-expressed in HE at all sites of hematopoietic emergence during embryonic development and that interaction between these two transcription factors blocks hematopoietic commitment in HE. SOX7 binds to RUNX1 and hinders its interaction with CBFb, which is an essential co-factor of RUNX1 for stabilization and DNA binding. Together, this prevents the transcriptional activation of hematopoietic targets by RUNX1, a critical event for hematopoietic specification. She also presented data on engrafting cells derived from in vitro differentiated mouse ESCs, demonstrating that these cells emerge rapidly from the mesoderm and are very transient. Finally, she presented data exploring how cell fate specification of the HE can be modulated by culture conditions.
– Pablo Menendez (IJC, Barcelona, Spain) presented the attempts his lab has done over the past decade to get hematopoietic cells from hPSCs. He delineated his working model based on a population of hemogenic precursors/hemogenic endothelium, which is solely responsible to give rise to both blood and mature endothelial cells. Playing with different transcription factors, signalling pathways and microenvironment signals, one can regulate the dynamics and output of hemogenic precursors-blood cells. Interestingly, his lab has been exploring hPSCs as a tool to understand the developmental impact of fusion genes linked to childhood-infant acute leukemia, which is known to have a prenatal origin.
– Jim Palis (university of Rochester Medical Center, Rochester, USA) discussed the emergence of hematopoiesis in the murine embryo prior to the emergence of HSCs, which consists of two robust multilineage (primitive and EMP/definitive) waves of hematopoietic progenitors. EMPs are necessary and sufficient (together with primitive hematopoiesis) for survival of mouse embryos lacking HSCs. EMPs emerge from HE in the YS beginning at E8.5, independently of arterio-venous identity or blood flow. Recent unpublished studies indicate that Myb regulates lineage output of EMP and pre-HSCs, but not their initial emergence from HE in the YS and embryo proper, respectively. In addition, single cell analysis of EMP provides evidence that they contain definitive erythroid/megakaryocyte and myeloid lineage subpopulations. Studies of HSC-independent hematopoiesis will help to guide the differentiation of induced pluripotent stem cells and to increase our understanding of embryonic cell populations that persist in the adult.
– Thierry Jaffredo (UPMC, Paris, France) discussed first the formation of the vascular and blood systems in the early embryo and discussed the ontogeny of the aorta, the contribution of two distinct lineages in its formation and the role of the sub-aortic mesenchyme in triggering Runx1 expression and formation of the hematopoietic clusters. In a second part, he presented an unpublished story reporting the discovery of a HE in BM of the late fetus/young adult animal producing a cohort of hematopoietic stem and progenitor cells (HSPCs) in situ. The demonstration uses experimental embryology and genetic approaches in birds and mice respectively tagging either the somite or the vascular endothelium. Taken together, this reveals for the first time the existence of late fetal/early adult HE, suggesting that HSPCs can be generated de novo, past embryonic stages. Understanding the molecular events controlling this production will be critical to devise innovative therapies.
In addition to the scientific presentations and discussions on the control of the cellular identity, an outreach event was organised by Héloïse Dufour, from the “Cercle FSER” (“Fondation Schlumberger pour l’Education et la Recherche”). During this event, several groups of high school students from the area came to the Foundation des Treilles to meet with the participants of this meeting. Sixteen scientists participated in a « speed dating » format with small groups of students, and answered questions related to their research or gave career planning advices. The aim of this session was to encourage informal exchanges between scientists and the public to help bridge the gap between science and society and foster interest in fundamental science in the new generation.