Blood forming precursors in bone marrow
Larry H. Bernstein, MD, FCAP, Curator
LPBI
Blood stem cells study could pave the way for new cancer therapy
People with leukaemia could be helped by new research that sheds light on how the body produces its blood supply.
Scientists are a step closer to creating blood stem cells that could reduce the need for bone marrow transplants in patients with cancer or blood disorders.
Enabling scientists to grow the stem cells artificially from pluripotent stem cells could also lead to the development of personalised blood therapies, researchers say.
Blood stem cells are found in bone marrow and produce all blood cells in the body. These cells – known as haematopoietic stem cells (HSCs) – help to restore blood supply in patients who have been treated for leukaemia.
Researchers used a mouse model to pinpoint exactly how HSCs develop in the womb. They showed for the first time how three key molecules interact together to generate the cells, which are later found in adult bone marrow.
The discovery could help scientists to recreate this process in the lab, in the hope that HSCs could one day be developed for clinical use.
Scientists say this fundamental understanding of early development may also have an impact on other diseases that affect blood formation and supply.
###
The research has been published in Nature Communications.
Professor Alexander Medvinsky, of the University of Edinburgh’s MRC Centre for Regenerative Medicine said: “There is a pressing need to improve treatments for diseases like leukaemia and this type of research brings us a step closer to that milestone. The more we understand about how embryos develop these blood stem cells, the closer we come to being able to make them in the lab.”
http://www.ed.ac.uk/news/2016/stem-cells-100316
Céline Souilhol, Christèle Gonneau, Javier G. Lendinez, Antoniana Batsivari, Stanislav Rybtsov, Heather Wilson, Lucia Morgado-Palacin, David Hills, Samir Taoudi, Jennifer Antonchuk, Suling Zhao, Alexander Medvinsky. Inductive interactions mediated by interplay of asymmetric signalling underlie development of adult haematopoietic stem cells. Nature Communications, 2016; 7: 10784 DOI: 10.1038/ncomms10784
During embryonic development, adult haematopoietic stem cells (HSCs) emerge preferentially in the ventral domain of the aorta in the aorta–gonad–mesonephros (AGM) region. Several signalling pathways such as Notch, Wnt, Shh and RA are implicated in this process, yet how these interact to regulate the emergence of HSCs has not previously been described in mammals. Using a combination of ex vivo and in vivo approaches, we report here that stage-specific reciprocal dorso–ventral inductive interactions and lateral input from the urogenital ridges are required to drive HSC development in the aorta. Our study strongly suggests that these inductive interactions in the AGM region are mediated by the interplay between spatially polarized signalling pathways. Specifically, Shh produced in the dorsal region of the AGM, stem cell factor in the ventral and lateral regions, and BMP inhibitory signals in the ventral tissue are integral parts of the regulatory system involved in the development of HSCs.
Haematopoietic stem cells (HSCs) lie at the foundation of the adult haematopoietic system, and give rise to cells of all blood lineages throughout the lifespan of an organism. An important property of adult (definitive) haematopoietic stem cells (dHSCs) is that they are capable of long-term reconstitution of the haematopoietic system upon transplantation into irradiated recipients. In the mouse, such cells develop by embryonic stages E10–E11 in the aorta–gonad–mesonephros (AGM) region1, 2, 3, 4. An ex vivo approach showed that the AGM region has a robust autonomous capacity to generate dHSCs1. The AGM region comprises the dorsal aorta flanked on both sides by the urogenital ridges (UGRs), which contain embryonic rudiments of kidney and mesonephros. HSCs develop in a polarized manner, predominantly in the ventral floor of the dorsal aorta (AoV), more rarely in the dorsal domain of the dorsal aorta (AoD), and are absent in the UGRs2, 5, 6, 7. Localization of dHSCs to the AoV in mouse and human embryos was shown by long-term reconstitution experiments5, 6.
Abundant evidence indicates that during development, a specialized embryonic endothelial compartment known as haematogenic (or haemogenic) endothelium gives rise to haematopoietic stem and progenitors cells7, 8, 9, 10. The haematopoietic programme in various vertebrate models is executed predominantly in the AoV, and is recognized by the expression of essential haematopoietic transcription factors, for example, Runx1 and cKit, and the appearance of clusters of haematopoietic cells budding from the endothelium of the dorsal aorta6, 8, 9, 11, 12, 13, 14.
It is broadly accepted that HSCs develop from the haematogenic endothelium within intra-aortic clusters. This transition involves several consecutive maturation steps of HSC precursors: pro-HSCs
pre-HSC type Ipre-HSC type IIdHSC15, 16, 17. All these precursors express endothelial markers, such as vascular-endothelial cadherin (VC) and CD31, and sequentially upregulate haematopoietic surface markers: CD41 (pro-HSCs), CD43 (pre-HSC type I) and finally CD45 (pre-HSC type II). This maturation process occurs in the dorsal aorta between E9 and E11. Specifically, pro-HSCs emerge at E9, pre-HSCs Type I appear at E10 and pre-HSCs type II predominantly at E11. Unlike dHSCs, pre-HSCs cannot reconstitute the adult haematopoietic system by direct transplantation and require prior maturation in an embryonic or neonatal environment15, 16, 17, 18,19.
A number of signalling pathways (Notch, Wnt, retinoic acid, interleukin-3 and inflammatory) have been implicated in HSC development; however, a coherent picture is yet to be elucidated15, 17, 20,21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31. HSC precursors (pro-HSCs, pre-HSCs type I and pre-HSCs type II) express cKit17 from early developmental stages. A recent study has shown that the cKit ligand, known as stem cell factor (SCF), is a key regulator driving maturation of these HSC precursors into dHSCs in the AGM region17, which is in agreement with the marked decline of HSC activity in SCF mutant mice32, 33. In the adult, SCF is critically important for HSC maintenance in the bone marrow niche, mainly in the endothelial compartment32. Sonic Hedgehog (Shh) and bone morphogenetic protein 4 (BMP4) pathways are also important mediators; in zebrafish, these two morphogenes are involved in arterial specification and haematopoietic patterning, respectively34,35. In the mouse, subaortic BMP4 and Shh/Indian Hedgehog derived from gut were also proposed to be responsible for HSC development36, 37.
During development, interactions between spatially segregated compartments are essential for tissue patterning and specification, and are often mediated by gradients of secreted molecules38,39, 40. Molecules secreted by distant tissues, such as somites, can influence HSC development in the AGM region41, 42, 43, 44, 45. Developing HSCs are embedded in the complex AGM microenvironment, suggesting that HSC development may require signals derived from different compartments of the AGM region. We sought to test this hypothesis. However, the analysis of HSC development in vivo is significantly hampered by low accessibility of embryos developing in utero, fast maturation of dHSCs, lack of uniquely specific markers for HSC precursors and their low numbers in the AGM region. Therefore, we employed here a robust ex vivo culture system that models HSC development in the embryo in combination with functional HSC analysis using in vivolong-term reconstitution assay15, 16, 17. Specifically, to study interactions between AGM subregions, we took advantage of the in vitro reaggregation system that enables close juxtaposition of cell types15.
We show that interactions between three compartments of the AGM, the AoV, the AoD and the UGRs, are necessary for efficient generation of dHSCs. First, we show that dHSC activity in the isolated E10.5 AoV is limited but can be significantly enhanced by co-culture with the AoD, and that this is mediated at least partly by Shh, secreted dorsally in vivo. Second, while HSC activity in isolated E11.5 AoD is limited, co-culture with a competent AoV microenvironment activates dHSC generation in the AoD. We found that this effect is mediated by SCF, which is secreted abundantly by the AoV stroma in vivo as shown here. Third, we show that downregulation of BMP4 signalling by BMP antagonist Noggin, which is present at high levels in the AoV and especially in intra-aortic clusters as revealed here by in vivo observations, is required for HSC development. Fourth, UGRs, which express high levels of SCF, also enhance HSC development in the dorsal aorta.
Our results based on in vivo observations and ex vivo modelling strongly suggest that juxtaposed, anatomically distinct domains within the AGM region create a complex landscape of interactive signals that underpins HSC development.
Pre-HSCs localize preferentially to the AoV
As dHSCs mature from pre-HSCs, we investigated whether the emergence of dHSC predominantly in the AoV6 is a result of asymmetric (ventralized) distribution of pre-HSCs. Dorsal aortae were separated from UGRs and bisected into AoV and AoD (including notochord) as described previously6 (Supplementary Fig. 1a). The different domains were then directly transplanted into irradiated mice to detect dHSCs. We first confirmed our previous observation that at E11.5 dHSCs appear almost exclusively in the AoV, although some dHSCs were in the AoD and engrafted few recipients at high level (Supplementary Fig. 1b). Limiting dilution analysis showed that dHSCs are approximately four times more frequent in the AoV compared with AoD. UGRs did not contain HSCs in line with previous reports2, 6.
We then investigated the spatial distribution of pre-HSCs type I and pre-HSCs type II in E10.5–E11.5 embryos using the OP9 co-culture system supplemented with Il3+SCF+Flt3 (termed 3GF), which allows pre-HSCs (which do not engraft by direct transplantation) to mature into dHSC that become detectable by long-term repopulation assay as described previously16. Doses of transplanted cells (expressed in embryo equivalents, e.e.) were chosen based on the requirements of individual experiments (explained in Methods section). In these experiments (Fig. 1), the dose injected was high (1–2 e.e.) to detect potentially low dHSC numbers in AoD and UGRs.
(a) E10.5 AoV, AoD and UGRs were co-aggregated with OP9 and cultured for 5 days, and the formation of dHSCs was then tested by transplantation into irradiated mice (2 e.e. per recipient; AoV: six independent experiments; AoD: four independent experiments; UGRs: two independent experiments). Dashed line indicates the cutoff for high-level engraftment (>70% donor chimaerism). (b) E11.5 aortas and UGRs were transplanted after reaggregate culture (Ao: 0.2 e.e. per recipient and UGRs: 1 e.e. per recipient; two independent experiments). (c,d) Pre-HSCs type I (VC+CD45−) (c) or type II (VC+CD45+) (d) sorted from E11.5 AoV and AoD were co-aggregated with OP9 cells and transplanted after culture (1 e.e. per recipient; two independent experiments). (a–d) Levels of engraftment are plotted, and number of repopulated versus total number of transplanted mice are shown in brackets. Number of embryo equivalents (ee) injected in each experiment are indicated on the graphs. (*P<0.05; ***P<0.005; Mann–Whitney U-test). In all these experiments, tissues were cultured with three growth factors (Flt3I, Il3 and SCF). AGM, aorta–gonad–mesonephros region; Ao, dorsal Aorta; AoV, ventral domain of the dorsal aorta; AoD, dorsal domain of the dorsal aorta; UGRs, urogenital ridges.
We have shown previously that E10.5 AGM region mainly contains type I pre-HSCs, whereas at E11.5, type I and type II pre-HSCs co-exist16. Dissected E10.5 AGM regions co-cultured with OP9 in 3GF for 5 days were transplanted into adult irradiated recipients. Out of 21 recipients that received cultured AoV, 20 showed high levels (>70%) of donor-derived long-term haematopoietic chimerism (Fig. 1a). In contrast, only 7 out of 16 recipients of cultured AoD were repopulated at high levels (>70%), while the remaining recipients showed lower or no repopulation (7 and 2, respectively). Cultured UGRs did not produce dHSCs (Fig. 1a). Thus, we conclude that the E10.5 AoD does contain pre-HSCs but at significantly lower numbers than the AoV.
We then investigated whether pre-HSCs localization changes in E11.5 embryos and found that pre-HSCs were still exclusively localized to the dorsal aorta; UGRs carefully separated from the lateral mesenchyme adjacent to the dorsal aorta did not give any repopulation after culture (Fig. 1b). To establish the location of pre-HSCs within the E11.5 dorsal aorta, cell populations enriched for pre-HSCs type I (VC+CD45–) and pre-HSCs type II (VC+CD45+) were sorted from AoV and AoD, and co-cultured with OP9 stromal cells in the presence of 3GF as described previously16. We again were able to detect pre-HSC activity in AoD although at lower levels than in AoV. After maturation ex vivo, pre-HSCs type I from AoV and AoD repopulated 7 of 11 and 2 of 8 recipients, respectively (Fig. 1c). Similarly, cultured pre-HSCs type II from AoV and AoD repopulated 11 out of 12 and 4 out of 10 recipients, respectively (Fig. 1d). In all cases, multilineage engraftment was confirmed (Supplementary Fig. 2). These data show that pre-HSCs are significantly enriched in AoV.
Reciprocal inductive interactions between AoD and AoV
To explore hypothetical interactions between AoD and AoV, we made use of a dissociation–reaggregation system that recapitulates HSC development ex vivo15. This system allowed us to integrate AGM domains in a three-dimensional tissue-like organoid15 and study their interactions in HSC development. To track the origin of dHSCs, AoV and AoD from wild-type (WT) and green fluorescent protein (GFP) embryos with constitutive expression of GFP46 were co-aggregated (termed AoV//AoD co-aggregates) and cultured for 5 days in the presence of 3GF before transplantation (Fig. 2a). Mice transplanted with AoV//AoD co-aggregates can be reconstituted by dHSCs coming from AoD and AoV. The presence of GFP allowed the individual contributions of AoV and AoD to the total repopulation level within the same mouse to be assessed (Fig. 2b,c). This is presented in two separate columns in the graph. Namely, while columns 1 and 3 represent the same recipient mice, the former shows exclusively the contribution of the AoD and the latter shows exclusively the contribution of the AoV into each recipient. To assess the influence of AoD and AoV interaction on HSC development, the repopulation by co-aggregated AoD (column 1) or AoV (column 3) can then be compared with repopulation by independently cultured AoD (column 2) or AoV (column 4). All experiments included reciprocal use of WT and GFP tissues in AoV//AoD co-aggregates, and we observed no difference in repopulation properties between WT and GFP embryos. Homotypic AoV//AoV and AoD//AoD co-aggregates were always used as controls. Note that in these experiments, only 0.2 e.e. were injected per recipient, to ensure that the repopulation levels were not saturated and to allow any inductive effects to be revealed.
(a) Experimental design: the ventral domain (AoV) and the dorsal domain (AoD) of the aorta, and the urogenital ridges (UGRs) from wild-type (WT) and GFP+ embryos were subdissected, and chimeric reaggregates from tissues of these two genotypes were generated. Left column: to test interactions between AoV and AoD, chimeric AoV//AoD reaggregates were generated and transplanted into irradiated recipients after 4–5 days of culture (b,c). Right column: to test interactions between Ao and UGRs, chimeric Ao//UGR reaggregates were generated and transplanted into irradiated recipients after 4–5 days culture (d). GFP+ and/or GFP− donor-derived long-term repopulation allowed us to conclude whether dHSCs originated from AoV, AoD or UGRs. Accordingly, the tissue of origin of donor dHSCs is indicated below each graph. (b) E10.5 aortas from WT and GFP embryos were used to generate chimeric reaggregates as depicted schematically above plots. The reciprocal combination of WT and GFP tissues was used to generate AoV//AoD reaggregates. The tissue source of dHSCs is shown separately in the leftmost (AoD) and rightmost (AoV) columns as indicated below the plot (0.2 e.e. per recipient; two independent experiments). (c) E11.5 aortas from WT and GFP embryos were used to generate chimeric reaggregates. The tissue source of dHSCs is shown separately in the leftmost (AoD) and rightmost (AoV) columns as indicated below the plot (0.2 e.e. per recipient; two independent experiments). (d) E11.5 aortas (Ao) and UGRs from WT and GFP embryos were used to generate Ao//UGR chimeric reaggregates. As depicted schematically above the plot, the reciprocal combination of WT and GFP tissues was used to generate Ao//UGR reaggregates. The tissue source of dHSCs is shown separately in left (Ao) and right (UGRs) columns as indicated below the plot (0.01 e.e. per recipient; six independent experiments). (e) Reaggregation of WT Ao with UGRs generate more dHSCs than Ao alone (0.05 e.e. per recipient; two independent experiments). (b–e) In all these experiments, tissues were cultured with three growth factors.
Using this approach, we found that the E10.5 AoV generates more dHSCs when combined with AoD than on its own (Fig. 2b, compare two rightmost columns). One day later, E11.5 AoD had no positive influence on dHSC generation by AoV (Fig. 2c, compare two rightmost columns). Conversely, the E11.5 AoD produced more HSCs when reaggregated with the AoV than on its own (Fig. 2c, compare two leftmost columns). This inductive effect of AoV on AoD was not observed at E10.5 (Fig. 2b, compare two leftmost columns). These ex vivo modelling experiments revealed reciprocal stage-specific effects of AoV and AoD on HSC development, which could be explained by the differential release of factors by the two regions and/or by differences in the competency of the target cells to respond to signals.
UGRs enhance HSC development in the dorsal aorta
SCF expression is involved in polarized HSC development
Figure 3: Involvement of polarized stem cell factor in HSC development.
(a) qRT–PCR on fresh AoV, AoD and UGRs at E10.5 and E11.5 showed high expression levels of stem cell factor (SCF) in AoV and UGRs, compared with AoD (data are mean±s.e.m; *P<0.05, **P<0.01, t-test; three independent experiments). No significant difference was observed between E10.5 and E11.5 expression level in any of the tissues. (b) Expression of SCF-GFP and CD31 determined by immunostaining on thick section (300 μm) of SCF-GFP-positive E10.5 AGM region and on SCF-GFP-negative control. Bars, 50 μm. (c) Expression of SCF in sorted populations from fresh E10.5–E11.5 AoV (V) and AoD (D) determined by qRT–PCR. Endo, endothelial population (VC+CD45−CD43−); type I, pre-HSCs type I (VC+CD45−CD43+); type II, pre-HSCs type II (VC+CD45+); stroma, stromal population (VC−CD45−CD43−). (*P<0.05, t-test; five independent experiments). (d) E10.5 AoD were cultured as reaggregates in the presence of Il3 and Flt3L with or without SCF and human SCF antagonist (SCF-Rh). (0.5 e.e. per recipient; three independent experiments). (e,f) E11.5 AoD (two independent experiments) (e) and E10.5 AoV (two independent experiments) (f) were cultured as explants with or without SCF (no other cytokines); (0.2 e.e. per recipient).
Shh signalling enhances dHSC generation
Figure 4: Sonic Hedgehog is a positive modulator of pre-HSC type I.
(a) Expression level of Sonic Hedgehog (Shh) in E10.5 and E11.5 AGM region determined by qRT–PCR. (data are mean±s.e.m; *P<0.05, t-test; E10.5: three independent experiments and E11.5: two independent experiments). (b) Patched1 and Gli1 expression in endothelial cells (endo: VC+CD45−CD43−), pre-HSCs type I (I: VC+CD45−CD43+) and type II (II: VC+CD45+) sorted from E11.5 AoV and AoD (two independent experiments). (c) E10.5 AoV and AoD explants were cultured in presence of Shh recombinant protein before transplantation (AoV: 0.1 e.e. per recipient; two independent experiments and AoD: 0.2 e.e. per recipient; three independent experiments). (d) E10.5 AoV and doxycyline-inducible OP9-Shh were co-aggregated and cultured in presence or absence of doxycycline and/or Hedgehog (Hh) antagonist (200 nM) before transplantation (0.2 e.e. per recipient; two independent experiments). (e) 10.5 AoV and AoD co-aggregated with OP9 were cultured in presence of three growth factors with Hh antagonist before transplantation; (0.2 e.e. per recipient; two independent experiments). (f) E11.5 AoV explants were cultured in presence of Shh recombinant protein before transplantation; (0.2 e.e. per recipient; two independent experiments). (g): E11.5 AGM reaggregates were cultured in presence of Hh antagonist before transplantation; (0.1 e.e. per recipient; two independent experiments). (c,d,f,g) In all these experiments, tissues were cultured without cytokines. Hh anta, Hh antagonist; Dox, doxycycline.
BMP signalling is downregulated in the dHSC lineage
Figure 5: Bone morphogenetic protein signalling is downregulated in dHSC lineage.
(a) Expression of bone morphogenetic protein 4 (BMP4) at E10.5 determined by qRT–PCR; (data are mean±s.e.m.; *P<0.05, t-test; three independent experiments). (b) Expression of BMP4 in the E10.5 AGM region determined by immunostaining on frozen sections. Bars, 50 μm. Zoomed image shows the subendothelial localization of BMP4 (arrowheads). Bars, 10 μm. (c) Expression of phosphorylated-Smad (P-Smad) in the E10.5 AGM region determined by immunostaining on frozen sections. Bars, 50 μm. (d) Id genes expression in endothelial cells, pre-HSCs type I and type II directly isolated from E10.5 and E11.5 AoV determined by qRT–PCR. Endo, endothelial population (VC+CD45−CD43−); type I, pre-HSCs type I (VC+CD45−CD43+); type II, pre-HSCs type II (VC+CD45+); stroma, stromal population (VC−CD45−CD43−). (Data are mean±s.e.m.; *P<0.05, **P<0.01; t-test; five independent experiments). (e–g) Expression of P-Smad, CD31 and CD45 in the endothelium and haematopoietic clusters of E10.5 dorsal aorta. White arrowheads indicate cells with pre-HSC type II phenotype (CD31+CD45+); green arrows show (CD31+CD45−/low) cells budding out of the dorsal aorta and expressing P-Smad; asterisks indicate CD31+CD45− cells expressing P-Smad within the endothelium. Bars, 10 μm. A positive control showing P-Smad staining in the dorsal part of the neural tube can be found in h.
Figure 6: Haematopoietic clusters are exposed to low concentration of BMP4 and high levels of Noggin.
(a) Expression of BMP antagonists at E10.5 determined by qRT–PCR (data are mean±s.e.m.; *P<0.05,***P<0.005; t-test; three independent experiments). (b) Expression of Noggin in the E10.5 AGM region determined by immunostaining on frozen sections. Note the expression of Noggin in the notochord (Nt) as expected. Bar, 50 μm. (c) Expression of Noggin and BMP4 in intra-aortic clusters characterized by cKit and CD31 expression. Note that BMP4 is mainly expressed underneath the dorsal aorta (arrowheads), while Noggin is expressed in the cluster (arrows). Bars, 10 μm. (d) Expression of Noggin in isolated populations from E10.5 and E11.5 AoV (V) and AoD (D) determined by qRT–PCR. (*P<0.05, t-test; five independent experiments). (e) Model showing downregulation of BMP activity in dHSC lineage. BMP4 is mainly expressed in the ventral mesenchyme, while Noggin is found in haematopoeitic clusters. Accordingly, BMP activity, assessed by the phosphorylation of Smad1,5 and 8 (P-Smad), is high in mesenchymal cells underneath the aortic endothelium and in some endothelial cells (CD31+CD45−) of the aortic endothelium and decreases in the haematopoeitic clusters. While some pre-HSC type I cells (CD31+CD45−/low) exhibit BMP signalling at a low level, acquisition of CD45 (shown in red) is accompanied by a complete loss of BMP activity. EC, endothelial cells; MC, mesenchymal cells; I, pre-HSC type I; II, pre-HSC type II.
BMP signalling inhibits HSC development
http://www.nature.com/ncomms/2016/160308/ncomms10784/images_article/ncomms10784-f7.jpg
Interactions between SCF, Shh and BMP signalling pathways
We have shown previously that during murine embryo development definitive HSCs emerge predominantly in the ventral domain of the dorsal aorta (AoV)6. This spatially polarized production of HSCs might be explained by different origins of dorsal and ventral endothelium and/or by asymmetric production of key factors involved in HSC development37, 52, 53 and we reasoned that directional inductive interactions between AGM compartments could be involved. Great insight into inductive interactions in various organs has previously been obtained through in vitro modelling39. Here we modelled interactions between AGM domains in a co-culture system, which supports HSC development15. Using this ex vivo system, we demonstrate that at early stages (E10.5) HSC maturation in the AoV region is enhanced by the presence of the AoD. One day later (E11.5), the AoV microenvironment is able to induce HSC development in the AoD, previously thought to be mostly devoid of HSC activity6. We also found that UGRs can enhance HSC production from the dorsal aorta, but cannot generate dHSCs themselves, even under influence of the dorsal aorta. Thus, our data strongly suggest that reciprocal stage-specific inductive AoD//AoV interactions and involvement of UGRs are required for execution of the robust development of HSCs in vivo.
Our data indicate that previously established dorso–ventrally polarized HSC development6 is defined by two main factors. First, our current data show that although the AoD contains pre-HSCs (both type I and type II), their numbers are lower than in AoV, in line with lower intra-aortic cluster formation previously described in mouse AoD6, 13. Second, as shown here, dHSCs can be induced in the AoD by the AoV, and therefore the dHSCs deficiency in AoD cannot be explained solely by asymmetric pre-HSC distribution, but may also be influenced by differences in the microenvironment.
To study this, we focused on SCF, Shh and BMP4, whose expressions are dorso–ventrally polarized in the AGM region36, 47, 49 (and current data). We found that SCF is an inductive signal that is expressed at high levels in the AoV and UGRs, and can stimulate HSC development in isolated AoD, a region which had previously been considered to be mostly devoid of HSC activity. This is in agreement with a key role of SCF in HSC maturation17. We found that the aortic endothelial compartment expresses high levels of SCF, suggesting its important role in HSC development comparable to the bone marrow microenvironment of adult HSCs32. Importantly, we found that the pre-HSC type I population expresses SCF suggesting a positive-autocrine loop, which could promote HSC development.
Shh signalling in zebrafish is required for aortic angioblast migration and subsequent arterial specification of the dorsal aorta34, 54. We found that in mouse Shh stimulates and a Hh antagonist inhibits the development of HSCs at E10.5 but not at E11.5, in keeping with a previous study37. The induction of dHSCs in AoV by AoD is also limited to the E10.5 stage. Since Shh is secreted by the notochord (which is included in AoD-dissected tissue), this stage specificity is likely defined by the predominant presence of pre-HSCs type I at E10.5, which express higher levels of Shh signalling components (Ptch1 and Gli1) compared with pre-HSCs type II. By E11.5, the pre-HSC population is mainly represented by type II cells15. Stage-specific loss of sensitivity to Hh signalling was also described in the developing neural tube55. Notably, the poor ability of AoD to develop HSCs despite abundant presence of Shh can also be explained by lower levels of Ptch1 and Gli1 detected in AoD- compared with AoV-derived pre-HSC type I. Our ex vivo modelling data indicate that AoD-derived Shh is an active inducer of HSC development in the AGM region. This conclusion does not exclude the possibility that Shh secreted by the gut could also reach the dorsal aorta37, although by E10.5 these sites are separated by an extended mesentery.
BMP4 signalling is a key factor involved during differentiation of ventral mesoderm and its further specification into haematopoietic cells. In zebrafish, BMP signalling is clearly required during the patterning of the dorsal aorta and for the emergence of dHSCs in the ventral wall34. Its role in mouse is less clear due to the early lethality of BMP mutants56. Several lines of evidence point to BMP4 as a good candidate regulating HSC development. Indeed, BMP4 is highly expressed in the ventral mesenchyme underneath the dorsal aorta34, 36, 49; some reports suggested its role in controlling dHSC emergence36, 57, 58. However, the in vitro systems used likely assayed the maintenance of dHSCs, rather than their maturation. It was also reported that BMP4 signalling can define their differentiation potential59. BMP4 is also involved in the regulation of essential haematopoietic transcription factors such as Scl/Gata2/Fli1 and Runx1 (refs 60, 61). Here we analysed BMP signalling activity in the dHSC lineage in the AGM region. We show that in vivo the pre-HSC type I to type II transition is accompanied by a downregulation of BMP targets (Id genes). This correlates with our data demonstrating that BMP activity is downregulated in intra-aortic clusters and the observations of others that Runx1 expression is attenuating in the developing HSC lineage60, 62, 63. How is this decrease of BMP activity achieved in vivo, despite the presence of BMP4 in AoV? It has previously been noted that in amphibian embryos several BMP inhibitors are also expressed in AoV34. Similarly, our analysis of the embryo showed high expression of a number of BMP antagonists as well as inhibitory Smad6 and Smad7 in mouse AoV that may counteract BMP4 action in HSC lineage. Furthermore, we found that in the AGM region BMP4 and Noggin are spatially segregated: Noggin being present in haematopoietic clusters and BMP4 being mainly expressed underneath the aortic endothelium. Therefore, maturing HSCs in clusters are exposed to low BMP4 concentration and high concentration of the BMP antagonist Noggin. Furthermore, our qRT–PCR analysis shows that the pre-HSC type I population expresses Noggin, which possibly creates a very effective shield that protects them from BMP4. Accordingly, our ex vivo analysis strongly suggests that downregulation of BMP signalling is functionally important for HSC development in the embryo. Indeed, forced BMP signalling activation by the addition of BMP4, strongly inhibits HSC development, and conversely the addition of Noggin stimulates HSC development in E10.5–E11.5 AGM cultures. These results are in line with recent observation that deletion of Smad4, a common transducer for BMP4/TGFβ signalling, markedly augments the formation of intra-aortic clusters64. Our data do not exclude the possibility that BMP4 is essential for specification of mouse dHSCs at earlier stages, as described in the zebrafish model, where BMP signalling is required for HSC development at stages closer to mouse E8.5 (ref. 34).
Our analysis indicates that all three signalling pathways studied can cooperate for HSC development (Fig. 8c). Notably, the interplay of Shh and BMP pathways is broadly involved in development. For example, counter gradients of polarized Shh and BMP signalling in the developing spinal cord specify neuronal subsets along the dorso–ventral axis65, and the dorsal aorta resembles the neural tube with inverse orientation of Shh- and BMP-secreting centres34. However, we detected an antagonistic relationship between Shh and BMP pathways. At the molecular level, Shh can induce Noggin and Smad6 expression, thus inhibiting BMP4 signalling. In turn, BMP4 suppresses and, accordingly, Noggin enhances Shh signalling. Cooperation between Shh and Noggin has been previously described as critically important for developmental specification of somitic, neural and hair follicle cells66, 67, 68. Our in vitro data suggest that the feed-forward loop ShhNoggin/NogginShh is also involved in HSC development in vivo.
We propose a model where the polarized secreted factors form complex fields of gradients in vivo, which define an effector zone for optimal HSC development in the dorsal aorta and lead to the ventrally shifted appearance of dHSCs (Fig. 8c). Of interest, intra-aortic clusters are abundant in ventro–lateral positions69, which may reflect the position of this zone. The dissection close to such a zone could lead to accidental inclusion of powerful dHSCs in AoD samples observed here. Furthermore, it is possible that spatial segregation of co-operating and spatial overlap of antagonizing factors may also be important for adjustment of HSC development in vivo. Indeed, although the pool of pre-HSCs in the AGM region markedly expands during E9.5–11.5 (Rybtsov et al., submitted), complete maturation of the HSC pool is limited: while the majority of cells reach the pre-HSC type II stage, only one or two dHSCs are generated by the end of E11. Such controlled dynamics of HSC development may be needed to prevent a burst of active haematopoiesis in the AGM region. How exactly HSC maturation dynamics depend on overlapping concentrations of factors requires further analysis. Although ex vivo modelling is a powerful tool to dissect mechanisms of HSC development in vivo, there will likely be some variation in details. For example, spatial polarization in the developing HSC niche may define kinetics of HSC development in vivo.While we have demonstrated spatial polarization in vivo of the factors driving HSC development in our model system, it is currently unclear whether any factors become expressed in a polarized manner within the reaggregates and as such, whether polarization is also a pre-requisite for HSC maturation. Alternatively, if polarization is not required, the entire reaggregate may replicate the optimal zone for HSC development, resulting in massive generation of dHSCs. The distinction between these two scenarios will require further investigation.
In summary, our ex vivo modelling experiments suggest that HSC development in the embryo involves stage-dependent interactions between dorsal, ventral and lateral domains of the AGM region, mediated at least partly by the interplay of SCF, Shh, BMP4 and Noggin. Further detailed analysis will be required to better understand the complexity of the AGM signalling landscape in which HSC development takes place. Such knowledge may lead to development of novel protocols for the generation of definitive HSCs in vitro for clinical applications.
Integrated genomic DNA/RNA profiling of hematologic malignancies in the clinical setting
Key Points
-
Novel clinically-available comprehensive genomic profiling of both DNA and RNA in hematologic malignancies.
-
Profiling of 3696 clinical hematologic tumors identified somatic alterations that impact diagnosis, prognosis, and therapeutic selection
The spectrum of somatic alterations in hematologic malignancies includes substitutions, insertions/deletions (indels), copy number alterations (CNAs) and a wide range of gene fusions; no current clinically available single assay captures the different types of alterations. We developed a novel next-generation sequencing-based assay to identify all classes of genomic alterations using archived formalin-fixed paraffin-embedded (FFPE), blood and bone marrow samples with high accuracy in a clinically relevant timeframe, which is performed in our CLIA-certified CAP-accredited laboratory. Targeted capture of DNA/RNA and next-generation sequencing reliably identifies substitutions, indels, CNAs and gene fusions, with similar accuracy to lower-throughput assays which focus on specific genes and types of genomic alterations. Profiling of 3696 samples identified recurrent somatic alterations that impact diagnosis, prognosis and therapy selection. This comprehensive genomic profiling approach has proved effective in detecting all types of genomic alterations, including fusion transcripts, which increases the ability to identify clinically-relevant genomic alterations with therapeutic relevance.
Cohesin Ring Rules Blood Stem Cells, Binds Them to Renewal or Expansion
A genome-wide RNAi screen was used to assess the effects of 15,000 genes on the balance between self-renewal and differentiation of human hematopoietic stem cells (HSCs). The screen identified candidate genes whose knockdown maintained the HSC phenotype during culture. Such findings could lead to better protocols to grow these cells outside the body, potentially making bone marrow transplants more available to patients suffering blood cancers, or even identifying novel genes to target during the treatment of leukemia (left and right panels). Four genes in particular implicated cohesin, a ring-like protein complex that binds to the DNA in all of our cells, in the control of self-renewal versus differentiation in HSCs. Deficiency of cohesin causes an increase in self-renewal and a decrease in differentiation of HSCs. [Cell Reports]
Best known for its ability to regulate the separation of sister chromatids during cell division, the cohesin protein complex, a ring-shaped structure, has shown that it has other powers, such as the facilitation of DNA repair and the modification of transcription. And now, according to scientists based at Lund University, there is evidence that the cohesin complex controls the growth of blood stem cells. More to the point, the cohesin complex determines whether blood stem cells self-renew or differentiate.
The new finding is significant because it can help scientists improve the expansion of blood stem cells outside the body, thus increasing the supply of blood stem cells to patients suffering leukemia or hereditary blood disorders. Besides making bone marrow transplant material more available, the new finding could point scientists to new points of attack for the treatment of blood cancer, which is a disruption between blood stem cell multiplication and maturation.
The Lund University scientists, led by Jonas Larsson, presented their results March 17 in the journal Cell Reports, in an article entitled “Genome-wide RNAi Screen Identifies Cohesin Genes as Modifiers of Renewal and Differentiation in Human HSCs.” The article describes how a genome-wide RNA interference (RNAi) screen was performed in primary human CD34+ cells. This screen enabled the scientists to identify candidate genes whose knockdown maintained the HSC phenotype during culture.
“A striking finding was the identification of members of the cohesin complex (STAG2, RAD21, STAG1, and SMC3) among the top 20 genes from the screen,” wrote the authors. “Upon individual validation of these cohesin genes, we found that their knockdown led to an immediate expansion of cells with an HSC phenotype in vitro.”
A similar expansion, the authors added, was observed in vivo following transplantation to immunodeficient mice.
“Transcriptome analysis of cohesin-deficient CD34+ cells showed an upregulation of HSC-specific genes,” the authors continued. This finding, the authors asserted, demonstrates that when cohesin is deficient, transcription shifts to a more stem cell–like pattern.
“The research is unique as the study of so many genes alongside one another is unprecedented,” said Dr. Larsson. “In addition, we have used human blood stem cells, which is difficult in itself as it is requires the gathering of a large amount of material.”
Of the 15,000 genes that were tested, the Lund team found around 20 candidates with a strong capacity to affect the balance of growth in the blood stem cells. What was striking was that four of these 20 genes were physically connected through cooperation in a protein complex.
“The discovery showed that this protein complex is crucial and has an overarching function in the growth of the blood stem cells,” emphasized Dr. Larsson.
The cohesin complex acts as a sort of brace that holds different parts of the DNA strand together in the cell. The researchers believe that this allows the cohesin complex to control access to the “on/off switches” in DNA and to change the impulses the blood stem cells receive from various genes, thereby affecting cell division. The blood stem cell either multiplies or matures to become a specialized cell with other tasks.
Independently of the Lund researchers’ discovery, other research in the field of blood cancer has recently identified mutations in exactly the same four genes in patients with various forms of blood cancer.
“This is incredibly exciting! Together with the results from our study, this indicates that the cohesin genes are directly and crucially significant in the development of blood cancer,” exclaimed the study’s lead author, Ph.D. candidate Roman Galeev. “Our findings entail a new understanding of how the expansion of blood stem cells is controlled. Eventually, this can lead to new ways of affecting the process, either to prevent the development of cancer or to expand the stem cells for transplant.”
UNPRECEDENTED PRECISION STUDY IDENTIFIES THE FOUR GENES RESPONSIBLE FOR BLOOD STEM CELL DEVELOPMENT.
Article Info- •A genome-wide RNAi screen was performed in primary human CD34+ cells
- •Several cohesin genes were identified as modifiers of renewal and differentiation
- •Cohesin-deficient HSCs show enhanced reconstitution capacity in vivo
- •Cohesin deficiency induces immediate HSC-specific transcriptional programs
Summary
To gain insights into the regulatory mechanisms of hematopoietic stem cells (HSCs), we employed a genome-wide RNAi screen in human cord-blood derived cells and identified candidate genes whose knockdown maintained the HSC phenotype during culture. A striking finding was the identification of members of the cohesin complex (STAG2, RAD21, STAG1, andSMC3) among the top 20 genes from the screen. Upon individual validation of these cohesin genes, we found that their knockdown led to an immediate expansion of cells with an HSC phenotype in vitro. A similar expansion was observed in vivo following transplantation to immunodeficient mice. Transcriptome analysis of cohesin-deficient CD34+ cells showed an upregulation of HSC-specific genes, demonstrating an immediate shift toward a more stem-cell-like gene expression signature upon cohesin deficiency. Our findings implicate cohesin as a major regulator of HSCs and illustrate the power of global RNAi screens to identify modifiers of cell fate.
Human hematopoiesis is maintained by a small number of hematopoietic stem cells (HSCs) that are capable of generating all blood cell lineages at an extremely rapid pace for the entire lifespan of a human being (Orkin and Zon, 2008). HSCs have been studied extensively during the last four decades and are probably the best functionally characterized adult stem cells. However, despite this, the regulatory mechanisms that govern different cellular fate options in HSCs have remained incompletely defined. In particular, it has been challenging to understand the molecular basis of the inherent ability of HSCs to self-renew and preserve their undifferentiated state, which has hampered efforts to expand HSCs ex vivo for therapeutic benefit (Dahlberg et al., 2011). Ex vivo expansion of HSCs would allow for critical improvements of bone marrow transplantation procedures in treatment of malignant and inherited hematological diseases (Chou et al., 2010). Defining the genetic and molecular basis of self-renewal of HSCs is thus important to enhance current cell-therapy strategies, but it is also essential in order to better understand mechanisms behind dysregulated hematopoiesis that may cause leukemia. Genes and pathways balancing cell-fate options between renewal and differentiation in stem cells are often key players in cancer development (Orkin and Zon, 2008).
Figure 1
Genome-wide RNAi Screen in Primitive Human Hematopoietic Cells Defines Genes and Pathways Associated with Cancer Progression and Cell Proliferation
(A) Overview of the experimental outline for the primary screen. 60 million cord blood-derived CD34+ cells were transduced with a pooled lentiviral library containing 75,000 shRNAs across six transduction replicates in total. A fraction of the cells were isolated after 72 hr, and proviral inserts were deep sequenced to determine the initial library distribution. Following 20 days of culture, CD34+ cells were magnetically isolated and proviral inserts were sequenced again to determine the changes in distribution for all shRNAs.
(B) Relative distribution of shRNAs following 20 days of in vitro culture, ranked from the most enriched to the most depleted. The y axis shows the average enrichment value across six replicate screens.
(C) Gene ontology analysis for all genes represented by multiple shRNAs in the most enriched (10%) fraction.
(D) KEGG pathway analysis showing strong enrichment for cancer-associated pathways among the top-scoring genes.
- A Secondary Screen Targeted to CD34+CD38−CD90+CD45RA−Cells Identifies the Cohesin Complex as a Candidate Regulator of HSC Renewal and Differentiation
- Knockdown of Cohesin Genes Restricts Differentiation and Expands Primitive Cells In Vitro
- Cohesin-Deficient HSCs Show Delayed Differentiation but Overall Increased Reconstitution Potential following Transplantation to Primary and Secondary Immunodeficient Mice
- Silencing of Cohesin Genes Results in Immediate Transcriptional Changes with a Shift toward a More HSC-like Gene Expression Signature
We report here on the successful development of a genome-wide RNAi screening approach targeted to primary human hematopoietic stem and progenitor cells to define genes and pathways associated with self-renewal and differentiation. Based on findings from the screen, we implicate the cohesin complex as a crucial regulator of cell-fate decisions influencing self- renewal and differentiation in HSCs both in vitro and in vivo.
These efforts represent a genome-wide RNAi screen targeted to primary human HSPCs. The main limiting factor when performing functional screens in primary human cells is cell number. This obviously becomes even more challenging when rare cell subsets, such as stem and progenitor cells, are studied. Through unique access to cord blood with daily deliveries from several local hospitals, we were able to gather the necessary quantities to perform a screen in enriched primary HSPCs with reasonable coverage (300X).
2012pharmaceutical
Amandeep Kaur
Aashir Awan, Phd
Abhisar Anand
Adina Hazan
Alan F. Kaul, PharmD., MS, MBA, FCCP
alexcrystal6
anamikasarkar
apreconasia
aptubman
Arav Gandhi
aviralvatsa
David Orchard-Webb, PhD
danutdaagmailcom
Demet Sag, Ph.D., CRA, GCP
Dror Nir
dsmolyar
Ethan Coomber
evelinacohn
Gail S Thornton
Irina Robu
jayzmit48
jdpmd
jshertok
kellyperlman
Ed Kislauskis
larryhbern
Madison Davis
marzankhan
megbaker58
ofermar2020
Dr. Pati
pkandala
ritusaxena
Rick Mandahl
sjwilliamspa
Srinivas Sriram
stuartlpbi
Dr. Sudipta Saha
tildabarliya
zraviv06
zs22
Leave a Reply