Portal de Periódicos da CAPES

Haematopoietic stem cell niches: new insights inspire new questions

2013; Springer Nature; Volume: 32; Issue: 19 Linguagem: Inglês

10.1038/emboj.2013.201

1460-2075

Fernando Ugarte, E. Camilla Forsberg,

Mesenchymal stem cell research

Review10 September 2013free access Haematopoietic stem cell niches: new insights inspire new questions Fernando Ugarte Fernando Ugarte Department of Biomolecular Engineering, Institute for the Biology of Stem Cells, University of California Santa Cruz, Santa Cruz, CA, USA Search for more papers by this author E Camilla Forsberg Corresponding Author E Camilla Forsberg Department of Biomolecular Engineering, Institute for the Biology of Stem Cells, University of California Santa Cruz, Santa Cruz, CA, USA Search for more papers by this author Fernando Ugarte Fernando Ugarte Department of Biomolecular Engineering, Institute for the Biology of Stem Cells, University of California Santa Cruz, Santa Cruz, CA, USA Search for more papers by this author E Camilla Forsberg Corresponding Author E Camilla Forsberg Department of Biomolecular Engineering, Institute for the Biology of Stem Cells, University of California Santa Cruz, Santa Cruz, CA, USA Search for more papers by this author Author Information Fernando Ugarte1 and E Camilla Forsberg 1 1Department of Biomolecular Engineering, Institute for the Biology of Stem Cells, University of California Santa Cruz, Santa Cruz, CA, USA *Corresponding author. Department of Biomolecular Engineering, Institute for the Biology of Stem Cells, University of California Santa Cruz, 1156 High Street, Santa Cruz, CA 95064, USA. Tel.:+1 831 459 2111; E-mail: [email protected] The EMBO Journal (2013)32:2535-2547https://doi.org/10.1038/emboj.2013.201 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Haematopoietic stem cell (HSC) niches provide an environment essential for life-long HSC function. Intense investigation of HSC niches both feed off and drive technology development to increase our capability to assay functionally defined cells with high resolution. A major driving force behind the desire to understand the basic biology of HSC niches is the clear implications for clinical therapies. Here, with particular emphasis on cell type-specific deletion of SCL and CXCL12, we focus on unresolved issues on HSC niches, framed around some very recent advances and novel discoveries on the extrinsic regulation of HSC maintenance. We also provide ideas for possible paths forward, some of which are clearly within reach while others will require both novel tools and vision. What drives the desire to understand HSC niches? Haematopoietic stem cell (HSC) niches are an intensely studied topic that aims at understanding the elusive cues from the microenvironment that supports HSC function in vivo. Combined with the intrinsic properties of HSCs, these signals are capable of maintaining a perfect balance of HSC quiescence and proliferation, self-renewal and differentiation to maintain haematopoietic homeostasis under drastically dynamic conditions throughout life. In a field of many 'firsts'—prospective isolation of the stem cell, single cell transplantation, stem cell therapy—HSC biology also suffers from many unknowns. It is frustrating to admit that despite decade-old knowledge that adult HSCs reside primarily in the bone marrow (BM), we do not have a clear mechanistic understanding of the cues that attract them to and retain them in the BM, and we do not understand the complex inputs that enable HSC self-renewal within these niches (Figure 1). These gaps in understanding the basic biology of HSC regulation hamper our clinical treatment strategies. The difficulty in achieving robust ex vivo expansion of functional HSCs—considered the 'holy grail' of haematopoiesis (Sauvageau and Humphries, 2010)—and inability to derive robustly engrafting HSCs from pluripotent cells means that we cannot provide a reliable supply of HSCs for all patients, although many lives could be saved and painful diseases cured if we did. We also lack protocols for specific manipulation of HSC dislodgment out of and engraftment into BM niches. Thus, more than 50 years after the first successful haematopoietic transplantation in humans, we still have to use non-specific, broadly damaging and, too often, lethal approaches to enable transplanted HSCs to engraft and thrive long term in the recipient. By comprehending the function of HSC niches, we hope to solve these issues and many more, including how extrinsic cues affect lineage output and contribute to leukaemogenesis and other haematopoietic disorders. Figure 1.Visualization of HSC niches at increasing resolution. Recent findings have contributed to our knowledge of structural, cellular, and molecular features important for HSC localization and maintenance. (A) Although HSCs can, at least temporarily, circulate and take up residence in organs such as spleen and liver, the majority of functional adult HSCs reside in the bone marrow. (B) Within the bone marrow, HSCs can be located in either perivascular or endosteal regions, where they interact directly or indirectly with different types of cells that comprise HSC niches (C). (D) These niche cells secrete factors that maintain haematopoietic homeostasis by favouring HSC quiescence, while allowing self-renewal and differentiation upon demand. This figure illustrates how the definition of HSC location is becoming increasingly more detailed, from organ to region, cellular, and molecular environment, by the use of novel tools and increasingly sophisticated approaches. Download figure Download PowerPoint Several excellent and comprehensive reviews on HSC niches have been published recently (Isern and Méndez-Ferrer, 2011; Frenette et al, 2013; Krause et al, 2013; Smith and Calvi, 2013). Therefore, although many factors are involved in HSC maintenance, we restrict the scope of this review to selected unresolved issues, using recent discoveries on the structural, cellular, and molecular regulation of HSC maintenance by extrinsic factors as the framework for discussion and future directions. Technical and conceptual advances by recent publications Defining HSC location relative to other components of the BM is a challenging task, with technical difficulties ranging from preparation of BM sections, cell-specific labelling, and high-resolution detection methods. Nevertheless, major advances have been made in the past years to narrow down the location of HSC niches (Figure 1). While HSCs are known to circulate and reside in multiple tissues, they primarily localize to the BM (Figure 1A). Within the BM, both perivascular and endosteal niches have been described (Kiel and Morrison, 2008), providing potentially different cellular contexts for HSCs (Figure 1B and C). Increasingly, specific molecules in different cellular environments are being probed for their contribution to HSC function (Figure 1D). A breakthrough in the ability to investigate HSC location was made when the Morrison laboratory established the SLAM markers to identify stringently defined HSCs by two-colour analysis, and locate them close to sinusoidal endothelial cells (SECs) in BM and spleen (Kiel et al, 2005). One technical hurdle in moving forward from those studies has been limitations in microscopy capability. High-resolution analysis usually limits examination to only a few areas of a BM section, making the search for the rare HSCs prohibitively labour intensive and time consuming. The analysis of only a fraction of HSCs that reside in each bone may miss features of the comprehensive environment. A recent study by Nombela-Arrieta et al (2013) used novel technology to overcome these limitations. Laser scanning cytometry (LSC) is a slide-based microscopy technique that allows for the quantification of light emitted by all cells in a relatively large section of tissue, therefore not limiting analysis to a small field of view (see LSC review; Harnett, 2007). In their study, Nombela-Arrieta et al used a combination of LSC and confocal imaging to describe the three-dimensional vascular architecture of the BM at a more global level, revealing that the BM environment is highly vascularized, especially at the bone-proximal regions. They also comprehensively quantify the location and distribution of haematopoietic cells, demonstrating an enrichment of haematopoietic stem and progenitor cells (HSPCs) in perivascular areas of bone-proximal regions. This observation suggests that vascular and endosteal niches should not be viewed as two different compartments, but rather that highly vascularized endosteal regions may provide the complex cellular and molecular environment necessary for HSC maintenance. This finding also fits the idea of different types of niches in close proximity, supporting either quiescence or proliferation of HSPCs depending on demand (Kiel and Morrison, 2008). Interestingly, Nombela-Arrieta et al also observed that hypoxic HSPCs are evenly distributed throughout the BM, in contrast to previous observations that highlighted the endosteum as a hypoxic compartment favourable for HSC maintenance (Eliasson and Jönsson, 2010; Mohyeldin et al, 2010; Takubo et al, 2010; Suda et al, 2011). The ability of this new technology to visualize single cells and structures in large, three-dimensional space will facilitate the creation of more comprehensive, architectural maps of the BM vasculature, with distribution of niche cells and HSCs. How such maps change under different conditions will be important for understanding the dynamics of HSC regulation by these environments. New discoveries on the spatial organization of HSC niches were also made recently by Wang et al (2012). Using a dual-colour reporter mouse (mT/mG mice; Muzumdar et al, 2007) in combination with an endothelial-specific inducible allele (cdh5(PAC)-creERT2), they were able to distinguish SECs from perivascular stromal cells (PVCs). This led to the identification of a zone between SECs and PVCs containing clusters of HSCs, which they named hemospheres. These distinct hemosphere compartments were of variable size, but highly enriched for HSCs on the abluminal surface of a sinusoidal vessel (Figure 2). Deletion of VEGFR2 from endothelial cells disrupted the formation of hemosphere structures and led to reduced numbers of HSCs in the BM. This exciting new finding adds a spatial dimension to the cellular organization of HSC niches. Importantly, it provides a conceptual framework for testing how the BM is compartmentalized to provide the specialized context necessary for HSC function. Figure 2.The hemosphere niche model. A novel view of the perivascular niche was proposed by Wang et al, where HSCs and haematopoietic progenitors are contained in defined areas termed hemospheres (Wang et al, 2012), delimited by SECs and perivascular stromal cells. This area may be highly enriched for molecular cues that promote HSC maintenance and self-renewal, such as SCF and CXCL12. Broad acceptance of this model will likely require validation by alternative genetic model and imaging techniques. Download figure Download PowerPoint In addition to advances in structural organization, our understanding of both the cellular and molecular specificity of HSC maintenance has taken great strides forward. Several recent studies have capitalized on the well-established roles of the ckit and CXCR4 receptors and the corresponding ligands, SCF and CXCL12, in HSC function (Chabot et al, 1988; Williams et al, 1990; Nagasawa et al, 1996; Broudy, 1997; Zou et al, 1998; Peled et al, 1999; Nie et al, 2008; Kimura et al, 2011); to define the cellular and molecular environment necessary for HSC maintenance. Building on the pioneering studies of Sugiyama et al who used a CXCL12-GFP reporter mouse to identify a small fraction of perivascular cells in close proximity to putative HSCs, named CXCL12-abundant reticular (CAR) cells (Sugiyama et al, 2006), several groups have used similar strategies to label and functionally test the importance of potential niche cells and molecules. Mendez-Ferrer et al used a Nestin-GFP reporter mouse to identify a rare population of mesenchymal stem/progenitor cells (MSPCs) highly enriched for factors implicated in HSC maintenance, including CXCL12 and SCF (Méndez-Ferrer et al, 2010). The Nestin+ cells were located in perivascular regions close to putative HSCs. Deletion of the Nestin+ population significantly reduced the number of HSCs in the BM, demonstrating that these perivascular MSPCs play important roles in HSC maintenance. Then, the Morrison and Link labs turned the question around by focussing on specific molecules—SCF and CXCL12—and asking which cells are responsible for their production (Ding et al, 2012; Ding and Morrison, 2013; Greenbaum et al, 2013). Intriguingly, the two groups independently demonstrated that deletion of either SCF or CXCL12 from endothelial and perivascular cells, but not from other stromal or haematopoietic cell types, led to impaired HSC maintenance (see Table I for Cre recombinase mouse lines used in these studies). In an impressive demonstration that cell type-specific production of secreted factors matters, they also observed that deletion of CXCL12 from osteoblast lineage cells (OBCs) resulted in decreased numbers of lymphoid progenitors, but not of HSCs. Although a few differences between the findings from the two groups need to be reconciled, their studies converge on a model where different niche cells support different stem and progenitor populations: HSCs are supported by SCF- and CXCL12-expressing vascular and perivascular cells, whereas lymphoid progenitors are maintained by CXCL12-expressing OBCs. Table 1. Cre recombinase mouse lines used in recent HSC niche studies Cre-tg SCF deletion Cxcl12 deletion Ding et al (2012) Ding and Morrison (2013) Greenbaum et al (2013) UBC-CreER X X — Vav-Cre X X — Col2.3-Cre X X — Nestin-Cre X X — Lepr-Cre X X — Tie2-Cre X X X Prx1-Cre — X X Oc-Cre — — X Osx-Cre — — X We anticipate that the powerful tools and concepts established in these studies will be applied to the many additional secreted molecules implicated in HSC function (for a more comprehensive list of factors, see Wilson and Trumpp, 2006). Integration of the findings on microvessel organization (Nombela-Arrieta et al, 2013), hemosphere structures (Wang et al, 2012), and cell type-specific production of SCF and CXCL12 (Ding et al, 2012; Ding and Morrison, 2013; Greenbaum et al, 2013) promises to advance our understanding of HSC functions far beyond our current knowledge. In this context, we discuss some of the conceptual and technical challenges of addressing the complexities of HSC-supporting niches, and provide suggestions for further progress. Niche cell nomenclature, isolation, and classification strategies The genetic reporter strategies mentioned above delineated several new cell populations (e.g., CAR cells and Nestin+ cells), while other BM cells have been categorized by functional characteristics (e.g., MSPCs), by location (e.g., PVCs), or based on the markers typically expressed by a cell lineage (e.g., OBCs). This has led to a rapidly growing list of BM cells suggested to play a role in supporting HSC maintenance (Table II). The different strategies used to identify new cell populations, combined with inconsistent nomenclature, have precluded clear delineation of the relationship, and in many cases significant overlap, between these populations. Improvements in stromal cell naming conventions and isolation protocols would avoid unnecessary confusion on the identity of BM cell types. Table 2. Bone marrow niche cells Cell type Identity/Markersa Mouse modelb Niche function Perivascular stromal cell-Nes-GFP-SCF-GFP-Cxcl12-GFP/DsRed(CAR) Cxcl12hiSCFhighNG2Pdgfr-aPdgfr-bCD146Sca-1Vcam1CFU-F potentialMesenchymal lineage differentiation Nestin-GFP (Méndez-Ferrer et al, 2010; Ding et al, 2012)Prx1-cre (Ding and Morrison, 2013; Greenbaum et al, 2013)Cxcl12-GFP (Sugiyama et al, 2006)/DsRed (Ding and Morrison, 2013)Cxcl12-DTR-GFP (Omatsu et al, 2010)Osx-cre (Greenbaum et al, 2013)Lepr-cre (Ding et al, 2012; Ding and Morrison, 2013)Scf-GFP (Ding et al, 2012) HSC maintenance and retention in BMDirect supply of soluble and non-soluble factors for HSCsPrecursors of other niche cells—osteoblasts, adipocytes, chondrocytes, etc. Sinusoidal endothelial cells CD31EndomucinVE-cadherinVcam1LamininSca-1MECA-32Endoglin Tie2-cre (Ding et al, 2012; Ding and Morrison, 2013; Greenbaum et al, 2013)Cdh5-CreER (Wang et al, 2012) HSC maintenance Osteoprogenitors OsterixRunx2CD146 Osx-cre (Greenbaum et al, 2013)Prx1-cre (Ding and Morrison, 2013; Greenbaum et al, 2013) Lymphoid progenitor cell niche Osteoblast (OB) OsteocalcinOsteopontinALPLN-cadCxcl12low Prx1-cre (Ding and Morrison, 2013; Greenbaum et al, 2013)Col2.3-cre (Ding et al, 2012; Ding and Morrison, 2013)Osteocalcin-cre (Greenbaum et al, 2013)Osx-cre (Ding and Morrison, 2013; Greenbaum et al, 2013) Lymphoid progenitor cell niche Adipocytes Fabp4AdiponectinPerilipin A-ZIP/F1 (Naveiras et al, 2009) Negative regulators of haematopoiesis Monocytes/Macrophages Gr-1CD169CD11bF4/80 CD169-DTR (Chow et al, 2011)Gr-1-DTR (Chow et al, 2011)MAFIA (Winkler et al, 2010; Chow et al, 2011)Clodronate liposomes (Winkler et al, 2010; Chow et al, 2011)CD68:G-CSFR (Christopher et al, 2011) Regulating MSC and OB function. Active player in G-CSF mobilization. Non-myelinating Schwann cells Active TGF-BGFAPNestin Tgfbr2fl/− (Yamazaki et al, 2011) Maintenance of HSC quiescence a aCommonly used markers to describe the respective cell population. b bMouse models used to interrogate the respective cell population. Inconsistencies in nomenclature range from simple naming conventions to more complex issues. For example, the term BM 'stromal cell' sometimes excludes osteoblasts (Shestopalov and Zon, 2012) and sometimes is used to refer specifically to the progeny of mesenchymal stem cells (MSCs) (Frenette et al, 2013). In turn, the term 'MSC' has widely been used to describe the heterogeneous populations of cells obtained from various tissues, often without rigorous assessment of the self-renewal or lineage potential expected of true MSCs (Cao et al, 2013). Similarly, 'perivascular stromal cell' has been used to describe several different types of cells located in proximity to SECs, including CAR, Nestin-GFP, SCF-GFP, and Lepr-expressing cells. The extent of overlap between these cell types remains to be determined, but is likely significant. Clarifying the lineage relationship between the different cell populations will also be important, and is facilitated by the recent establishment of new Cre/lox reporter lines. Unfortunately, such studies are not without caveats. For example, it was proposed recently that haematopoietic progenitors can give rise to osteoblasts upon transplantation (Hofmann et al, 2013), contradicting the dogma that osteoblasts originate from mesenchymal progenitors. Additionally, the specificity of Nestin-mediated reporter activity varies between different transgenic lines (Ding et al, 2012), and it is clear that the populations of cells targeted by the different Cre lines (Tables I and II) are not yet completely mapped and may in some cases target several cell types or lineages (Hanoun and Frenette, 2013). To date, some of the newly identified PVCs are believed to be mesenchymal progenitors, based on their potential to differentiate into osteo- and adipocyte lineages, as shown for CAR cells (Omatsu et al, 2010), or based on the CFU-F potential and in vivo ossicle formation, as shown for Nestin-GFP cells (Méndez-Ferrer et al, 2010). In contrast, lineage tracing experiments showed that Lepr-expressing cells did not give rise to osteoblasts (Ding et al, 2012). The functional properties of other types of PVCs remain unknown. A productive approach to this issue was used by Sacchetti et al (2007), who showed that a clonal population of human CD146+ stromal cells were capable of self-renewal and formation of a haematopoietic microenvironment upon transplantation into heterotopic sites in mice. Whether PVCs have this ability and whether HSCs can colonize these ectopic areas remain to be tested. An additional source of variation is caused by differences in cell extraction protocols. Recovery and relative frequency of cell types greatly depend on the isolation method, and the methods established for harvesting haematopoietic cells from BM preclude recovery of some of the more adherent cell types. Different techniques such as flushing, crushing, and enzymatic digestion will yield different numbers and types of stromal cells. For example, we detected few, if any, endothelial cells (defined as CD45-Ter119-CD31+ cells) using standard crushing methods, but the recovery increased substantially with collagenase treatment (Smith-Berdan et al, 2012). Similarly, GFAP+ cells, though detected under the microscope in BM sections, were not recovered in BM cell suspensions (Yamazaki et al, 2011). Suboptimal and inconsistent cell isolation protocols make it difficult to accurately quantitate the numbers of distinct cell type and therefore, as discussed below, their relative contribution to secreted ligands. In addition, reliable isolation of viable cells is important in order to couple phenotype with function, as has been so valuable for understanding the haematopoietic system. Differences in isolation protocols may account for some of the inconsistencies in expression levels of HSC-supporting factors by different cell types. For example, Ding et al found the highest levels of CXCL12 expression in PVCs, followed by endothelial cells and then by OBCs (Ding and Morrison, 2013), whereas others have reported that OBCs express higher or equivalent levels of CXCL12 compared to endothelial cells (Semerad et al, 2005; Smith-Berdan et al, 2012; Greenbaum et al, 2013). In addition, failure to recover viable cells corresponding to the phenotype described in BM sections hampers genome-wide molecular characterization and precludes assessment of their lineage potential. Although HSC niches are a rapidly evolving research area, a consistent nomenclature based on the lineage and functional characteristics would be hugely beneficial in building a comprehensive picture of BM stromal cells. Here, in an attempt to adhere to a consistent, yet admittedly incomplete, nomenclature, we use the term 'stromal cell' to refer to all non-haematopoietic cells (generally defined as CD45−Ter119− or CD45−Lin− cells) in the BM. Stromal cells will therefore include MSCs, perivascular and endothelial cells, osteoblasts, adipocytes, chondrocytes, non-myelinating Schwann cells, and other cell populations of these lineages yet to be defined (Figure 3). We refer to cells that have committed to the osteoblast lineage as OBCs, inclusively referring to osteoprogenitors as well as mature osteoblasts (Marie, 2003; Stein et al, 2004). We also use the term MSPC, as in Frenette et al 2013, for cell populations believed to contain MSPCs, but not always well characterized or consistently referred to (for a recent and comprehensive review on mesenchymal cells and their role in HSC regulation, see Frenette et al, 2013). Figure 3.Putative lineage relationships of some of the BM cells that have been implicated in HSC function. Lineage hierarchy of the distinct candidate niche cells. Perivascular stromal cells described using different genetic models significantly overlap in terms of function and molecular profile. Some of these cells have mesenchymal stem/progenitor properties, giving rise to osteo-, adipo-, and chondrocyte lineages. Other BM cells implicated in HSC function include haematopoietic-derived osteoclasts and macrophages, sinusoidal and non-SECs, adipocytes and non-myelinating Schwann cells. Download figure Download PowerPoint Considering the high complexity of the BM microenvironment, we believe that it will be productive to establish the developmental hierarchy of different types of PVCs and other BM stromal cells. A recent study using genetic lineage tracing for splenic stromal cells identified a common, multipotent precursor of different types of spleen stromal subsets capable of initiating the formation of an artificial lymphoid structure upon transplantation (Castagnaro et al, 2013). It will also be important to establish protocols capable of dissociating viable cells without destruction of epitopes that define these cells in situ. GFP-driven transgenes can be a productive approach, but additional strategies that are less time consuming and labour intensive, and more amenable to the multi-marker cell labelling that will likely be necessary to understand the complexity of HSC niches, would be extremely valuable. Secreted factors may affect HSC function by both direct and indirect mechanisms The recent reports by Ding et al and Greenbaum et al confirmed the previously established roles for ckit and CXCR4 as essential regulators of HSC function. However, HSCs are not the only cell type responding to SCF and CXCL12, and it is therefore possible that indirect effects, by these ligands acting on niche cells, also play important roles in HSC maintenance (Figure 4). Figure 4.CXCL12 may affect HSC function by both direct and indirect mechanisms. In (A), the primary function of CXCL12 (yellow triangles) is to support HSC function by acting directly on HSCs. In (B), CXCL12 also affects HSC function by indirect mechanisms, by acting on CXCR4-expressing niche cells (effect number 1). This action of CXCL12 may function to reinforce an HSC-supportive environment by either cellular mechanisms, such as niche cell function or organization, and/or by molecular mechanisms, by triggering release of additional HSC-supporting factors, such as SCF or angiopoietin (blue circles), from niche cells (effect number 2). Download figure Download PowerPoint Germline deletion of either CXCL12 or CXCR4 results in severely impaired BM haematopoiesis and late embryonic lethality (Nagasawa et al, 1996; Zou et al, 1998; Tachibana et al, 1998). The similar phenotypes of CXCL12 and CXCR4 deletions have contributed to the prevailing notion that CXCL12 serves as the main ligand for CXCR4, and it is clear that HSCs directly respond to CXCL12 by expression of CXCR4 (Peled et al, 1999; Smith-Berdan et al, 2011; Wright et al, 2002). Similarly, the rapid mobilization of HSCs from BM to blood upon in vivo administration of either anti-ckit antibodies (Czechowicz et al, 2007) or CXCR4 inhibitors such as AMD3100 (Broxmeyer et al, 2005; De Clercq, 2010; Smith-Berdan et al, 2011) provides convincing evidence that ckit and CXCR4 function directly in HSC retention to BM niches. However, several lines of evidence suggest that SCF/ckit and CXCL12/CXCR4 interactions on other cells affect HSC function indirectly. For example, two models of conditional CXCR4 deletion in adulthood, one in all cell types (Rosa26-CreER; Nie et al, 2008) and the other selectively (but not exclusively; Park et al, 2012) in haematopoietic cells (Mx1-Cre; Sugiyama et al, 2006; Nie et al, 2008) resulted in different effects on HSC numbers. In addition, Sugiyama et al (2006) reported that, surprisingly, HSC location next to CXCL12-expressing cells did not depend on CXCR4 expression by HSCs. Lastly, global CreER-mediated deletion of CXCL12 in adulthood appeared to be more severe than the analogous deletion of CXCR4 (Sugiyama et al, 2006; Nie et al, 2008). The extent to which other surface receptors offset the loss of CXCR4 to ensure proper HSC function is not clear (Forsberg and Smith-Berdan, 2009), but such intrinsic compensatory mechanisms may be more efficient than the ability to compensate for the loss of a ligand secreted by neighbouring cells. It is also possible that CXCL12 acts on non-haematopoietic cells expressing CXCR4, and that these stromal cells are important for HSC function (Figure 4B). Indeed, there is evidence for expression and function of the CXCR4 receptor by BM stromal cells. Importantly, CXCR4 deletion from osteoprogenitors, using an Osx-Cre transgene, resulted in impaired osteoblast differentiation and proliferation, accompanied by altered bone formation and extracellular matrix composition (Zhu et al, 2011). CXCR4 was also shown to be essential for vascular endothelial cell development, as mice lacking CXCR4 have defective vessel formation, in addition to impaired haematopoiesis (Tachibana et al, 1998). Thus, deletion of CXCL12 may strongly influence the function of osteoblasts or endothelial cells, including their ability to support HSCs. To resolve these issues, it will be important to determine which stromal cells express the CXCR4 and ckit receptors and whether these cells are affected, directly or indirectly, by deletion of SCF or CXCL12 (Figure 4). While Ding et al (2012) demonstrated that the frequency of SCF-GFP BM cells was normal in Lepr-Cre;SCFfl/gfp mice, it is possible that the organization or expression profile of SCF-expressing cells is altered upon SCF deletion. By capitalizing on the technical and conceptual approaches used by Nombela-Arrieta et al and Wang et al, microvessel organization and hemosphere integrity can be investigated upon deletion of SCF and CXCL12. Determining the consequences of receptor deletion in specific stromal cell types, as already done for CXCR4 in osteoblasts in the report mentioned above (Zhu et al, 2011), and the potential impact on HSC maintenance will likely yield important insights into the crosstalk between different types of BM cells. Additionally, time courses in conditional deletion models may reveal whether one or more types of putative niche cells are affected prior to reduction in HSC numbers or function. While the issues at hand are com