Bone and Blood Vessels: The Hard and the Soft of Hematopoietic Stem Cell Niches
2009; Elsevier BV; Volume: 4; Issue: 6 Linguagem: Inglês
10.1016/j.stem.2009.05.011
ISSN1934-5909
AutoresRussell Garrett, Stephen G. Emerson,
Tópico(s)Hematological disorders and diagnostics
ResumoOsteoblasts are important regulators of myelopoiesis and lymphopoiesis, producing several necessary soluble and membrane-associated factors. In addition, they may play important roles, along with other mesenchymal populations, in constructing an environment that is suitable for development of sinusoidal niches capable of supporting hematopoietic stem cells. Osteoblasts are important regulators of myelopoiesis and lymphopoiesis, producing several necessary soluble and membrane-associated factors. In addition, they may play important roles, along with other mesenchymal populations, in constructing an environment that is suitable for development of sinusoidal niches capable of supporting hematopoietic stem cells. Hematopoietic stem cells (HSCs) are tasked with the de novo generation and lifelong maintenance of all the cells comprising the vertebrate blood system. They do so through the balanced production of new HSCs, referred to as HSC self-renewal, and more mature and increasingly lineage-restricted progenitors, referred to as differentiation. The functions of HSCs and those downstream progenitors and precursors ultimately responsible for the maturation and proliferation of mature cell numbers are controlled by factors both intrinsic and extrinsic to the specific cells. It is hypothesized that specific microenvironments, or "niches," exist within the bone marrow that organize these factors in order to support specific aspects of hematopoiesis, such as HSC survival, self-renewal, and differentiation. Given that, at steady state, adult hematopoiesis takes place exclusively in bone, a tie between the formation of blood cells and some unique character of bone has long been suggested. After all, while definitive HSCs arise in the aorta-gonad-mesonephros of the mammalian embryo and shortly thereafter move to the fetal liver, there is a rapid and permanent perinatal shift of the entire hematopoietic system to the bone cavity. While the exact combination of cellular and noncellular components that define these hematopoietic microenvironments (HMEs) is slowly becoming clearer, it is almost certain that bone- and/or marrow-specific elements drive much of this HSC localization. A specific role for osteoblasts in supporting the survival and/or differentiation of hematopoietic stem and progenitor cells (HSPCs) was hypothesized based on the proximity of both cell types to the bone endosteum (Taichman and Emerson, 1994Taichman R. Emerson S. J. Exp. Med. 1994; 1: 1677-1682Crossref Scopus (289) Google Scholar). Osteoblasts and bone marrow stromal cells (BMSCs) exist at the endosteum, playing critical roles in bone homeostasis, including the following: (1) formation of new bone through the mineralization of a type I collagen matrix; and (2) regulation of osteoclastogenesis, bone resorbing cells. Uncovering this latter role demonstrated that osteoblasts, in addition to their primary osteogenic function, also regulate the fate of hematopoietic cells, as osteoclasts are derived from a monocytic precursor. More recently, the production of bone matrix, the regulation of free calcium concentrations, and the presence of specific adhesive niches on osteoblast or osteoprogenitor surfaces have all been implicated in the adhesion and quiescence of HSCs. The first direct role discovered for osteoblasts in blood cell production was the support of terminal granulopoiesis. While it was known that bone marrow fibroblasts were capable of producing myeloid-specific cytokines such as granulocyte colony-stimulating factor (G-CSF) and granulocyte monocyte colony-stimulating factor (GM-CSF) in response to inflammation, the source of basal granulocyte production was unknown. Taichman and Emerson, 1994Taichman R. Emerson S. J. Exp. Med. 1994; 1: 1677-1682Crossref Scopus (289) Google Scholar demonstrated that osteoblast monolayers induced the differentiation of some CD34+ bone marrow cells to mature neutrophils, at least in part due to G-CSF production. Most recently, in vivo studies confirmed that loss of osteoblasts leads to a rapid and significant reduction in the number of cKit+ promyelocytes, as well as more mature granulocytes and monocytes (Zhu et al., 2007Zhu J. Garrett R. Jung Y. Zhang Y. Kim N. Wang J. Joe G. Hexner E. Choi Y. Taichman R. Emerson S. Blood. 2007; 109: 3706-3712Crossref PubMed Scopus (259) Google Scholar). Other hematopoietic-relevant molecules, such as IL-6, are expressed by human osteoblasts and contribute to their support of in vitro hematopoiesis (Taichman and Emerson, 1994Taichman R. Emerson S. J. Exp. Med. 1994; 1: 1677-1682Crossref Scopus (289) Google Scholar, Taichman et al., 1997Taichman R. Reilly M. Verma R. Emerson S. Blood. 1997; 15: 1165-1172Google Scholar). Interestingly, secretion of certain cytokines such as IL-6 and GM-CSF appears to be dynamically regulated. Stimulation of osteoblasts with parathyroid hormone (PTH) increases their production of IL-6 and GM-CSF. Additionally, it was found that coculturing primary cells enriched for HSCs with osteoblasts increased the secretion of leukemia inhibitory factor, IL-6, and stromal-derived factor 1 (SDF-1) by osteoblasts (Jung et al., 2008Jung Y. Song J. Shiozawa Y. Wang J. Wang Z. Williams B. Havens A. Schneider A. Ge C. Franceschi R. et al.Stem Cells. 2008; 26: 2042-2051Crossref PubMed Scopus (125) Google Scholar, Taichman et al., 1997Taichman R. Reilly M. Verma R. Emerson S. Blood. 1997; 15: 1165-1172Google Scholar). This principle of retrograde HSC-to-osteoblast communication was recently reinforced by Jung et al., 2008Jung Y. Song J. Shiozawa Y. Wang J. Wang Z. Williams B. Havens A. Schneider A. Ge C. Franceschi R. et al.Stem Cells. 2008; 26: 2042-2051Crossref PubMed Scopus (125) Google Scholar, who found that HSPCs could increase the proliferation and differentiation of BMSCs into osteoblastic cells through secretion of bone morphogenic proteins. In vivo data using neonatal vertebral bone transplantation supported the finding, and data from aging or myeloablated mice demonstrate the dynamic nature of this phenomenon (Jung et al., 2008Jung Y. Song J. Shiozawa Y. Wang J. Wang Z. Williams B. Havens A. Schneider A. Ge C. Franceschi R. et al.Stem Cells. 2008; 26: 2042-2051Crossref PubMed Scopus (125) Google Scholar). Osteoblasts are also required for B lymphopoiesis. Taking advantage of an osteoblast-specific thymidine kinase linked to 2.3KbCol1A1, which permits targeted destruction of osteoblasts following ganciclovir administration, Visnjic et al., 2004Visnjic D. Kalajzic Z. Rowe D. Katavic V. Lorenzo J. Aguila H. Blood. 2004; 103: 3258-3264Crossref PubMed Scopus (558) Google Scholar found a complete loss of B cells 25 days after drug treatment. Expanding on those preliminary results, Zhu et al., 2007Zhu J. Garrett R. Jung Y. Zhang Y. Kim N. Wang J. Joe G. Hexner E. Choi Y. Taichman R. Emerson S. Blood. 2007; 109: 3706-3712Crossref PubMed Scopus (259) Google Scholar demonstrated that pre-pro-B and pro-B cells are reduced by greater than 60% and 80%, respectively, an effect that precedes loss of HSCs. In addition, osteoblasts were able to support the generation of all stages of B lymphopoiesis, including the sIgM+ immature B cell stage, in vitro (Zhu et al., 2007Zhu J. Garrett R. Jung Y. Zhang Y. Kim N. Wang J. Joe G. Hexner E. Choi Y. Taichman R. Emerson S. Blood. 2007; 109: 3706-3712Crossref PubMed Scopus (259) Google Scholar). The necessity for specific adhesion and signaling molecules such as VCAM-1, α4 integrin, and IL-7 was demonstrated. Interestingly, while supplementation with PTH was not necessary for osteoblast-supported in vitro B lymphopoiesis, it was found to be stimulatory, possibly working through some combination of SDF-1 and IL-7 (Zhu et al., 2007Zhu J. Garrett R. Jung Y. Zhang Y. Kim N. Wang J. Joe G. Hexner E. Choi Y. Taichman R. Emerson S. Blood. 2007; 109: 3706-3712Crossref PubMed Scopus (259) Google Scholar). The absolute requirement for osteoblasts in B lymphopoiesis was recently confirmed using the Osterix promoter to delete Gsα, a downstream mediator of PTH/PTH-related peptide receptor (PPR), resulting in a loss of early B cell precursors including pro-B cells and a concomitant loss of IL-7 production. Interestingly, the numbers of pre-pro-B cells were normal, indicating a role for PTH in the transition to the pro-B stage, but not earlier (Wu et al., 2008Wu J. Purton L. Rodda S. Chen M. Weinstein L. McMahon A. Scadden D. Kronenberg H. Proc. Natl. Acad. Sci. USA. 2008; 4: 16976-16981Crossref Scopus (174) Google Scholar). Together, these three studies show a critical role for osteoblasts during in vivo B lymphopoiesis, and the ability for osteoblasts to support in vitro B lymphopoiesis. While osteoblasts are the site of HSC differentiation through the earliest stages of B lymphopoiesis, later maturation appears to occur preferentially at bone marrow sinusoids, specialized thin-walled blood vessels. Recent data show that the homing to and engraftment in the sinusoidal niche requires the action of cannabinoid receptor 2 (CB2), another G protein-coupled receptor, and interaction between α4 integrin and its receptor VCAM-1 (Pereira et al., 2009Pereira J. An J. Xu Y. Huang Y. Cyster J. Nat. Immunol. 2009; 10: 403-411Crossref PubMed Scopus (153) Google Scholar). Interruption of CB2 via gene knockout or receptor antagonism reduced the retention time in the bone marrow sinusoids and resulted in a decreased number of λ+ B cells in the periphery, interpreted as increased cell death due to inappropriate repertoire development/selection (Pereira et al., 2009Pereira J. An J. Xu Y. Huang Y. Cyster J. Nat. Immunol. 2009; 10: 403-411Crossref PubMed Scopus (153) Google Scholar). The above studies demonstrate the importance of multiple niches involved at different stages of a common developmental program and may prove to be important in understanding other hematopoietic niches in the bone marrow. Whereas evidence for osteoblasts' roles in B lymphopoiesis and granulopoiesis is clear and consistent, what do we know about the role of osteoblasts in HSC maintenance, self-renewal, and proliferation? As stated earlier, the hypothesis that osteoblasts represent an important regulatory component of HSC biology is not new, but recent experiments demonstrating that increasing the numbers of osteoblasts, either through inactivation of the BMP receptor type 1A (BMPR1A) or constitutive activation of the PPR, leads to a concomitant increase in the number of HSCs, and that donor HSCs can be detected adjacent to osteoblasts (Calvi et al., 2003Calvi L. Adams G. Weibrecht K. Weber J. Olson D. Knight M. Martin R. Schipani E. Divieti P. Bringhurst F. et al.Nature. 2003; 425: 841-846Crossref PubMed Scopus (2669) Google Scholar, Zhang et al., 2003Zhang J. Niu C. Ye L. Huang H. He X. Tong W. Ross J. Haug J. Johnson T. Feng J. et al.Nature. 2003; 425: 836-841Crossref PubMed Scopus (2282) Google Scholar), has implicated osteoblasts as the site of HSC homing and survival. Much has been written about these two studies, and we believe an in-depth discussion in this document is not necessary. However, we note that these data should be interpreted in light of the particulars of the experimental systems used. For example, in order to introduce constitutively active PPR to osteoblasts, Calvi et al., 2003Calvi L. Adams G. Weibrecht K. Weber J. Olson D. Knight M. Martin R. Schipani E. Divieti P. Bringhurst F. et al.Nature. 2003; 425: 841-846Crossref PubMed Scopus (2669) Google Scholar used 2.3KbCol1A1, which potentially targets more immature osteoprogenitors. Zhang et al., 2003Zhang J. Niu C. Ye L. Huang H. He X. Tong W. Ross J. Haug J. Johnson T. Feng J. et al.Nature. 2003; 425: 836-841Crossref PubMed Scopus (2282) Google Scholar knocked out BMPRIA using an inducible Mx1-Cre system, arguing that the nonspecificity of the promoter did not confound their results because they could not find detectable levels of BMPR1A mRNA in HSCs. Since then, others have found that human CD34+ bone marrow cells do express BMPR1A and that signaling through this receptor had some, but limited, capacity to regulate the numbers of primitive HSPC in vitro (Grassinger et al., 2007Grassinger J. Simon M. Mueller G. Drewel D. Andreesen R. Hennemann B. Cytokine. 2007; 40: 165-171Crossref PubMed Scopus (18) Google Scholar). This caveat does not rule out the possibility that a relationship does exist between the numbers of osteoblasts and the numbers of HSCs. In fact, our lab and others found that while losses were detected in the lymphocytic and myelocytic compartments very early on following in vivo depletion of osteoblasts, there was an eventual loss of LSK cells, which include HSCs, from the marrow of those mice after 28 days. Additionally, osteoblasts were able to support LSK cells in vivo, with the addition of external factors including SCF and interleukins (Visnjic et al., 2004Visnjic D. Kalajzic Z. Rowe D. Katavic V. Lorenzo J. Aguila H. Blood. 2004; 103: 3258-3264Crossref PubMed Scopus (558) Google Scholar, Zhu et al., 2007Zhu J. Garrett R. Jung Y. Zhang Y. Kim N. Wang J. Joe G. Hexner E. Choi Y. Taichman R. Emerson S. Blood. 2007; 109: 3706-3712Crossref PubMed Scopus (259) Google Scholar). Very recently, two papers have used similar approaches to identify where purified Flk2− LSKs home in the bone marrow (Lo Celso et al., 2009Lo Celso C. Fleming H. Wu J. Zhao C. Miake-Lye S. Fujisaki J. Côté D. Rowe D. Lin C. Scadden D. Nature. 2009; 457: 92-96Crossref PubMed Scopus (634) Google Scholar, Xie et al., 2009Xie Y. Yin T. Wiegraebe W. He X. Miller D. Stark D. Perko K. Alexander R. Schwartz J. Grindley J. et al.Nature. 2009; 457: 97-101Crossref PubMed Scopus (402) Google Scholar). In the case of Lo Celso et al., 2009Lo Celso C. Fleming H. Wu J. Zhao C. Miake-Lye S. Fujisaki J. Côté D. Rowe D. Lin C. Scadden D. Nature. 2009; 457: 92-96Crossref PubMed Scopus (634) Google Scholar, the entire experiment was performed in vivo with live animal imaging of calvaria. Xie et al., 2009Xie Y. Yin T. Wiegraebe W. He X. Miller D. Stark D. Perko K. Alexander R. Schwartz J. Grindley J. et al.Nature. 2009; 457: 97-101Crossref PubMed Scopus (402) Google Scholar harvested leg bones shortly after transferring HSCs and imaged them in semisolid media. In both studies, imaging occurred within 5 hr of HSC transfer, and both irradiated and nonirradiated mice were used as recipients. Neither study attempted to identify the location of host HSCs, as had been done previously (Kiel et al., 2005Kiel M. Yilmaz O. Iwashita T. Yilmaz O. Terhorst C. Morrison S. Cell. 2005; 121: 1109-1121Abstract Full Text Full Text PDF PubMed Scopus (2255) Google Scholar). Striking similarities were found between the two studies regarding to where HSCs home under normal and myeloablated scenarios. In both cases, HSCs homed to areas away from the endosteum and closer to vascular endothelium in nonirradiated mice. Lo Celso et al., 2009Lo Celso C. Fleming H. Wu J. Zhao C. Miake-Lye S. Fujisaki J. Côté D. Rowe D. Lin C. Scadden D. Nature. 2009; 457: 92-96Crossref PubMed Scopus (634) Google Scholar found that after 5 hr, donor HSCs were 0–16 μm from vasculature but more than 30 μm from the endosteum. Likewise, in the second study, 25 of 32 (78%) cells counted were located more than four cell depths (defined by the investigators as one cell depth = 6–8 μm) from the endosteum. No information was given regarding proximity to vasculature, but it was noted that the vasculature was distributed throughout the bone marrow cavity, including at the endosteum (Xie et al., 2009Xie Y. Yin T. Wiegraebe W. He X. Miller D. Stark D. Perko K. Alexander R. Schwartz J. Grindley J. et al.Nature. 2009; 457: 97-101Crossref PubMed Scopus (402) Google Scholar). Likewise, Lo Celso et al., 2009Lo Celso C. Fleming H. Wu J. Zhao C. Miake-Lye S. Fujisaki J. Côté D. Rowe D. Lin C. Scadden D. Nature. 2009; 457: 92-96Crossref PubMed Scopus (634) Google Scholar found that there is extensive vascularization of the endosteum, with 60% and 90% of osteoblasts within 10 or 20 μm of vasculature, respectively. Hooper et al., 2009Hooper A. Butler J. Nolan D. Kranz A. Iida K. Kobayashi M. Kopp H. Shido K. Petit I. Yanger K. et al.Cell Stem Cell. 2009; 4: 263-274Abstract Full Text Full Text PDF PubMed Scopus (420) Google Scholar imaged VEGFR3+ sinusoidal endothelium in very close proximity to osteoblasts. These studies agree with a previously published report identifying HSCs in untreated mice as being adjacent to sinusoidal endothelium. Using the refined HSC phenotype of Lin−CD41−CD48−CD150+, only 14% of HSCs were found in close contact with the endosteum, while 60% were in contact with sinusoidal endothelium in the bone marrow, and 62% were in contact with splenic sinusoids. Additionally, these areas were significantly enriched for HSCs over other hematopoietic populations (Kiel et al., 2005Kiel M. Yilmaz O. Iwashita T. Yilmaz O. Terhorst C. Morrison S. Cell. 2005; 121: 1109-1121Abstract Full Text Full Text PDF PubMed Scopus (2255) Google Scholar). Following irradiation and destruction of vascular sinusoids, both studies noted that donor HSCs homed closer to the endosteum. In the first study, fluorescently labeled HSCs were located between 0 and 40 μm of the endosteum when transferred to irradiated recipients (Lo Celso et al., 2009Lo Celso C. Fleming H. Wu J. Zhao C. Miake-Lye S. Fujisaki J. Côté D. Rowe D. Lin C. Scadden D. Nature. 2009; 457: 92-96Crossref PubMed Scopus (634) Google Scholar). No information was given regarding changes in distance to vasculature. In the second study, 28 of 43 (65%) donor HSCs counted were now located within four cell depths of the endosteum, with only 15 of 43 (35%) farther out (Xie et al., 2009Xie Y. Yin T. Wiegraebe W. He X. Miller D. Stark D. Perko K. Alexander R. Schwartz J. Grindley J. et al.Nature. 2009; 457: 97-101Crossref PubMed Scopus (402) Google Scholar). Similar to the Lo Celso et al., 2009Lo Celso C. Fleming H. Wu J. Zhao C. Miake-Lye S. Fujisaki J. Côté D. Rowe D. Lin C. Scadden D. Nature. 2009; 457: 92-96Crossref PubMed Scopus (634) Google Scholar study, there was a more variable distribution of donor HSC endosteal distance following irradiation. A possible interpretation of these investigations is that under steady-state conditions, the physiological HSC niche exists at the heavily vascularized endosteum but is saturated with host HSCs, forcing transplanted cells to adhere to alternate niches, including the immature B cell niche identified in the sinusoids (Pereira et al., 2009Pereira J. An J. Xu Y. Huang Y. Cyster J. Nat. Immunol. 2009; 10: 403-411Crossref PubMed Scopus (153) Google Scholar). Following irradiation and the destruction of host HSCs, these endosteal niches are available for engraftment by transplants. Notably, Lo Celso et al. also demonstrated that HSCs transplanted into WWv recipients, which do not require irradiation to yield high chimerism, localized close to osteoblasts, correlating homing distances with the degree of donor cell reconstitution. Due to the strong evidence that HSCs are located adjacent to sinusoids and away from the endosteum at steady state, it is more likely that transplanted cells home to the perivascular location under nonirradiated conditions because it represents the preferred microenvironment. Due to low HSC engraftment into normal nonmyeloablated hosts, and a normally low frequency of HSCs entering the cell cycle, transplanted cells are incapable of generating detectable blood cell chimerism. Under this hypothesis, paraosteoblastic niches normally occupied by early B cell progenitors may also serve as sites for HSC homing and proliferation following irradiation-induced loss of sinusoids, enabling hematopoietic recovery. It is also possible, given the extensive endosteal vascularization noted by both studies, that HSCs home to surviving sinusoids or other stromal cells near the endosteum (Lo Celso et al., 2009Lo Celso C. Fleming H. Wu J. Zhao C. Miake-Lye S. Fujisaki J. Côté D. Rowe D. Lin C. Scadden D. Nature. 2009; 457: 92-96Crossref PubMed Scopus (634) Google Scholar, Xie et al., 2009Xie Y. Yin T. Wiegraebe W. He X. Miller D. Stark D. Perko K. Alexander R. Schwartz J. Grindley J. et al.Nature. 2009; 457: 97-101Crossref PubMed Scopus (402) Google Scholar). Slayton et al., 2007Slayton W. Li X. Butler J. Guthrie S. Jorgensen M. Wingard J. Scott E. Stem Cells. 2007; 25: 2945-2955Crossref PubMed Scopus (48) Google Scholar found that SDF-1, a potent HSC chemokine, was expressed at the endosteum following lethal radiation, providing a plausible mechanism for this transit. It would be of keen interest to know the location of surviving HSCs following sublethal irradiation and how signaling to and from osteoblasts changes following irradiation. HSCs have been imaged residing adjacent to bone marrow sinusoids (Kiel et al., 2005Kiel M. Yilmaz O. Iwashita T. Yilmaz O. Terhorst C. Morrison S. Cell. 2005; 121: 1109-1121Abstract Full Text Full Text PDF PubMed Scopus (2255) Google Scholar). Bone marrow sinusoids are single-cell walled structures with discontinuous pericytes and basal lamina, void of tight junctions, and which receive some of their structural support from the hematopoietic marrow, making them sensitive to treatments that cause a rapid loss of bone marrow cellularity. As early as 7 days postirradiation, bone marrow sinusoids are enlarged and irregular in shape, with extensive intramarrow hemorrhage and limited sinusoidal endothelial cell (SEC) DNA damage and necrosis (Li et al., 2008Li X. Hu Z. Jorgenson M. Wingard J. Slayton W. Exp. Hematol. 2008; 36: 1143-1156Abstract Full Text Full Text PDF PubMed Scopus (49) Google Scholar, Slayton et al., 2007Slayton W. Li X. Butler J. Guthrie S. Jorgensen M. Wingard J. Scott E. Stem Cells. 2007; 25: 2945-2955Crossref PubMed Scopus (48) Google Scholar). Repopulation and repair of the sinusoids results from the proliferation of surviving SECs, a process that begins shortly after irradiation and is accelerated by infusion of endothelial progenitor cells (EPCs) (Li et al., 2008Li X. Hu Z. Jorgenson M. Wingard J. Slayton W. Exp. Hematol. 2008; 36: 1143-1156Abstract Full Text Full Text PDF PubMed Scopus (49) Google Scholar, Salter et al., 2009Salter A. Meadows S. Muramoto G. Himburg H. Doan P. Daher P. Russell L. Chen B. Chao N. Chute J. Blood. 2009; 113: 2104-2107Crossref PubMed Scopus (110) Google Scholar, Slayton et al., 2007Slayton W. Li X. Butler J. Guthrie S. Jorgensen M. Wingard J. Scott E. Stem Cells. 2007; 25: 2945-2955Crossref PubMed Scopus (48) Google Scholar). Intriguingly, EPC infusion also accelerated hematopoietic recovery in irradiated mice, as evidenced by near-normal numbers of colony-forming cells (CFCs) and a notable elevation in the number of phenotypic and functional HSCs (Salter et al., 2009Salter A. Meadows S. Muramoto G. Himburg H. Doan P. Daher P. Russell L. Chen B. Chao N. Chute J. Blood. 2009; 113: 2104-2107Crossref PubMed Scopus (110) Google Scholar). A complementary study blocking angiogenesis using antibodies directed against VEGFR2 showed that concomitant with reduced restoration of the bone marrow sinusoids was a near-total block of donor HSPC engraftment and hematopoietic bone marrow recovery, as measured by flow cytometry (Hooper et al., 2009Hooper A. Butler J. Nolan D. Kranz A. Iida K. Kobayashi M. Kopp H. Shido K. Petit I. Yanger K. et al.Cell Stem Cell. 2009; 4: 263-274Abstract Full Text Full Text PDF PubMed Scopus (420) Google Scholar). These combined data clearly link hematopoietic recovery with sinusoidal recovery. The positive relationship between BMSC and the HME was first demonstrated more than 30 years ago (Friedenstein et al., 1974Friedenstein A. Chailakhyan R. Latsinik N. Panasyuk A. Keiliss-Borok I. Transplantation. 1974; 17: 331-340Crossref PubMed Scopus (1023) Google Scholar). Some of the most exciting recent investigations have looked at the ability of specific cell populations isolated from bone to establish the HME. These studies collectively show that the appearance of hematopoietic cells in the bone marrow first requires the assembly of sinusoids, a process that is, at least in part, controlled by osteoprogenitors. Subcutaneous transplantation of human CD45−CD146+ (additional markers include CD105+ alkaline phosphatase+ angiopoietin-1+) colony-forming unit fibroblasts (CFU-Fs), which appeared in situ as subendothelial reticular fibroblasts, into mice resulted in the formation of ectopic bone, complete with sinusoids and hematopoietic cells (Sacchetti et al., 2007Sacchetti B. Funari A. Michienzi S. Di Cesare S. Piersanti S. Saggio I. Tagliafico E. Ferrari S. Robey P. Riminucci M. Bianco P. Cell. 2007; 131: 324-336Abstract Full Text Full Text PDF PubMed Scopus (1557) Google Scholar). Detection of donor-derived bone and fibroblastic tissue preceeded formation of host-derived vascular sinusoids. Shortly thereafter, hematopoietic cells were located in the area surrounding these sinusoids (Sacchetti et al., 2007Sacchetti B. Funari A. Michienzi S. Di Cesare S. Piersanti S. Saggio I. Tagliafico E. Ferrari S. Robey P. Riminucci M. Bianco P. Cell. 2007; 131: 324-336Abstract Full Text Full Text PDF PubMed Scopus (1557) Google Scholar). Interestingly, a small population of CD146+ reticular cells, functional in secondary transplantations, was located adjacent to the sinusoids. Although no attempt was made to determine whether HSCs were supported, the investigators noted that a rare population of hematopoietic cells expressing Tie-2, the ligand for angiopoietin-1, and expressed by HSCs, was adjacent to CD146+ reticular cells (Sacchetti et al., 2007Sacchetti B. Funari A. Michienzi S. Di Cesare S. Piersanti S. Saggio I. Tagliafico E. Ferrari S. Robey P. Riminucci M. Bianco P. Cell. 2007; 131: 324-336Abstract Full Text Full Text PDF PubMed Scopus (1557) Google Scholar). More recently, Chan et al., 2009Chan C. Chen C. Luppen C. Kim J. DeBoer A. Wei K. Helms J. Kuo C. Kraft D. Weissman I. Nature. 2009; 457: 490-494Crossref PubMed Scopus (301) Google Scholar identified a similar CD105+ population of murine stromal cells. When transplanted underneath the renal capsule, these cells were capable of generating an ectopic HME, including donor-derived bone formed through a cartilaginous intermediate, host-derived vasculature, and hematopoietic cells. Unlike Sacchetti et al., 2007Sacchetti B. Funari A. Michienzi S. Di Cesare S. Piersanti S. Saggio I. Tagliafico E. Ferrari S. Robey P. Riminucci M. Bianco P. Cell. 2007; 131: 324-336Abstract Full Text Full Text PDF PubMed Scopus (1557) Google Scholar, the investigators confirmed the presence of HSCs but did not identify their location (Chan et al., 2009Chan C. Chen C. Luppen C. Kim J. DeBoer A. Wei K. Helms J. Kuo C. Kraft D. Weissman I. Nature. 2009; 457: 490-494Crossref PubMed Scopus (301) Google Scholar). Gene-silencing experiments revealed that while SCF production by the donor-derived elements was not necessary for the initiation and maintenance of hematopoiesis, Osterix, associated with endochondral ossification, was (Chan et al., 2009Chan C. Chen C. Luppen C. Kim J. DeBoer A. Wei K. Helms J. Kuo C. Kraft D. Weissman I. Nature. 2009; 457: 490-494Crossref PubMed Scopus (301) Google Scholar). As was found following irradiation-induced damage, VEGFR2-mediated signaling is required for the formation of bone marrow sinusoids (Chan et al., 2009Chan C. Chen C. Luppen C. Kim J. DeBoer A. Wei K. Helms J. Kuo C. Kraft D. Weissman I. Nature. 2009; 457: 490-494Crossref PubMed Scopus (301) Google Scholar, Hooper et al., 2009Hooper A. Butler J. Nolan D. Kranz A. Iida K. Kobayashi M. Kopp H. Shido K. Petit I. Yanger K. et al.Cell Stem Cell. 2009; 4: 263-274Abstract Full Text Full Text PDF PubMed Scopus (420) Google Scholar). Unfortunately, no attempt was made to identify whether the CD105+ cells represented a self-renewing stem cell, as found by Sacchetti et al., 2007Sacchetti B. Funari A. Michienzi S. Di Cesare S. Piersanti S. Saggio I. Tagliafico E. Ferrari S. Robey P. Riminucci M. Bianco P. Cell. 2007; 131: 324-336Abstract Full Text Full Text PDF PubMed Scopus (1557) Google Scholar. In both studies, bone formation preceded vascularization, which itself was necessary for population by hematopoietic cells. Intriguingly, mature cells populated this new HME before HSCs, suggesting that structurally sound sinusoids, receiving support from hematopoietic marrow, are a prerequisite for HSC colonization. In addition, both studies found that generation of bone alone was not sufficient for the establishment of hematopoiesis (Chan et al., 2009Chan C. Chen C. Luppen C. Kim J. DeBoer A. Wei K. Helms J. Kuo C. Kraft D. Weissman I. Nature. 2009; 457: 490-494Crossref PubMed Scopus (301) Google Scholar, Sacchetti et al., 2007Sacchetti B. Funari A. Michienzi S. Di Cesare S. Piersanti S. Saggio I. Tagliafico E. Ferrari S. Robey P. Riminucci M. Bianco P. Cell. 2007; 131: 324-336Abstract Full Text Full Text PDF PubMed Scopus (1557) Google Scholar). Importantly, these studies represent experimental models, which can be used to dissect the molecular signature of the HSC niche, as well as other lineage-specific niches. Combined, we believe these studies reveal a genesis of the HME that begins with a population of stromal cells constructing a proangiogenic environment that includes critical vascular sinusoids, capable of recruiting and sustaining HSPCs. Following initiation of the HME, HSCs are largely maintained at these vascular sinusoids, with osteoblasts critically supporting myelopoisis and lymphopoiesis through the production of soluble and membrane-associated factors such as G-CSF and IL-7. Additionally, osteoblasts provide a home for surviving and transplanted HSPCs following myeloablative therapies, maintaining those cells while the sinusoidal niche regenerates (Figure 1). It is important to note that this model is but one possible interpretation of the studies discussed above. Additionally, these studies underscore that a complex array of cells with overlapping and unclear function and morphology are present in the marrow and may also perform critical roles in hematopoiesis. Understanding how specific hematopoietic niches support the maintenance of the blood system under normal and stressed conditions is of both academic and practical importance. Recently, much information has been gathered using innovative approaches, such as in vivo imaging and the formation of ectopic niches. These tools and the information gathered using them will be refined and reinterpreted during the next few years. As we do this, we must be mindful of the possibility that there is not a single and simply constructed niche but that it is the totality of the bone marrow environment that is essential for hematopoiesis, with a combination of distinct and overlapping roles being performed by multiple cellular and extracellular partners. These environments are likely dynamic, altered by routinely used experimental protocols that perturb normal in vivo hematopoiesis. At present, it is clear that osteoblasts and stromal cells, endothelial cells and hematopoeitic cells interact, through combinations of adherent and soluble factors, to create as-yet ill-defined niches that support HSC self-renewal and differentiation. Future experiments, which combine genetic tools with high-resolution in vivo imaging, should shed more light on the specific interactions that define each of these microniches. We would like to thank the members of the Emerson lab, past and present, who have helped to shape this discussion. We apologize for omitting many important contributions to the field due to a strict limit on the number of references.
Referência(s)