sign in sign up
<<
Home , Hematopoietic stem cell

Oncogene (2007) 26, 6687–6696 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc REVIEW Hematopoietic developmental pathways: on cellular basis

H Iwasaki1 and K Akashi1,2

1Center for Cellular and Molecular Medicine, Kyushu University Hospital, Fukuoka, Japan and 2Department of and AIDS, Harvard Medical School, Dana-Farber Cancer Institute, Boston, MA, USA

To elucidate the molecular mechanisms underlying normal counterpart, the common myeloid progenitor (CMP) and malignant hematopoietic development, it is critical to that can be a source of all myeloid cell types, supports identify developmental intermediates for each lineage the concept that lymphoid and myeloid lineages develop downstream of hematopoietic stem cells. Recent advances independently downstream of HSCs (Kondo et al., 1997; in prospective isolation of hematopoietic stem and Akashi et al., 2000). progenitor cells, and efficient xenogeneic transplantation Recent progresses in the fluorescence-activated cell systems have provided a detailed developmental map in sorting analysis using additional surface markers for both mouse and human hematopoiesis, demonstrating that early hematopoiesis, however, have provided more surface phenotypes of mouse stem–progenitor cells and detailed developmental map downstream of HSCs. their human counterparts are considerably different. Importantly, prior to proceed into the classical myeloid Here, we summarize the phenotype and functional vs lymphoid pathways, HSCs appear to form myelo- properties and their differences of hematopoietic stem erythroid vs myelo-lymphoid progenitors (Arinobu and populations between mouse and et al., 2007). Furthermore, it is now clear that there human. are considerable differences in distribution of surface Oncogene (2007) 26, 6687–6696; doi:10.1038/sj.onc.1210754 markers between human and mouse hematopoiesis, which makes identification of human counterparts of Keywords: hematopoietic ; lineage commitment; mouse progenitors difficult. In this review, we summar- transcription factor; leukemic stem cell ize what we have learned from recent studies concerning the developmental sequences in murine and human hematopoiesis.

Introduction

Hematopoietic stem cells (HSCs) can self-renew, and Developmental pathways in murine hematopoiesis differentiate into all types of cells throughout life. Mature hematopoietic cells are traditionally categorized In murine hematopoiesis, cells with multipotent activity into two distinct lineages: the lymphoid and the myeloid. reside in a small fraction of marrow, which lacks The lymphoid lineage consists of T, B and natural killer the expression of lineage-affiliated surface markers (Lin) (NK) cells, while the myeloid lineage includes a number but expresses high levels of Sca-1 and c-Kit (Spangrude of morphologically, phenotypically and functionally et al., 1988; Ikuta and Weissman, 1992). Within the LSK distinct cell types such as different subclasses of population, the most primitive self-renewing HSCs with (, and ), long-term reconstituting activity (LT-HSCs) reside in lo À À –, erythrocytes, the Thy1.1 , CD34 , fms-like tyrosine kinase-3 (Flt3) low and mast cells. These two lineages have been thought to or Hoechst (side population; SP) fraction (Morrison use separate differentiation pathways. and Weissman, 1994; Osawa et al., 1996; Goodell et al., lo À HSCs were identified within the ‘LSK’ (LinÀSca-1 þ c-Kit þ ) 1997; Adolfsson et al., 2001). The Thy1.1 or the CD34 B population of the mouse (Spangrude LSK fraction constitutes only 0.01% of total bone B lo et al., 1988; Morrison and Weissman, 1994; Osawa marrow cells. At the single cell level, 20% of Thy1.1 B À et al., 1996), and subsequently lineage-restricted pro- or 35% of CD34 LSK cells display multilineage genitors at various developmental stages have been isolated long-term reconstitution in competitive reconstitution downstream of HSCs by using the multicolor fluorescence- assays (Osawa et al., 1996; Wagers et al., 2002). À þ þ activated cell sorting system. The successful isolation of On the other hand, the Thy1.1 , CD34 or Flt3 LSK the common lymphoid progenitor (CLP) that can generate fraction is capable of only transient reconstitution, all lymphoid types but not any myeloid cells, and its thereby contains short-term (ST) HSCs or multipotent progenitors (MPPs) (Morrison and Weissman, 1994; Osawa et al., 1996; Adolfsson et al., 2001; Christensen Correspondence: Professor K Akashi, Department of Cancer Immuno- logy and AIDS, Harvard Medical School, Dana-Farber Cancer Institute, and Weissman, 2001). Several reports have tried to Boston, MA 02115, USA. discriminate ST-HSCs and MPPs by using low or E-mail: [email protected] negative expression of CD4, CD11b, and/or Thy1.1, Hematopoietic developmental pathways H Iwasaki and K Akashi 6688 and showed some difference in duration and magnitude FcgRII/IIIloCD34 þ CMPs, FcgRII/IIIhiCD34 þ granu- of reconstitution (Morrison and Weissman, 1994; Morrison locyte– progenitors (GMPs) and FcgRII/ et al., 1997). These reports, however, did not provide IIIloCD34À –erythrocyte progenitors clear-cut definition or functional difference of ‘ST-HSC’ (MEPs) are isolatable (Akashi et al., 2000). The CMP and ‘MPP’. We will therefore use MPPs instead of can generate all types of myeloid colonies, while the ST-HSCs throughout this review. GMP or the MEP produces only macro- phage (GM) or megakaryocyte erythrocyte (MegE) The myeloid vs lymphoid pathways downstream of MPP lineage cells, respectively. The CMP differentiates into within the LSK fraction the GMP and the MEP in vitro, upregulating the Because the LSK population contains all multipotent FcgRII/III and downregulating the CD34 expression, precursors such as LT-HSCs and MPPs, lineage-restrict- respectively, indicating that the CMP is the precursor ed progenitor populations were first sought outside for the GMP and the MEP (Akashi et al., 2000). Upon the LSK fraction. The CLP is the earliest population in vivo transfer, these populations display short-term that upregulates the receptor for -7 (IL-7) production of lineages corresponding to their in vitro (Kondo et al., 1997), an essential for both activities, indicating that they do not appreciably self- T- and B-cell development (Peschon et al., 1994; von renew (Na Nakorn et al., 2002). Freeden-Jeffry et al., 1995). The IL-7 receptor signaling The existence of the CLP and the CMP outside the plays a critical role in survival through LSK fraction leads to the idea that lymphoid and maintenance of Bcl-2 (Akashi et al., 1997), and in myeloid development starts at the CLP and the CMP the rearrangement of immunoglobulin heavy chain V stage, respectively (Figure 1a). These stem and progeni- segments through the activation of the Pax5 gene tor populations are widely used for targeted analysis and (Corcoran et al., 1998). In human, the genetic loss of manipulation of cells at a specific hematopoietic stage. IL-7Ra chain is found in patients with severe combined immunodeficiency (SCID) (Puel et al., 1998). The Initial hematopoietic lineage commitment starts within the IL-7Ra þ fraction concentrates lymphoid potential in LSK ‘MPP’ fraction prior to the conventional CMP or the bone marrow, and IL-7Ra þ LinÀSca-1loc-Kitlo cells, CLP stages termed as the CLP, possess clonogenic T- and B-cell, Although the CMP and the CLP were found outside the and NK-cell potential, but lack myelo-erythroid differ- LSK fraction, recent studies have revealed that the MPP entiation activity (Kondo et al., 1997). population is heterogeneous, and contains progenitor The CMP was also isolated outside the LSK fraction. subsets committed to myelo-erythroid or myelo-lym- The IL-7RaÀ LinÀSca-1Àc-Kit þ population, which con- phoid lineages (Figures 1b and 2) (Arinobu et al., 2007). tains the majority of myelo-erythroid colony-forming The question regarding the simple myeloid vs activity in the bone marrow, is further fractionated lymphoid diversification model initially arose from the based on the expression of FcgRII/III and CD34. finding that a fraction of MPPs have primed to the Three distinct myeloid progenitor subsets such as lymphoid lineage (Iwasaki and Akashi, 2007). In mice

abCMP vs. CLP model CMP vs. GMLP model

HSC HSC

Flt3+ LMPP MPP MPP

CMP CLP CMP GMLP

MEP GMP proT proB MEP GMP CLP

proT proB

Figure 1 Cellular pathways in adult murine hematopoiesis based on prospective purification of lineage-restricted progenitors. (a) The conventional myeloid vs lymphoid developmental pathways. (b) The new developmental pathway with integration of newly identified GMLP within the MPP population. CLP, common lymphoid progenitor; CMP, common myeloid progenitor; GMLP, granulocyte– monocyte–lymphoid progenitor; GMP, granulocyte–monocyte progenitor; LMPP, lymphoid-primed multipotent progenitor; MEP, megakaryocyte–erythrocyte progenitor; MPP, multipotent progenitor.

Oncogene Hematopoietic developmental pathways H Iwasaki and K Akashi 6689 percent of ELPs give rise to GM and rarely MegE colonies. In our hands, the ELP resides within the CD34 þ

HSC CD34 - MPP population (Osawa et al., 1996), and almost 40% of ELPs can form GM but not MegE colonies (Arinobu et al., 2007). These data suggest that the ELP population contains cells with GM and lymphoid but not MegE lineage potential. Additional evidence for the myelo-lymphoid commit- MPP CD34 + ment within the MPP fraction has been shown by fractionating LSK cells by the expression of Flt3, an HSC-specific receptor tyrosine kinase (Mackarehtschian et al., 1995). Flt3-deficient mice display loss or reduction Flt3 + of early T, B, NK and dendritic cells (Mackarehtschian CMP et al., 1995; McKenna et al., 2000; Sitnicka et al., 2002). Mice deficient for Flt3 ligand (FL) lack CLPs, but GATA-1+ PU.1+ possess normal numbers of CMPs (Adolfsson et al., 2001), suggesting that Flt3 signaling plays a critical role RAG1+ in murine lymphoid development. In murine hemato- poiesis, Flt3 is not expressed in LT-HSCs, but is upregulated in a fraction of CD34 þ MPPs (Adolfsson et al., 2001; Christensen and Weissman, 2001). A recent report has shown that Flt3 þ MPPs largely lack MegE potential, whereas they have clonal and robust GM/ lymphoid potential (Adolfsson et al., 2005), further supporting the concept that LSK cells initially commit to the myelo-lymphoid lineage. This population expres- MEP GMP CLP ses lymphoid-related Flt3 although still retaining the GM potential, and therefore was termed as the lymphoid- primed multipotent progenitor (LMPP) (Adolfsson et al., 2005). Although a minor fraction of the LMPP may still retain MegE potential as reported (Forsberg et al., Erythrocyte 2006), the vast majority of the LMPP appears to be Megakaryocyte Monocyte committed into the GM/lymphoid lineage. More recently, further fractionation of the LMPP population was performed by utilizing the expression of Dendritic cell vascular cell adhesion molecule-1 (VCAM-1) (Lai and Kondo, 2006). The Thy1.1À LSK population, which is Figure 2 Heterogeneity of the murine ‘MPP’ population. The CD34 þ almost equal to the CD34 þ MPPs, contains cells LSK population, termed as the MPP, is composed of multiple expressing Flt3. The Flt3 þ LMPP is divided into progenitor populations. The LMPP is the MPP population þ À expressing Flt3, and is largely committed to the myelo-lymphoid VCAM-1 and VCAM-1 populations. Myelo-ery- lineage. The upregulation of GATA-1 or PU.1 marks the lineage throid potential are gradually lost as LMPPs lose the specification of the MPP into the myelo-erythroid or the myelo- VCAM-1 expression: VCAM-1À LMPPs display strong lymphoid lineages, respectively. GATA-1 þ MPPs are functional þ þ CLP activity in vivo, but similar to the ELP, around CMPs, whereas PU.1 MPPs are functional GMLPs. RAG1 À ELPs are committed to the lymphoid lineage but retaining a minor 10% of VCAM-1 LMPPs give rise to GM, but not GM potential. The LMPP contains PU.1 þ GMLPs and RAG1 þ MegE colonies (Lai and Kondo, 2006). Thus, like ELPs, ELPs, and these progenitor populations might represent continuous the majority of VCAM-1À LMPPs have committed to developmental steps from the MPP to the lymphoid-committed the lymphoid lineage, but a fraction of them likely have CLP. CMP, common myeloid progenitor; ELP, earliest GM/lymphoid potential. This study describes the progenitor; GMLP, granulocyte–monocyte–lymphoid progenitor; HSC, ; LMPP, lymphoid-primed multi- heterogeneity of the LMPP, but basically supports the potent progenitor; MMP, multipotent progenitor. concept that toward the lymphoid lineage development, the loss of myeloid potential occurs first in the MegE and then in the GM lineage (Akashi et al., 2005). carrying a green fluorescence protein (GFP) gene knocked into the Rag1 gene locus (Kuwata et al., Myelo-erythroid vs myelo-lymphoid diversification 1999; Igarashi et al., 2001), a fraction (B5%) of LSK at the ‘MPP’ stage can be visualized by utilizing cells express GFP (Igarashi et al., 2002). This popula- transcription-factor reporters tion, called as the earliest lymphocyte progenitor (ELP), The limitation of the surface-marker-based fractiona- possesses potent T-, B- and NK-differentiation potential tion is that the upregulation or downregulation of with a weak myeloid colony-forming activity (Igarashi surface markers should occur as a result of lineage et al., 2002). The original report showed that a few decision: Lineage-restricted progenitors have been

Oncogene Hematopoietic developmental pathways H Iwasaki and K Akashi 6690 purified based on the difference in the expression level of progeny as compared to those by the same number of antigens that may not have significant functions in the original Sca-1À CMPs, and are capable of differ- lineage fate decision, such as Sca-1, CD34 and Thy1. entiation into the original Sca-1À CMPs as well as Recent data have shown that more rigorously com- GMPs and MEPs, confirming that GATA-1 þ Sca-1lo mitted progenitors are isolatable by the aid of transcrip- MPPs is the earliest, potent CMP within the LSK tion-factor reporters. A high-resolution map containing fraction. Thus, the CD34 þ MPP population expressing new lineage-restricted progenitor populations within the PU.1 or GATA-1 reporters represents granulocyte– LSK MPP fraction has been obtained by utilizing mice monocyte–lymphoid progenitors (GMLPs) or CMPs, having PU.1 or GATA-1 transcriptional factor reporters respectively (Arinobu et al., 2007). (Figure 2) (Arinobu et al., 2007). The new developmental model is schematized in PU.1 and GATA-1 transcription factors exert in- Figure 1b. In this model, HSCs express low PU.1, structive signals for GM and MegE lineage commit- perhaps to maintain self-renewal activity (Iwasaki et al., ment, respectively (Kulessa et al., 1995; Nerlov and 2005b). The subsequent upregulation of GATA-1 occurs Graf, 1998; Seshasayee et al., 1998; Heyworth et al., in a fraction of CD34 þ MPPs to establish the bipotent 2002; Iwasaki et al., 2003). PU.1 transactivates a CMP stage where GATA-1 and PU.1 are primed to play number of GM- and lymphoid-related genes (Zhang a competitive interaction (Miyamoto et al., 2002), et al., 1996; DeKoter et al., 1998, 2002; Iwama et al., whereas the upregulation of PU.1 results in the 1998), and is necessary for generation of early GM and formation of the GMLP. Then, when PU.1 becomes lymphoid progenitors such as CMPs, GMPs and CLPs more dominant, the CMP and the GMLP (or the (Iwasaki et al., 2005b), whereas GATA-1 is an essential LMPP) give rise to GMPs plus CLPs, whereas when transcription factor for MegE development (Fujiwara GATA-1 becomes dominant, CMPs produce MEPs at et al., 1996). Furthermore, PU.1 and GATA-1 mutually the expense of GMPs. Thus, the reciprocal activation of inhibit each other’s expression and transactivation GATA-1 and PU.1 primarily organizes hematopoietic functions (Nerlov and Graf, 1998; Rekhtman et al., lineage fate decision to form the earliest hematopoietic 1999; Zhang et al., 1999, 2000; Nerlov et al., 2000; branchpoint that comprises isolatable myelo-erythroid Walsh et al., 2002). On the basis of these findings, the and myelo-lymphoid progenitor populations (Arinobu competitive interplay of PU.1 and GATA-1 might play et al., 2007). a critical role in early hematopoietic fate decision such Importantly, the new model allows flexibility for GM as the GM or lymphoid vs the MegE lineage commit- development as the GMP can be generated from both ment (Graf, 2002; Iwasaki and Akashi, 2007). the GMLP and the CMP stages within the LSK The expression of PU.1 and GATA-1 reporters within fraction. The GMP is the source of a number of myeloid the LSK fraction was tested together with the expression subclasses because it is capable of producing all of Flt3, the critical marker for the LMPP (Arinobu granulocyte progenitors such as eosinophil progenitors, et al., 2007). Murine long-term HSCs do not express basophil progenitors, basophil–mast cell progenitors Flt3 (Adolfsson et al., 2001; Christensen and Weissman, (BMCPs) and mast cell progenitors (MCPs) (Arinobu 2001). In a mouse line harboring the GFP gene knocked et al., 2005; Iwasaki et al., 2005a, 2006). The detailed into the PU.1 locus (Back et al., 2005), CD34À long-term developmental scheme and its molecular mechanisms HSCs are entirely Flt3À and PU.1-GFPlo (PU.1lo) downstream of the GMPs are summarized in our recent (Iwasaki et al., 2005b). A high level of PU.1-GFP is review (Iwasaki and Akashi, 2007). expressed in a fraction of CD34 þ MPPs expressing a high level of Flt3, suggesting that this population represents the earliest hematopoietic stage initiating The murine lymphoid cells develop via multiple lymphoid PU.1 upregulation. In contrast, in a mouse line having progenitor subsets the transgenic GATA-1 reporter tagged with GFP The existence of a series of GM–lymphoid progenitors (Iwasaki et al., 2005a), GATA-1-GFP þ (GATA-1 þ ) such as the LMPP, the PU.1 þ Sca-1lo GMLP and the cells are found only in a fraction of CD34 þ MPPs that RAG1 þ ELP within the CD34 þ Flt3 þ MPP fraction do not express Flt3. Thus, the initial upregulation of strongly suggest that they may represent continuous steps PU.1 or GATA-1 occurs independently at the CD34 þ for HSCs to differentiate into the lymphoid lineage where MPP stage in Flt3 þ or Flt3À sub-populations, respec- cells progressively lose MegE than GM potential. The tively. Both populations express Sca-1 at lower levels as progenitor population that is completely committed to compared to the Sca-1 level in CD34À (Flt3À) LT-HSCs the lymphoid lineage is marked by the expression of (Arinobu et al., 2007). measurable amount of IL-7Ra protein on the cell surface. Almost all single PU.1 þ Sca-1lo MPPs form GM The CLP has the IL-7Ra þ LinÀSca-1loc-Kitlo phenotype colonies but not MegE colonies. This population gives (Kondo et al., 1997). Although the majority of CLP rise to GM and lymphoid cells at the single cell level, activity resides in cells within the originally described and differentiates into the IL-7Ra þ CLP (Kondo et al., IL-7Raþ LinÀSca-1loc-Kitlo fraction, similar activity is found 1997) and the GMP (Akashi et al., 2000). In contrast, within B220 þ CD19ÀCD43 þ proB cell fraction tran- the GATA-1 þ Sca-1lo MPPs give rise to myelo-erythroid scribing pTa. The IL-7Ra þ pTa þ B220 þ CD19Àc-KitÀ colonies at a high efficiency, but are incapable of population, called ‘CLP-2’, is isolatable in mice harbor- generating lymphoid cells. GATA-1 þ Sca-1lo MPPs ing the human (h) CD25-pTa transgenic reporter produce >20-fold higher number of GM or MegE (Gounari et al., 2002; Martin et al., 2003).

Oncogene Hematopoietic developmental pathways H Iwasaki and K Akashi 6691 RAG1-expressing ELPs still possess a minor GM the CLP and the ELP probably occurs at the GMLP potential, suggesting that ELPs have just committed to stage. The developmental scheme for the lymphoid the lymphoid lineage. In addition, ELPs express a high lineage based on these data is shown in Figure 3. level of c-Kit, whereas CLPs express only a low level. Upon seeding the , bone marrow progenitors ELPs have IL-7Ra mRNA but not a measurable level of have to enter in the circulation. The LMPP, the ELP IL-7Ra protein (Igarashi et al., 2002), whereas CLPs and the CLP are detectable in peripheral blood, whereas and CLP-2 are isolated based on a high level of IL-7Ra the CLP-2 has not been found in the blood (Schwarz expression. In the steady-state thymus, the c-Kit þ early and Bhandoola, 2004; Perry et al., 2006; Krueger and T-cell progenitor fraction (Allman et al., 2003) is the von Boehmer, 2007; Umland et al., 2007). Recently, a earliest thymic precursor (ETP), and this population novel T-lineage-committed progenitor subset presum- express only negative to low levels of IL-7Ra. These ably downstream of the CLP-2 was found in adult data have been interpreted that the ELP is upstream of murine blood. This circulating T-cell progenitor is the CLP and the CLP-2 (Bhandoola and Sambandam, isolatable by unique phenotype of LinÀhCD25/ 2006). pTa þ B220À (Sca-1 þ c-KitloThy1hiFlt3ÀIL-7Ra þ ), and is To formally test the lineal relationship between the distinguishable from other lymphoid progenitors (for ELP and the CLP, the history of RAG1 activation is example, LMPP, ELP, CLP or CLP-2) by the absence of tracked by using a RAG1 ‘ancestory’ mouse (unpub- Flt3 expression (Krueger and von Boehmer, 2007). The lished data), applying the similar strategy used to track circulating T-cell progenitor differentiates into only the history of lysozyme M gene transcription (Ye et al., T and NK cells, but not into B or myeloid cells. Thus, 2003). The mouse line harboring the Cre recombinase T and B cells could be derived from these multiple, knocked into the RAG1 locus was established, and was independent lymphoid progenitor populations. It is crossed with the ROSA26 yellow fluorescence protein critical to test the lineal relationship among these popula- (YFP) reporter mouse line. In this line, the Cre tions and their relative contribution toward T- and recombinase is activated in parallel with the endogenous B-cell homeostasis by performing competitive reconsti- activation of RAG1 transcription, and once RAG1 is tution analysis, for example. activated, cells are permanently labeled with YFP. The majority of CLPs are negative for RAG1-Cre/YFP. Only RAG1-Cre/YFPÀ CLPs give rise to T as well as B and NK cells in vivo, although RAG1-Cre/YFP þ CLPs Developmental pathways in human hematopoiesis give rise only to B and NK cells, indicating that true CLPs do not have a history of RAG1 activation. The expression patterns of surface markers specific for Furthermore, the RAG1-Cre/YFPÀ CLP rapidly ac- hematopoietic stem and progenitor populations are quires ETP phenotype under thymic microenvironment considerably different between human and mouse. after intrathymic injection (unpublished data). These Importantly, mouse LT-HSCs are CD34À/lo (Osawa et al., data suggest that the CLP is not a population down- 1996), CD38 þ (Randall et al., 1996) and CD90 (Thy1)lo stream of the ELP, but represents the T-cell pathway (Spangrude et al., 1988; Morrison and Weissman, 1994), independent of the ELP. Since the GMLP can generate whereas human HSCs are hCD34 þ (Okuno et al., 2002), at least the CLP (Arinobu et al., 2007), the divergence of hCD38À (Terstappen et al., 1991; Ishikawa et al., 2003)

Bone Marrow Thymus

LSK cells CD44+CD25- GMLP Myeloid HSC RAG1+IL-7R- ELP ETP T cell

CLP c-Kit c-Kit

pTα+ CLP2 proB B cell

Sca-1 Sca-1 Figure 3 T-cell development pathway from multiple murine lymphoid progenitors. RAG1 þ IL-7RÀ ELPs and RAG1ÀIL-7R þ CLPs independently exist in early hematopoiesis, and both of them have strong lymphoid reconstitution potential. PU.1 þ GMLPs might be the precursor for ELPs and CLPs. pTa þ CLP-2 might be derived from the CLP. These populations home the thymus to generate T cells via the earliest thymic precursor (ETP), or produce B cells in the bone marrow. The relative contribution of these progenitors toward lymphoid development is still unclear. CLP, common lymphoid progenitor; ELP, earliest lymphocyte progenitor; GMLP, granulocyte–monocyte–lymphoid progenitor; HSC, hematopoietic stem cell.

Oncogene Hematopoietic developmental pathways H Iwasaki and K Akashi 6692 and hCD90 þ (Baum et al., 1992). The fact that mouse the CLP phenotype and the requirement of IL-7 signaling LT-HSCs in steady-state bone marrow do not express may change during human ontogeny. significant levels of mCD34 (Osawa et al., 1996) raised In the myelo-erythroid pathway, CMP, GMP and an important question of whether hCD34 þ cells could MEP subsets are isolatable within the hCD34 þ hCD38 þ mark all long-term self-renewing human HSCs. Mouse fraction in both the bone marrow and (Manz lines harboring human genomic P1 artificial chromo- et al., 2002). All are negative for the early lymphoid some (PAC) clones containing the entire hCD34 gene, markers hCD10, hCD7 or hIL-7Ra. These myelo- including all exons, introns and more than 18 kb of 50- erythroid progenitors are prospectively isolatable ac- and 30-flanking sequences were established (Okuno cording to the expression of hCD45RA and hIL-3Ra. et al., 2002). In all transgenic mouse strains, mCD34À/lo hCD45RAÀhIL-3Ralo (CMPs), hCD45RA þ hIL-3Ralo LSK LT-HSCs expressed the hCD34 transgene. In (GMPs) and hCD45RAÀhIL-3RaÀ (MEPs) efficiently progenitor populations, the hCD34 transgene was formed distinct myelo-erythroid colony types according expressed in the majority of CLPs, CMPs and GMPs, to their definitions. CMPs give rise to MEPs and GMPs and B30% of MEPs (Okuno et al., 2002). These data in vitro, and a significant proportion of CMPs possess strongly support the notion that hCD34 þ human bone clonal GM and MegE potentials (Manz et al., 2002). marrow cells contain the vast majority of LT-HSCs, as Thus, the hierarchical myeloid progenitor relationships well as primitive myeloid and lymphoid progenitors. demonstrated in mice is well preserved in human The long-term reconstitution potential of human hematopoiesis. Phenotypic comparisons between mouse HSCs has been directly shown in SCID mouse and human subsets show that CD34, a marker positive repopulating assays. In the most advanced xenotrans- for only murine CMPs and GMPs, is uniformly plant models by utilizing the Rag2 and IL-2Rg double- expressed on all three human subsets, and that the deficient mouse (Traggiai et al., 2004) or the FcgRII/III (CD16/CD32), marking murine CMPs and NOD–SCID IL2Rg deficient mouse (Ishikawa et al., 2005), GMPs, was not detectable in none of human myeloid hCD34 þ hCD38À or the hCD34 þ hCD90 þ fractions of progenitors (Manz et al., 2002). human bone marrow and cord blood can reconstitute all All of these hematolymphoid progenitors develop human hematolymphoid lineage cells for a long term, from hCD34 þ hCD38À HSC population (Ishikawa et al., indicating that these hCD34 þ fractions contains normal 2005), but themselves have no self-renewal activity in human HSCs. However, it is still questionable that these xenogeneic transplantation models (Ishikawa et al., xenotransplant models correctly recapitulate the settings 2007), indicating that they are downstream of of the congenic murine transplant or the clinical human human LT-HSCs in both the bone marrow and the HSC transplant. For example, although mouse CD34À cord blood. HSCs reconstitute congenic animals for a long term even at the single cell level, injection of at least 1000 cells of the human hCD34 þ hCD38À bone marrow cells is The human counterpart for murine LMPP or GMLP have required to obtain multi-lineage, long-term reconstitu- not been identified in human hematopoiesis tion in xenotransplant models. Thus, it is unknown Another critical difference between human and mouse whether the HSC that is capable of reconstitution of all hematopoiesis is the expression pattern of Flt3. Around lineages at the single cell level exists in human hemato- 40–80% of hCD34 þ bone marrow and cord blood cells poiesis, and if so, what percent of hCD34 þ hCD38À express hFlt3 (Rappold et al., 1997; Sitnicka et al., or hCD34 þ hCD90 þ cells are multipotent, long- 2003). Although a fraction of both the hFlt3 þ and term HSCs. the hFlt3À populations gave rise to multilineage colo- nies containing all myelo-erythroid components, the hFlt3 þ hCD34 þ and hFlt3ÀhCD34 þ populations pre- Human lymphoid and myeloid progenitor subsets within dominantly formed GM and erythroid colonies, respec- the hCD34 þ hCD38 þ MPP fraction tively (Rappold et al., 1997; Gotze et al., 1998; Sitnicka Galy et al. (1995) first reported the existence of lymphoid- et al., 2003), and cells with NOD–SCID reconstitution committed progenitors in the human bone marrow with activity reside in the hCD34 þ hFlt3 þ fraction (Sitnicka the hCD34 þ hCD38 þ hCD45RA þ hCD10 þ phenotype. et al., 2003). Our recent experiments have revealed that Lymphoid potential was determined by reconstitution hFlt3 is expressed in the entire human bone marrow and of human bone and thymus fragments implanted into cord blood hCD34 þ hCD38À population, and purified SCID mice as well as by in vitro culture systems. Limiting hFlt3 þ hCD34 þ hCD38À cells could reconstitute dilution assays suggested that this population contains multilineage cells for a long term in the NOD-SCID/ clonal progenitors of B, NK and DC lineages (Galy et al., IL-2Rgnull xenogeneic transplantation system (unpub- 1995). In human cord blood, CLP activity is found in lished data). Therefore, unlike mouse hematopoiesis, the hCD7 þ fraction of hCD34 þ hCD38ÀhCD45RA þ human LT-HSCs express hFlt3. Furthermore, in strik- population (Hao et al., 2001). Interestingly, the hCD7 þ ing contrast to mouse hematopoiesis where Flt3 is hCD34 þ hCD38ÀhCD45RA þ population does not exist expressed in CLPs but not GMPs (D’Amico and Wu, in the adult bone marrow (Hao et al., 2001). IL-7Ra,a 2003; Karsunky et al., 2003), hFlt3 is expressed in critical marker for the murine CLP, is expressed in GMPs as well as in CLPs at a high level in human human hCD10 þ CLPs in the bone marrow, but not in hematopoiesis (unpublished data). Thus, the hFlt3 hCD7 þ CLPs in the cord blood. These data suggest that expression does not mark lymphoid-primed progenitor

Oncogene Hematopoietic developmental pathways H Iwasaki and K Akashi 6693 populations in human. The significant difference of Flt3 of fusion genes at the CMP or GMP stage is distribution in human and mouse hematopoiesis sug- sufficient for progenitors to acquire self-renewal activity gests that the critical role of Flt3 signaling in hemato- to develop into the LSC. More directly, retroviral poietic development could also be different. hFlt3 is transduction of leukemia fusions including MLL-ENL expressed in leukemic blasts in most cases with acute (Cozzio et al., 2003), MLL-AF9 (Krivtsov et al., 2006) myelogenous leukemia (AML) (Carow et al., 1996; and MOZ-TIF2 (Huntly et al., 2004) could confer LSC Rosnet et al., 1996), and FLT3 is one of the most property in self-renewing HSCs as well as non-self- frequently mutated genes in AML (Gilliland, 2002; renewing CMPs–GMPs. Thus, at least these leukemic Stirewalt and Radich, 2003). The FLT3 mutants fusion proteins can induce leukemic transformation at transduce constitutively active Flt3 signaling, which the level of myeloid progenitors without involving could be the cause of poor prognosis in AML with HSCs. However, in a practical scenario of sponta- FLT3 mutations (Kiyoi et al., 1999; Schnittger et al., neously arising human leukemia, it appears that HSCs 2002; Thiede et al., 2002). Because the signal from FLT3 accumulate the leukemogenic mutations, and that these mutations should be controlled under the regulation of genetic mutations exert their effects at the progenitor normal Flt3 expression machinery, signaling from FLT3 stage, leading to the generation of LSCs downstream of mutations should involve HSCs and GMPs, both of the HSCs. which are critical targets for leukemic transformation in The human LSC population capable of repopulating mouse AML models (see the next section) (Cozzio et al., AML in the NOD–SCID mouse model has the hCD34 þ 2003; So et al., 2003; Huntly et al., 2004; Wang et al., hCD38À normal HSC phenotype (Bonnet and Dick, 2005). Thus, special considerations are required in 1997). Only hCD34 þ hCD38À but not hCD34 þ CD38 þ utilizing mouse models to understand the role of FLT3 fractions of leukemia cells in each AML subtype mutations in human leukemogenesis. according to French–American–British (FAB) classifi- In summary, although the hierarchical precursor– cation except for M3 could reconstitute AML when they progeny relationships observed in mouse hematopoiesis are injected intravenously into sublethally irradiated appear to be generally preserved in human, expression NOD–SCID mice. The question is whether the patterns of a number of surface antigens at each hCD34 þ hCD38À phenotype of LSCs reflects the human developmental stage are significantly different between HSC origin of . Recent studies have shown human and mouse, and hematopoietic progenitors that the surface phenotype of hCD34 þ hCD38À LSCs is restricted to GM–lymphoid lineage have not been significantly different from that of normal human HSCs identified in human hematopoiesis. (Jordan, 2002). For example, in the majority of AML cases, LSCs do not express hCD90 (Miyamoto et al., 2000), but express a high level of IL-3Ra (Jordan et al., 2000) and hCD45RA (unpublished data), resembling Cellular targets in leukemogenesis the phenotype of normal human GMPs. Therefore, it is also possible that human AML LSCs may arise from Leukemias appear to depend on a small population of GMPs but lose the expression of hCD38 when they ‘leukemic stem cell (LSC)’ for their continued growth become LSCs. Similarly, at least in a fraction of acute and propagation (Passegue et al., 2003; Huntly and lymphoblastic leukemia (ALL), LSCs possess the Gilliland, 2005). The origin of LSCs is controversial, hCD34 þ hCD38À phenotype but express hCD19 (Castor and their phenotype and biology is still not fully et al., 2005). elucidated. The LSC is likely to be a crucial cellular As in mouse studies, it should be critical to elucidate target in the treatment of leukemias, and therefore, it is the cellular target for LSC formation by the targeted critical to understand difference of cellular properties transduction of leukemia-specific fusion genes into between LSCs and normal HSCs. purified human HSCs or progenitors, and by testing A variety of mouse AML models have been shown whether such treatment recapitulate the transformation that AML can develop from either HSCs or progenitor process of human leukemia. A recent study has shown populations lacking self-renewal capability. The MRP8 that the retroviral transduction of MLL-ENL or MLL- gene encodes a small calcium-binding protein of the AF9 into human bone marrow cells can establish AML S100 family, which is only expressed in CMPs and or ALL in NOD–SCID mouse recipients (Barabe et al., GMPs, but not in HSCs (Jaiswal et al., 2003). By 2007). Although the efficiency of gene transduction into enforcing the expression of leukemia-specific fusion human hematopoietic cells appears to be still unsatis- genes under the control of the human MRP8 promoter, factory, this type of experimental approach should be a number of AML models have been established: useful to clarify more detailed developmental mechan- MRP8-BCR-ABL transgenic mice develop a CML-like isms of human leukemias in future studies. disease with occasional progression into AML (Jaiswal et al., 2003), MRP8-PML-RARa transgenic mice exhibit a preleukemic state, which eventually progresses Conclusion to APL (Brown et al., 1997; Kogan et al., 2001), and MRP8-AML1-ETO transgenic mice develop AML with The ability to prospectively isolate lineage-restricted high frequency after mutagenesis treatment (Yuan et al., progenitors has greatly helped us understand the 2001). These studies suggest that the enforced expression hematopoietic developmental pathway from HSCs in

Oncogene Hematopoietic developmental pathways H Iwasaki and K Akashi 6694 both mouse and human hematopoiesis. It is now clear different from those in mice. Further phenotypic and that murine HSCs choose the myelo-lymphoid or the functional characterization of human hematopoietic myelo-erythroid pathway to develop hematolymphoid cells should be performed by utilizing improved efficient progeny, and therefore the GM lineage cells can develop xenotransplant models. These approaches will even- either of these pathways. The surface phenotype of tually help to identify LSCs that should be a critical human hematopoietic stem and progenitor cells is quite cellular target in leukemia treatment.

References

Adolfsson J, Borge OJ, Bryder D, Theilgaard-Monch K, Astrand- from self-renewing stem cells and short-lived myeloid progenitors. Grundstrom I, Sitnicka E et al. (2001). Upregulation of Flt3 Genes Dev 17: 3029–3035. expression within the bone marrow Lin(À)Sca1(+)c-(+) stem D’Amico A, Wu L. (2003). The early progenitors of mouse dendritic cell compartment is accompanied by loss of self-renewal capacity. cells and plasmacytoid predendritic cells are within the bone 15: 659–669. marrow hemopoietic precursors expressing Flt3. J Exp Med 198: Adolfsson J, Mansson R, Buza-Vidas N, Hultquist A, Liuba K, Jensen 293–303. CT et al. (2005). Identification of flt3(+) lympho-myeloid stem cells DeKoter RP, Lee H-J, Singh H. (2002). PU.1 regulates expression of lacking erythro-megakaryocytic potential a revised road map for the interleukin-7 receptor in lymphoid progenitors. Immunity 16: adult blood lineage commitment. Cell 121: 295–306. 297–309. Akashi K, Kondo M, von Freeden-Jeffry U, Murray R, Weissman IL. DeKoter RP, Walsh JC, Singh H. (1998). PU.1 regulates both (1997). Bcl-2 rescues T in interleukin-7 receptor- cytokine-dependent proliferation and differentiation of granulo- deficient mice. Cell 89: 1033–1041. cyte/ progenitors. EMBO J 17: 4456–4468. Akashi K, Traver D, Miyamoto T, Weissman IL. (2000). A clonogenic Forsberg EC, Serwold T, Kogan S, Weissman IL, Passegue E. (2006). common myeloid progenitor that gives rise to all myeloid lineages. New evidence supporting megakaryocyte–erythrocyte potential Nature 404: 193–197. of flk2/flt3+ multipotent hematopoietic progenitors. Cell 126: Akashi K, Traver D, Zon LI. (2005). The complex cartography of stem 415–426. cell commitment. Cell 121: 160–162. Fujiwara Y, Browne CP, Cunniff K, Goff SC, Orkin SH. (1996). Allman D, Sambandam A, Kim S, Miller JP, Pagan A, Well D et al. Arrested development of embryonic red cell precursors in mouse (2003). Thymopoiesis independent of common lymphoid progeni- lacking transcription factor GATA-1. Proc Natl Acad Sci tors. Nat Immunol 4: 168–174. USA 93: 12355–12358. Arinobu Y, Iwasaki H, Gurish MF, Mizuno S, Shigematsu H, Ozawa Galy A, Travis M, Cen D, Chen B. (1995). Human T, B, natural killer, H et al. (2005). Developmental checkpoints of the basophil/mast cell and dendritic cells arise from a common bone marrow progenitor lineages in adult murine hematopoiesis. Proc Natl Acad Sci USA cell subset. Immunity 3: 459–473. 102: 18105–18110. Gilliland DG. (2002). Molecular genetics of human leukemias: new Arinobu Y, Mizuno S-I, Chong Y, Shigematsu H, Iino T, Iwasaki H insights into therapy. Semin Hematol 39: 6–11. et al. (2007). Reciprocal activation of GATA-1 and PU.1 marks Goodell MA, Rosenzweig M, Kim H, Marks DF, DeMaria M, initial specification of hematopoietic stem cells into myelo-erythroid Paradis G et al. (1997). Dye efflux studies suggest that hematopoie- and myelo-lymphoid lineages. Cell Stem Cell (in press). tic stem cells expressing low or undetectable levels of CD34 antigen Back J, Allman D, Chan S, Kastner P. (2005). Visualizing PU.1 exist in multiple species. Nat Med 3: 1337–1345. activity during hematopoiesis. Exp Hematol 33: 395–402. Gotze KS, Ramirez M, Tabor K, Small D, Matthews W, Civin CI. Barabe F, Kennedy JA, Hope KJ, Dick JE. (2007). Modeling the (1998). Flt3high and Flt3low CD34+ progenitor cells isolated from initiation and progression of human acute leukemia in mice. Science human bone marrow are functionally distinct. Blood 91: 1947–1958. 316: 600–604. Gounari F, Aifantis I, Martin C, Fehling HJ, Hoeflinger S, Leder P Baum CM, Weissman IL, Tsukamoto AS, Buckle AM, Peault B. et al. (2002). Tracing lymphopoiesis with the aid of a pTalpha- (1992). Isolation of a candidate human hematopoietic stem-cell controlled reporter gene. Nat Immunol 3: 489–496. population. Proc Natl Acad Sci USA 89: 2804–2808. Graf T. (2002). Differentiation plasticity of hematopoietic cells. Blood Bhandoola A, Sambandam A. (2006). From stem cell to T cell: one 99: 3089–3101. route or many? Nat Rev Immunol 6: 117–126. Hao QL, Zhu J, Price MA, Payne KJ, Barsky LW, Crooks GM. Bonnet D, Dick JE. (1997). Human acute myeloid leukemia is (2001). Identification of a novel, human multilymphoid progenitor organized as a hierarchy that originates from a primitive hemato- in cord blood. Blood 97: 3683–3690. poietic cell. Nat Med 3: 730–737. Heyworth C, Pearson S, May G, Enver T. (2002). Transcription Brown D, Kogan S, Lagasse E, Weissman I, Alcalay M, Pelicci PG factor-mediated lineage switching reveals plasticity in primary et al. (1997). A PMLRARalpha transgene initiates murine committed progenitor cells. EMBO J 21: 3770–3781. acute promyelocytic leukemia. Proc Natl Acad Sci USA 94: Huntly BJ, Gilliland DG. (2005). Leukaemia stem cells and the 2551–2556. evolution of cancer-stem-cell research. Nat Rev Cancer 5: 311–321. Carow CE, Levenstein M, Kaufmann SH, Chen J, Amin S, Rockwell P Huntly BJ, Shigematsu H, Deguchi K, Lee BH, Mizuno S, Duclos N et al. (1996). Expression of the hematopoietic growth factor receptor et al. (2004). MOZ-TIF2, but not BCR-ABL, confers properties of FLT3 (STK-1/Flk2) in human leukemias. Blood 87: 1089–1096. leukemic stem cells to committed murine hematopoietic progenitors. Castor A, Nilsson L, Astrand-Grundstrom I, Buitenhuis M, Ramirez C, Cancer Cell 6: 587–596. Anderson K et al. (2005). Distinct patterns of hematopoietic Igarashi H, Gregory SC, Yokota T, Sakaguchi N, Kincade PW. stem cell involvement in acute lymphoblastic leukemia. Nat Med (2002). Transcription from the RAG1 locus marks the earliest 11: 630–637. lymphocyte progenitors in bone marrow. Immunity 17: 117–130. Christensen JL, Weissman IL. (2001). Flk-2 is a marker in Igarashi H, Kuwata N, Kiyota K, Sumita K, Suda T, Ono S et al. hematopoietic stem cell differentiation: a simple method to isolate (2001). Localization of recombination activating gene 1/green long-term stem cells. Proc Natl Acad Sci USA 98: 14541–14546. fluorescent protein (RAG1/GFP) expression in secondary lymphoid Corcoran AE, Riddell A, Krooshoop D, Venkitaraman AR. (1998). organs after immunization with T-dependent antigens in rag1/GFP Impaired immunoglobulin gene rearrangement in mice lacking the knocking mice. Blood 97: 2680–2687. IL-7 receptor. Nature 391: 904–907. Ikuta K, Weissman IL. (1992). Evidence that hematopoietic stem cells Cozzio A, Passegue E, Ayton PM, Karsunky H, Cleary ML, express mouse c-kit but do not depend on steel factor for their Weissman IL. (2003). Similar MLL-associated leukemias arising generation. Proc Natl Acad Sci USA 89: 1502–1506.

Oncogene Hematopoietic developmental pathways H Iwasaki and K Akashi 6695 Ishikawa F, Livingston AG, Minamiguchi H, Wingard JR, Ogawa M. Lai AY, Kondo M. (2006). Asymmetrical lymphoid and myeloid (2003). Human cord blood long-term engrafting cells are CD34+ lineage commitment in multipotent hematopoietic progenitors. CD38. Leukemia 17: 960–964. J Exp Med 203: 1867–1873. Ishikawa F, Niiro H, Iino T, Yoshida S, Saito N, Onohara S et al. Mackarehtschian K, Hardin JD, Moore KA, Boast S, Goff SP, (2007). The developmental program of human dendritic cells is Lemischka IR. (1995). Targeted disruption of the flk2/flt3 gene leads operated independently of conventional myeloid and lymphoid to deficiencies in primitive hematopoietic progenitors. Immunity 3: pathways. Blood (in press). 147–161. Ishikawa F, Yasukawa M, Lyons B, Yoshida S, Miyamoto T, Manz MG, Miyamoto T, Akashi K, Weissman IL. (2002). Prospective Yoshimoto G et al. (2005). Development of functional human isolation of human clonogenic common myeloid progenitors. Proc blood and immune systems in NOD/SCID/IL2 receptor {gamma} Natl Acad Sci USA 99: 11872–11877. chain(null) mice. Blood 106: 1565–1573. Martin CH, Aifantis I, Scimone ML, von Andrian UH, Reizis B, von Iwama A, Zhang P, Darlington GJ, McKercher SR, Maki R, Tenen Boehmer H et al. (2003). Efficient thymic immigration of B220+ DG. (1998). Use of RDA analysis of knockout mice to identify lymphoid-restricted bone marrow cells with T precursor potential. myeloid genes regulated in vivo by PU.1 and C/EBPalpha. Nucleic Nat Immunol 4: 866–873. Acids Res 26: 3034–3043. McKenna HJ, Stocking KL, Miller RE, Brasel K, De Smedt T, Iwasaki H, Akashi K. (2007). Myeloid lineage commitment from the Maraskovsky E et al. (2000). Mice lacking flt3 ligand have deficient hematopoietic stem cell. Immunity 26: 726–740. hematopoiesis affecting hematopoietic progenitor cells, dendritic Iwasaki H, Mizuno S, Arinobu Y, Ozawa H, Mori Y, Shigematsu H cells, and natural killer cells. Blood 95: 3489–3497. et al. (2006). The order of expression of transcription factors directs Miyamoto T, Iwasaki H, Reizis B, Ye M, Graf T, Weissman IL hierarchical specification of hematopoietic lineages. Genes Dev 20: et al. (2002). Myeloid or lymphoid promiscuity as a critical 3010–3021. step in hematopoietic lineage commitment. Dev Cell 3: Iwasaki H, Mizuno S, Wells RA, Cantor AB, Watanabe S, Akashi K. 137–147. (2003). GATA-1 converts lymphoid and myelomonocytic progeni- Miyamoto T, Weissman IL, Akashi K. (2000). AML1/ETO-expressing tors into the megakaryocyte/erythrocyte lineages. Immunity 19: nonleukemic stem cells in acute myelogenous leukemia with 451–462. 8;21 chromosomal translocation. Proc Natl Acad Sci USA 97: Iwasaki H, Mizuno SI, Mayfield R, Shigematsu H, Arinobu Y, 7521–7526. Seed B et al. (2005a). Identification of eosinophil lineage- Morrison SJ, Wandycz AM, Hemmati HD, Wright DE, Weissman IL. committed progenitors in the murine bone marrow. J Exp Med (1997). Identification of a lineage of multipotent hematopoietic 201: 1891–1897. progenitors. Development 124: 1929–1939. Iwasaki H, Somoza C, Shigematsu H, Duprez EA, Iwasaki-Arai J, Morrison SJ, Weissman IL. (1994). The long-term repopulating subset Mizuno S et al. (2005b). Distinctive and indispensable roles of PU.1 of hematopoietic stem cells is deterministic and isolatable by in maintenance of hematopoietic stem cells and their differentiation. phenotype. Immunity 1: 661–673. Blood 106: 1590–1600. Na Nakorn T, Traver D, Weissman IL, Akashi K. (2002). Jaiswal S, Traver D, Miyamoto T, Akashi K, Lagasse E, Weissman IL. Myeloerythroid-restricted progenitors are sufficient to confer radio- (2003). Expression of BCR/ABL and BCL-2 in myeloid progenitors protection and provide the majority of day 8 CFU-S. J Clin Invest leads to myeloid leukemias. Proc Natl Acad Sci USA 100: 109: 1579–1585. 10002–10007. Nerlov C, Graf T. (1998). PU.1 induces myeloid lineage com- Jordan CT. (2002). Unique molecular and cellular features of acute mitment in multipotent hematopoietic progenitors. Genes Dev 12: myelogenous leukemia stem cells. Leukemia 16: 559–562. 2403–2412. Jordan CT, Upchurch D, Szilvassy SJ, Guzman ML, Howard DS, Nerlov C, Querfurth E, Kulessa H, Graf T. (2000). GATA-1 interacts Pettigrew AL et al. (2000). The interleukin-3 receptor alpha chain is with the myeloid PU.1 transcription factor and represses PU.1- a unique marker for human acute myelogenous leukemia stem cells. dependent transcription. Blood 95: 2543–2551. Leukemia 14: 1777–1784. Okuno Y, Iwasaki H, Huettner CS, Radomska HS, Gonzalez DA, Karsunky H, Merad M, Cozzio A, Weissman IL, Manz MG. (2003). Tenen DG et al. (2002). Differential regulation of the human and Flt3 ligand regulates dendritic cell development from Flt3+ lymphoid murine CD34 genes in hematopoietic stem cells. Proc Natl Acad Sci and myeloid-committed progenitors to Flt3+ dendritic cells in vivo. USA 99: 6246–6251. J Exp Med 198: 305–313. Osawa M, Hanada K, Hamada H, Nakauchi H. (1996). Long-term Kiyoi H, Naoe T, Nakano Y, Yokota S, Minami S, Miyawaki S et al. lymphohematopoietic reconstitution by a single CD34- low/negative (1999). Prognostic implication of FLT3 and N-RAS gene mutations hematopoietic stem cell. Science 273: 242–245. in acute myeloid leukemia. Blood 93: 3074–3080. Passegue E, Jamieson CH, Ailles LE, Weissman IL. (2003). Normal Kogan SC, Brown DE, Shultz DB, Truong BT, Lallemand-Breiten- and leukemic hematopoiesis: are leukemias a stem cell disorder or a bach V, Guillemin MC et al. (2001). BCL-2 cooperates with reacquisition of stem cell characteristics? Proc Natl Acad Sci USA promyelocytic leukemia retinoic acid receptor alpha chimeric 100(Suppl 1): 11842–11849. protein (PMLRARalpha) to block neutrophil differentiation and Perry SS, Welner RS, Kouro T, Kincade PW, Sun XH. (2006). initiate acute leukemia. J Exp Med 193: 531–543. Primitive lymphoid progenitors in bone marrow with T lineage Kondo M, Weissman IL, Akashi K. (1997). Identification of reconstituting potential. J Immunol 177: 2880–2887. clonogenic common lymphoid progenitors in mouse bone marrow. Peschon JJ, Morrissey PJ, Grabstein KH, Ramsdell FJ, Maraskovsky E, Cell 91: 661–672. Gliniak BC et al. (1994). Early lymphocyte expansion is severely Krivtsov AV, Twomey D, Feng Z, Stubbs MC, Wang Y, Faber J et al. impaired in receptor-deficient mice. J Exp Med 180: (2006). Transformation from committed progenitor to leukaemia 1955–1960. stem cell initiated by MLL-AF9. Nature 442: 818–822. Puel A, Ziegler SF, Buckley RH, Leonard WJ. (1998). Defective IL7R Krueger A, von Boehmer H. (2007). Identification of a T lineage- expression in T(À)B(+)NK(+) severe combined immunodefi- committed progenitor in adult blood. Immunity 26: 105–116. ciency. Nat Genet 20: 394–397. Kulessa H, Frampton J, Graf T. (1995). GATA-1 reprograms avian Randall TD, Lund FE, Howard MC, Weissman IL. (1996). Expression myelomonocytic cell lines into eosinophils, thromboblasts, and of murine CD38 defines a population of long-term reconstituting erythroblasts. Genes Dev 9: 1250–1262. hematopoietic stem cells. Blood 87: 4057–4067. Kuwata N, Igarashi H, Ohmura T, Aizawa S, Sakaguchi N. (1999). Rappold I, Ziegler BL, Kohler I, Marchetto S, Rosnet O, Birnbaum D Cutting edge: absence of expression of RAG1 in peritoneal B-1 cells et al. (1997). Functional and phenotypic characterization of cord detected by knocking into RAG1 locus with green fluorescent blood and bone marrow subsets expressing FLT3 (CD135) receptor protein gene. J Immunol 163: 6355–6359. tyrosine kinase. Blood 90: 111–125.

Oncogene Hematopoietic developmental pathways H Iwasaki and K Akashi 6696 Rekhtman N, Radparvar F, Evans T, Skoultchi AI. (1999). and identification of subgroups with poor prognosis. Blood 99: Direct interaction of hematopoietic transcription factors PU.1 and 4326–4335. GATA-1: functional antagonism in erythroid cells. Genes Dev 13: Traggiai E, Chicha L, Mazzucchelli L, Bronz L, Piffaretti JC, 1398–1411. Lanzavecchia A et al. (2004). Development of a human adaptive Rosnet O, Buhring HJ, Marchetto S, Rappold I, Lavagna C, Sainty D in cord -transplanted mice. Science 304: et al. (1996). Human FLT3/FLK2 receptor tyrosine kinase is 104–107. expressed at the surface of normal and malignant hematopoietic Umland O, Mwangi WN, Anderson BM, Walker JC, Petrie HT. cells. Leukemia 10: 238–248. (2007). The blood contains multiple distinct progenitor populations Schnittger S, Schoch C, Dugas M, Kern W, Staib P, Wuchter C et al. with clonogenic B and T lineage potential. J Immunol 178: (2002). Analysis of FLT3 length mutations in 1003 patients with 4147–4152. acute myeloid leukemia: correlation to cytogenetics, FAB subtype, von Freeden-Jeffry U, Vieira P, Lucian LA, McNeil T, Burdach SE, and prognosis in the AMLCG study and usefulness as a marker for Murray R. (1995). Lymphopenia in interleukin (IL)-7 gene-deleted the detection of minimal residual disease. Blood 100: 59–66. mice identifies IL-7 as a nonredundant cytokine. J Exp Med 181: Schwarz BA, Bhandoola A. (2004). Circulating hematopoietic pro- 1519–1526. genitors with T lineage potential. Nat Immunol 5: 953–960. Wagers AJ, Sherwood RI, Christensen JL, Weissman IL. (2002). Little Seshasayee D, Gaines P, Wojchowski DM. (1998). GATA-1 dom- evidence for developmental plasticity of adult hematopoietic stem inantly activates a program of erythroid gene expression in factor- cells. Science 297: 2256–2259. dependent myeloid FDCW2 cells. Mol Cell Biol 18: 3278–3288. Walsh JC, DeKoter RP, Lee HJ, Smith ED, Lancki DW, Gurish MF Sitnicka E, Bryder D, Theilgaard-Monch K, Buza-Vidas N, Adolfsson et al. (2002). Cooperative and antagonistic interplay between PU.1 J, Jacobsen SE. (2002). Key role of flt3 ligand in regulation of and GATA-2 in the specification of myeloid cell fates. Immunity the common lymphoid progenitor but not in maintenance of the 17: 665–676. hematopoietic stem cell pool. Immunity 17: 463–472. Wang J, Iwasaki H, Krivtsov A, Febbo PG, Thorner AR, Ernst P et al. Sitnicka E, Buza-Vidas N, Larsson S, Nygren JM, Liuba K, Jacobsen SE. (2005). Conditional MLL-CBP targets GMP and models therapy- (2003). Human CD34+ hematopoietic stem cells capable of related myeloproliferative disease. EMBO J 24: 368–381. multilineage engrafting NOD/SCID mice express flt3: distinct flt3 Ye M, Iwasaki H, Laiosa CV, Stadtfeld M, Xie H, Heck S et al. (2003). and c-kit expression and response patterns on mouse and candidate Hematopoietic stem cells expressing the myeloid lysozyme gene human hematopoietic stem cells. Blood 102: 881–886. retain long-term, multilineage repopulation potential. Immunity 19: So CW, Karsunky H, Passegue E, Cozzio A, Weissman IL, Cleary 689–699. ML. (2003). MLL-GAS7 transforms multipotent hematopoietic Yuan Y, Zhou L, Miyamoto T, Iwasaki H, Harakawa N, Hether- progenitors and induces mixed lineage leukemias in mice. Cancer ington CJ et al. (2001). AML1-ETO expression is directly involved Cell 3: 161–171. in the development of acute myeloid leukemia in the presence of Spangrude GJ, Heimfeld S, Weissman IL. (1988). Purification additional mutations. Proc Natl Acad Sci USA 98: 10398–10403. and characterization of mouse hematopoietic stem cells. Science Zhang DE, Hohaus S, Voso MT, Chen HM, Smith LT, Hetherington 241: 58–62. CJ et al. (1996). Function of PU.1 (Spi-1), C/EBP, and AML1 in Stirewalt DL, Radich JP. (2003). The role of FLT3 in haematopoietic early : regulation of multiple myeloid CSF receptor malignancies. Nat Rev Cancer 3: 650–665. promoters. Curr Top Microbiol Immunol 211: 137–147. Terstappen LW, Huang S, Safford M, Lansdorp PM, Loken MR. Zhang P, Behre G, Pan J, Iwama A, Wara-Aswapati N, Radomska HS (1991). Sequential generations of hematopoietic colonies derived et al. (1999). Negative cross-talk between hematopoietic regulators: from single nonlineage-committed CD34+CD38- progenitor cells. GATA proteins repress PU.1. Proc Natl Acad Sci USA 96: Blood 77: 1218–1227. 8705–8710. Thiede C, Steudel C, Mohr B, Schaich M, Schakel U, Platzbecker U Zhang P, Zhang X, Iwama A, Yu C, Smith KA, Mueller BU et al. et al. (2002). Analysis of FLT3-activating mutations in 979 patients (2000). PU.1 inhibits GATA-1 function and erythroid differentia- with acute myelogenous leukemia: association with FAB subtypes tion by blocking GATA-1 DNA binding. Blood 96: 2641–2648.

Oncogene