Regulation of Hematopoietic Stem Cell Fate

Regulation of hematopoietic stem cell fate

Diane S Krause*,1

1Department of Laboratory Medicine, Yale University School of Medicine, New Haven, Connecticut, CT 06520-8035, USA

Oncogene (2002) 21, 3262 ± 3269. DOI: 10.1038/sj/onc/ by low level transcription of many genes that are 1205316 characteristic of multiple independent discreet lineages (Billia et al., 2001; Hu et al., 1997). By single cell RT ± Keywords: hematopoiesis; microenvironment; dieren- PCR analysis, both erythroid and myeloid gene tiation; niche; notch expression programs are initiated by the same progenitor cell prior to exclusive commitment toward the myeloid or erythroid lineage (Hu et al., 1997). Introduction Although many of the cells analysed shared a speci®c phenotype, the genes expressed in these cells varied. By de®nition, a hematopoietic stem cell (HSC) is For example, in CD34+lin- primary bone marrow capable both of unlimited self-renewal and dierentia- cells, approximately 50% of the cells expressed mRNA tion into all mature peripheral blood types. Hemato- for both b-globin (consistent with RBC dierentiation) poietic dierentiation can be viewed as a tree with the and myeloperoxidase (MPO, consistent with myeloid HSC represented by the trunk (or `stem') and each of dierentiation). These data suggest that the cells were the committed progenitors and their more committed `primed' to dierentiate down either the erythroid or progeny represented by the branches. The mechanisms myeloid lineage. None of the cells expressed b-globin in that control whether a HSC self-renews or dierentiates the absence of MPO. However, 40% expressed MPO in remain a mystery. However, many of the important the absence of b-globin. These latter ®ndings suggest players in this regulation have been identi®ed. The that some of the cells had already become committed transcriptional regulation of hematopoiesis has been to myeloid dierentiation, or that the expression of b- studied in several dierent organisms and many globin is variable in this population of equipotent transcription factors have been identi®ed that specify population cells. It is not yet known whether the whether hematopoietic stem cells dierentiate down the transcripts detected at the single cell level are myeloid, lymphoid, or erythromegakaryocytic lineages. translated. Current methods do not allow detection One of the themes that has grown more apparent over of low level protein expression for a single cell. years of study is that the critical pathways that control Another elegant study was performed using single cell HSC and their progeny are shared by other stem and gene analysis primarily for genes that encode growth progenitor cell types throughout the body and through- factor receptors (Billia et al., 2001). The cells analysed out phylogeny. The transcriptional regulation of were always derived from a four or eight cell cluster that erythropoiesis, lymphopoiesis and myelopoiesis is had grown from a single cell in culture. The potential already covered in detail in other manuscripts in this dierentiative stage of the cell analysed was predicted to issue of Oncogene. This review focuses on components be the same as that of the other cells in the cluster and of the cellular microenvironment that regulate HSC only those cells that were members of a cluster in which dierentiation, and on the gene expression patterns that each of the other cells had the same fate as one another play a role in maintaining a stem cell in its was used. In this way, the gene expression pattern could undierentiated state. These ®ndings are placed into be correlated with the dierentiation potential of the context with other tissue systems and other organisms. cells. While for many hematopoietic growth factor receptors, the pattern of gene expression was as expected with a restriction of lineage expression as the cells dierentiated, there was also `promiscuous' gene expres- Gene expression patterns sion in the most multipotent cells. For example, c-fms, which encodes the M-CSF receptor and is understood to Postulated mechanisms of lineage restriction play a role in monocytic dierentiation, was expressed As cells dierentiate, there is an orchestrated silencing not only in that lineage but also in multipotent cells and of some genes and activation of others. Experimental in partially committed precursors in the erythroid, data suggest that pluripotent stem/progenitor cells are megakaryocyte, and neutrophil lineages. Similarly, `primed' to dierentiate down several dierent lineages though IL-7 and IL-2 receptor transcripts are known to be expressed in B and T cells, respectively, they were also expressed in multipotent precursors (Billia et al., *Correspondence: DS Krause; E-mail: [email protected] 2001). Regulation of HSC fate DS Krause 3263 These two studies have been instrumental in proving or undetectable levels in the dierentiated progeny (e.g. that more pluripotent hematopoietic cells express genes scl). Type 1 genes would be implicated as playing a role previously thought to be unique to speci®c lineages. in maintaining the stem cell phenotype and preventing Therefore, two concepts coexist. The ®rst is that there dierentiation. A type 2 change occurs when a gene is can be a set of genes that function together to maintain expressed at very low or undetectable levels in stem stem-cell-ness (or pluripotentiality). Second, what cells and is expressed in both daughter cell types (e.g. makes stem cells unique from their more committed CD38). Type 3 genes are expressed at detectable levels progeny may be that they express low levels of many in stem cells and are completely inactivated in one dierent genes previously considered to be lineage daughter cell while being expressed at higher levels in speci®c, and these cells have not yet permanently the other daughter cell (e.g. MPO). Type 4 genes are turned o these genes. A corollary to this model of similar to type 3 except expression of these genes is lineage speci®cation in which unilineage commitment is undetectably in stem cells. Because the data suggest `prefaced' by a promiscuous phase of multilineage gene that promiscuous low-level gene expression is stochas- expression, is that once lineage commitment has been tic a gene that is expressed in some but not all stem achieved, the genes expressed in the lineage pathways cells could potentially evade detection by chance. A not taken are permanently silenced. This is consistent caveat regarding this schematic is that type 1 genes, with the relatively new understanding of hematopoietic which are stem cell speci®c for the three cell stages dierentiation as being due not just to speci®c shown, may be expressed at high levels at later stages transcription factors acting positively to determine in dierentiation (e.g. scl and notch). lineage commitment, but also due to those same lineage speci®c transcription factors actively repressing Stochastic versus directed stem cell regulation alternate lineage programs. This has been shown to be the case for several transcription factors including Hematopoietic stem cells either self-renew, thereby GATA1, myb and Pu.1 (Nerlov et al., 2000; Takahashi maintaining stem cell properties, or alternatively, give et al., 2000; Zhang et al., 1999; 2000). rise to cells increasingly committed to dierentiation Figure 1 demonstrates schematically the dierent into various hematopoietic lineages. It is unclear if this types of gene changes that can occur when a stem cell cell fate decision is controlled by a purely stochastic dierentiates into two dierent daughter cell types that mechanism or is the result of environmental cues represent commitment towards distinct lineages. This mediated at least in part through speci®c receptor same scheme would also be applicable to any bipotent ligand interactions. It is likely that cell fate is or pluripotent progenitor cell that dierentiates down in¯uenced both by the stochastic nature of gene at least two separate lineages. A type 1 change in gene expression and by soluble factors and cell-cell interac- expression pattern occurs when a gene is expressed at a tions. The levels of promiscuous gene expression vary relatively high level in stem cells and expressed at low in a random manner, but the presence of speci®c combinations of growth factor receptors, for example, will prime certain cells for induction when and if they are exposed to speci®c growth factor combinations present in the microenvironment.

Specific genes and their regulation In addition to the relatively low-level expression of multiple `promiscuous' genes, many laboratories have worked to identify genes that appear to be expressed uniquely in hematopoietic stem and progenitor cells and not in mature committed blood cells (Type I above). Multiple experimental techniques have been designed to identify these genes including, but not limited to, analysis of genes that are altered in leukemia and lymphoma by analysis of translocations in human disease (Okuda et al., 1996), insertional mutagenesis of murine retroviruses (Perkins et al., 1991), dierential display (Bond et al., 1998), sub- tractive PCR hybridization (Tsai et al., 2000), and microarrays followed by bioinformatic analysis (Phil- lips et al., 2000). These studies have progressed with the caveat that the stem cell phenotype is not conferred Figure 1 Shown above is a schematic of gene expression pro®les only by expression of speci®c genes that are uniquely that occur during dierentiation of a hematopoietic stem cell into two daughter cells of dierent hematopoietic lineages. A type 1 expressed in these cells, but also by the combination of change is shown in black, type 2 in slanted stripes, type 3 in gray these and other gene products in the cell. Genes that and type 4 in vertical stripes. See text for explanation were discovered via analysis of mutation or altered

Oncogene Regulation of HSC fate DS Krause 3264 expression patterns in leukemia are now known to play the yeast genome has been analysed for the occurrence critical roles in normal hematopoiesis. Such genes of speci®c combinations of transcription factor binding include myb and ets, which are induced by a domains in the promoters of genes that are known to leukemogenic avian retrovirus in which the myb and be expressed at the same time. Using bioinformatics, ets sequences are fused to form a chimeric protein. the investigators distinguished between coordinately Another critical transcription factor is myc, which is regulated genes and those that were expressed induced by a group of viruses that can transform simultaneously, but regulated by alternative mechan- monocytic cells. isms (Pilpel et al., 2001). One of the common themes Additional genes identi®ed by these techniques that has emerged from promoter analyses of many of include scl, rbtn2, GATA2 and AML1, each of which the critical genes for hematopoiesis is that most are has been shown, using murine knock-out experiments, TATA less and have GC rich promoter regions that to be necessary for de®nitive hematopoiesis (Okuda et interact with the ubiquitous transcription factor Sp1. al., 1996; Porcher et al., 1996; Tsai et al., 1994; Many also contain multiple potential binding domains Yamada et al., 2000; Yamada et al., 1998). Scl is a for the GATA and ets transcription factor families. helix-loop-helix domain transcription factor that was Given that there is some commonality amongst identi®ed as part of a translocation breakpoint. In hematopoietic promoters, gene expression speci®city is addition to a role in early hematopoiesis, scl is also achieved by complex protein-protein interactions expressed in mature lymphoid and megakaryocytic amongst multiple transcription factors that are them- cells. Rbtn2 (also known as lmo2 or ttg-2) is a lim selves expressed at dierent relative levels in a cell type domain containing protein that functionally interacts speci®c manner. with scl and GATA-2. GATA2, a member of the GATA family of zinc ®nger transcription factors, is postulated to play a role in self-renewal of HSC and to Stem cell niche prevent their dierentiation (Shivdasani and Orkin, 1996). AML1, a runt domain containing transcription Regulation of HSC fate involves multiple highly factor, synergizes with many dierent transcription orchestrated pathways that determine the cell cycle factors to enhance transcription of multiple genes status, and gene expression pro®le. This complexity is including, but not limited to, GM ± CSF, MPO, and due in large part to the fact that HSC's do not grow as the B and T cell antigen receptors. However, AML1 independent autonomous units. Rather, these cells are can also act in concert with additional proteins to surrounded in all dimensions by the marrow micro- inhibit transcription. Many dierent leukemogenic environment, which is de®ned by cell-cell interactions, translocations disrupt normal AML1 activity (Lutter- cell-ECM interactions, and exposure to variable bach and Hiebert, 2000). concentrations and combinations of soluble factors including cytokines. In addition, the proximity to endothelial cells will aect the local oxygen level. The Coordinate gene expression bone marrow microenvironment is comprised of The multiple approaches taken over many years of stromal cells, a diverse population consisting of study re¯ect the diculty of identifying a cohort of ®broblasts, smooth muscle cells, endothelial cells and coordinately regulated genes that are expressed others. These cells not only provide a scaold to the uniquely in the relatively rare hematopoietic stem cell developing stem and progenitor cells, but also produce subpopulation. As a further complication, many of the transmembrane ligands, extracellular matrix compo- genes that some may consider to be stem cell speci®c nents and soluble proteins. are also expressed in certain mature blood subpopula- tions. Even the classic murine stem cell `markers' sca Cell ± growth factor interactions and kit, both of which are expressed on hematopoietic stem and progenitor cells, are also expressed on T HSC are exposed in situ to many dierent growth lymphocytes and mast cells, respectively. Conversely, factors, some soluble, some bound to ECM and some some growth factors such as thrombopoietin, are both bound to adjacent cells. Due to the inherent diculty key regulators of lineage speci®c dierentiation and in studying these cells in their microenvironment, most appear to play a role in hematopoietic stem cell self- of what we know about how dierent soluble growth renewal. factors control HSC fate has been deduced from in Large-scale gene expression analyses have guided us vitro studies in which the growth factors are not toward the regulatory domains of genes that may be necessarily introduced at physiological concentrations, coordinately regulated as they are activated and/or in physiological ratios to one another, or for the repressed at similar stages of hematopoiesis. However, appropriate amount of time. The primary question that the analyses of the genome that will reveal the has been addressed is whether we can simulate in vitro underlying regulatory processes for these genes have the conditions that tell the HSC to self-renew. If this not yet been fully elucidated. Bioinformatics ap- were completely understood and could be performed in proaches, which are necessary to identify potentially vitro, we could then expand small numbers of highly shared regulatory domains of coordinately expressed pluripotent HSC for clinical transplantation as well as genes, are currently under development. For example, for other scienti®c studies.

Oncogene Regulation of HSC fate DS Krause 3265 Isolation and testing of soluble factors that mediate cells called the terminal ®lament. Amongst the control ex vivo maintenance and expansion of stem cells has signals received by the mitotically active GSC from continued to pose a major challenge (reviewed recently adjacent somatic cells are proteins called hedgehog, YB in McNiece and Briddell, 2001). Growth factors that and PIWI, each of which is required for GSC have been shown to enhance the proliferation of early maintenance (Xie and Spradling, 2000). hematopoietic stem and progenitor cells are ¯t3, SCF, There is evidence that homologues of these genes in erythropoietin, IL6 and thrombopoietin (Henschler et higher vertebrates play analogous roles in maintaining al., 1994; Zandstra et al., 1997). These eects are the pluripotentiality of hematopoietic stem cells in the dependent both on the concentration or each compo- marrow. For example, receptors and coreceptors for nent and on the speci®c combinations of growth hedgehog ligands (Patched1 and Smoothened) are factors that are present. For example, amongst the expressed on normal marrow cells and elements of very promising data on ex vivo expansion, the complex the hedgehog signaling pathway are involved in the of interleukin 6 (IL-6) and soluble IL-6 receptor (IL-6/ control of hematopoietic dierentiation (Detmer et al., sIL-6R), acts in synergy with SCF and Flt3 (Lyman et 2000). Similarly, a human homologue of piwi, called al., 1993; Small et al., 1994) to expand immature hiwi, is expressed by CD34+ hematopoietic stem and human HSCs. In order to prove that early hemato- progenitor cells but not by more dierentiated cell poietic stem/progenitor cells were expanded, the populations. Expression of hiwi in a human leukemia researchers used a xenogeneic transplant model in cell line results in decreased proliferation suggesting which human cells are transplanted into highly that it may play a role in HSC maintenance (Sharma et immunode®cient mice. The ex vivo expanded human al., 2001) cells were capable of repopulating the immunode®cient mice longterm. The proportion of human CD45+ cells Cell-cell interactions in recipient marrow was 10 times higher in animals that received culture expanded cord blood cells than In the bone marrow microenvironment, HSCs are in those that received comparable numbers of fresh cord cell-cell contact with stromal cells, committed hemato- blood CD34+ cells. poietic progenitor cells and other HSCs. Therefore, Some soluble growth factors have been shown to there is plenty of opportunity for cell-cell signaling. confer stem cell survival while inhibiting cell prolifera- Studies in which HSC were expanded in vitro have tion. A classic example of this is TGFb, which is shown that contact with stromal cells enhances HSC known to play a role in maintaining HSC in a proliferation (Brouard et al., 1998; Dorshkind, 1990). quiescent stem cell state (Eaves et al., 1991; Fortunel HSCs receive messages from adjacent cells via gap et al., 2000; Jacobsen et al., 1995; Liu et al., 1997; junctions (connexins are expressed by developing Sitnicka et al., 1996). For gene therapy studies using hematopoietic cells (for recent reviews see Montecino- retroviruses, proliferation of HSCs, and thus their Rodriguez and Dorshkind, 2001; Rosendaal and infection by retroviruses, is enhanced when TGFb Krenacs, 2000) and receptor/ligand interactions be- activity is inhibited (Dao and Nolta, 1999). The story tween adjacent cells. Here, the notch/jagged pathway(s) of TGFb also brings up another very interesting point are focused on. regarding the cell microenvironment and `priming of Notch and jagged are members of a superfamily of gene expression'. TGFb is a prototypical growth factor highly evolutionarily conserved transmembrane recep- that aects hematopoietic cells very dierently depend- tors that in¯uence numerous cell-fate decisions in both ing upon the state of dierentiation of the cells. For invertebrates and vertebrates. Throughout the phylo- myeloid committed progenitor cells, TGFb induces genetic tree, these superfamilies participate in binary dierentiation while for HSCs, it helps to maintain cell fate decisions. Members of both the notch and them in a quiescent undierentiated state. In both cases jagged superfamilies are predominantly transmembrane there is an inhibition of cell cycle activation, and the molecules. When a notch family member binds to its cell-speci®c eects are due to the dierent relative ligand (jagged family member), a signal is transferred expression levels of multiple transcription factors. to cleave the intracellular domain of notch, which then Additional soluble factors that help to maintain translocates to the nucleus where it interacts with HSCs in culture have been identi®ed via studies that cofactors to aect the transcriptional program. were based on discoveries in drosophila melanogaster Notch was originally ideniti®ed as playing a role in and c. elegans. In drosophila, surrounding support cells regulating cell fate decisions during neuronal end in the developing gonads restrict the self-renewal, and epidermal dierentiation in drosophila development. regulate the number of, germline stem cells. Germline Notch acts by a local feedback mechanism to induce stem cells are a self-renewing population of cells that dierentiation of one cell while simultaneously steering serve as source of gametes in diverse organisms. the surrounding cells to an alternative dierentiative Current work suggests that self-renewal is controlled fate. This feedback mechanism, which is not yet fully both by somatic signaling and by intracellular understood, thus acts to amplify small dierences in mechanisms including dierential gene expression, notch expression between adjacent cells thereby leading and the cell cycle machinery. In the drosophila ovary, them down dierentiative pathways. The speci®c fates each GSC division produces a daughter cell GSC that that are determined based on notch receptor signaling remains in contact with a cluster of somatic support are dependent upon the developmental history of the

Oncogene Regulation of HSC fate DS Krause 3266 notch expressing cell. In hematopoiesis, dierential and decreased notch allows for myeloid dierentiation notch expression plays a role in multiple bipolar cell (Bigas et al., 1998). In addition dierential notch fate decisions as shown in Figure 2. In this ®gure, the expression plays a role in multiple bifurcations in the cell that expresses more notch is shown in black and dierentiation tree of hematopoietic dierentiation. the default pathway in the absence of notch (or that For erythromegakaryocytic progenitors, higher levels cell type which dierentiates from the cell that of notch lead cell towards a megakaryocytic pathway expresses relatively lower levels of notch) is shown in and lack of notch leads cells toward the erythroid white. pathway (Lam et al., 2000). In lymphoid dierentia- HSC's have higher levels of notch than more tion, dierential notch expression guides cells toward committed progenitors (Milner et al., 1994). Jagged-1, the T cells versus B cell dierentiation pathway (Pui et presented both on the surface of bone marrow stromal al., 1999), the ab versus gd T cell pathway (Washburn cells in the microenvironment promotes formation of et al., 1997), and also plays a role in determining primitive precursor cell populations (Karanu et al., whether resultant T cells will dierentiate into CD4+ 2000; Milner and Bigas, 1999; Varnum-Finney et al., or CD8+ cells (Valdez and Robey, 1999). 1998; Varnum-Finney et al., 2000; Walker et al., 1999). The mechanism by which notch plays a role in these Therefore, the committed progenitors may be viewed bifurcations is not yet known. However, consistent as the laterally inhibited cells that have been directed to with the ®nding that speci®cation of cell fate seems to an alternate fate. Thus in myelopoiesis, higher notch involve the coordinated activation of genes related to levels maintain the cells in a less dierentiated state one fate and suppression of expression of genes of

Figure 2 Shown above is a schematic of several binary decision points at which notch expression has been shown to play a role. At each dierentiation step, cells can dierentiate down one of two dierent fates. Alternatively, the binary decision may be whether or not the stem/progenitor cell should dierentiate. The fate that is taken when notch expression is lacking or decreased is shown with a white arrow pointing to a white cell. The alternative pathway, shown with a gray arrow and a black cell, is taken by the daughter cell when notch expression is high. Small numbers refer to the references on which these data are based and are as follows: 1: Milner et al., 1994; 2: Lam et al., 2000; 3: Bigas et al., 1998; 4: Pui et al., 1999; 5: Washburn et al., 1997; 6: Valdez and Robey, 1999

Oncogene Regulation of HSC fate DS Krause 3267 alternative fates (Lehming et al., 1994; Nerlov et al., dierent heterodimers have dierent ligand speci®city 2000; Zhang et al., 1999; Zhang et al., 2000), notch has although dierent integrins share some ligands (Cou- been shown to interact with CBP/p300 a multifunc- lombel et al., 1997; Teixido et al., 1992). The best tional protein complex which is necessary for the characterized integrin on hematopoietic stem and activity of speci®c transcription factors (Florence and progenitor cells is alpha4/beta1, a beta1 integrin that McGinnis, 1998; Koyano-Nakagawa et al., 1999). For is also known as VLA-4 (very late activation antigen example, GATA1 and myb stimulate erythropoiesis 4). VLA-4 binds to ®bronectin in the ECM as well as and myelopoiesis, respectively and activation of to VCAM-1 (vascular cellular adhesion molecule 1) on GATA1 or myb inhibits the transcriptional activity adjacent stromal cells. Fibronectin consists of multiple of the other. Each of these tissue speci®c transcription domains, each with speci®c binding sites for other factors requires CBP as a coactivator of transcription matrix macromolecules and for receptors on the and there is evidence that levels of CBP/p300 are surfaces of cells. Fibronectin can thus bind simulta- limiting (Goodman and Smolik, 2000). It is possible neously to collagen, heparin in the ECM and integrins that notch signaling directly aects which of these two on the surface of various cells. transcription factors will prevail in a given cell and that As for other cell receptor/ECM ligand combinations, the directed stimulation of cell fate in one cell and the the binding of VLA-4 on hematopoietic stem and simultaneously lateral inhibition of that fate in progenitor cells to ®bronectin serves multiple functions adjacent cells is in part mediated by the ability of including adhesion of the HSC to the microenviron- notch to sequester CBP/p300 in cells in a dierential ment, homing of circulating hematopoietic stem and manner. progenitor cells, and also transduction of signals to the intracellular environment via the VLA-4 receptor. The critical role of VLA-4/®bronectin (and VLA-4/VCAM) Cell-ECM interactions in homing was shown initially by injecting antibodies Extracellular matrix (ECM) is composed of three against VLA-4 into the circulation of mice and major classes of molecules: structural proteins such as primates (Craddock et al., 1997; Papayannopoulou collagen and elastin, specialized proteins such as and Nakamoto, 1993). In these animals, the hemato- ®bronectin and laminin, and proteoglycans which poietic stem and progenitor cells exited the marrow consist of a protein core to which is attached long into the peripheral blood. As expected, co-injection of chains of repeating disaccharide units termed glycosa- anti-VLA4 antibodies along with marrow cells at the minoglycans (GAGs). ECM proteins in the BM include time of bone marrow transplantation greatly decreased ®bronectin, collagen, laminin and cytokines, which can the ability of hematopoietic stem and progenitor cells bind to ECM proteins. Similar to assays on the to home to the marrow. In addition to hematopoiesis function of various cytokines, ex vivo expansion in the bone marrow, analogous eects of integrins on protocols have been used to assess the role of ECM stem cell maintenance have been shown in other tissues in HSC regulation. Although this section is focused on including epidermal stem cells in the skin (Jensen et al., the eect of ECM on HSC's, it is important to note 1999; Zhu et al., 1999). The epidermal stem cells that the ECM acts in concert with cell-cell interactions express high levels of b1-integrins and binding of these and soluble factors to regulate the HSC. integrins to ECM proteins inhibits dierentiation. In general, adhesion of hematopoietic stem and progenitor cells to the marrow ECM inhibits cellular proliferation and prevents apoptosis, both of which Stage speci®c cell cycle regulation in early hematopoiesis would lead to longterm survival of quiescent hematopoietic stem cells. Clues are beginning to emerge The focus of this review has been on transcriptional regarding how these eects are mediated. For example, regulation and on aspects of the stem cell niche that binding of integrins on hematopoietic stem or play a role in controlling stem cell fate. However, the progenitor cells has been shown to lead to increased stem cells themselves have intrinsic characteristics that p27 expression, and p27 halts progression of the cell dier from other cells including speci®c combinations cycle by inhibiting a cyclin dependent kinase, as of nuclear factors, chromosomal modi®cations discussed in further detail below. (Cheung et al., 2000; Kouzarides, 2000; Redner et al., Although there are multiple molecules on the HSC 1999; Tsukiyama and Wu, 1997) and mitotic clocks. membrane that can mediate signals from the ECM These intrinsic clocks, which may in part determine the including integrins, immunoglobulin-like molecules, frequency and number of cell divisions, include cadherins, selectins, and mucins, I focus here on the telomerase (reviewed in Broccoli et al., 1995; Hay¯ick, integrin molecules (Becker et al., 1999; Coulombel et 1998; Notaro et al., 1997) and cyclin dependent kinase al., 1997; Levesque and Simmons, 1999; Teixido et al., inhibitors, which have been shown to regulate the cell 1992). Integrins consist of heterodimers of two cycle dierently in stem cells than in more committed transmembrane glycoprotein subunits that are called progenitor cells. alpha and beta. There are many dierent alpha and Cyclin dependent kinases (CDK) regulate the cell beta subunits and more are yet to be discovered. cycle at dierent checkpoints. Just as there are several Several dierent integrins have been identi®ed on cyclin dependent kinases, there are several CDK hematopoietic stem and progenitor cells, and the inhibitors including p21, p27, and p18. These dierent

Oncogene Regulation of HSC fate DS Krause 3268 CDK inhibitors play distinct roles in the regulation of explanation for these data is that there are more hematopoietic self-renewal and dierentiation. For progenitor cells per stem cell derived from the marrow example, p21 acts to maintain a stem cell in a quiescent of a p277/7 mouse. In contrast, mice lacking p21 stem cell state, while p27 and p18 more speci®cally have a larger than normal stem cell pool that, unlike inhibit cycling of committed progenitor cells (Cheng et the stem cell pool in wildtype animals, can undergo al., 2000a, Cheng et al., 2000b). Inhibition of p27 or exhaustion when stressed (Cheng et al., 2000b). This p18 function leads to a decrease in the cell cycle suggests that p21 normally acts to limit the cycling control of committed hematopoietic progenitors and status of hematopoietic stem cells. does not aect hematopoietic stem cells. In vivo competitive repopulation studies show that bone marrow cells from p277/7 mice achieved a higher Acknowledgments % engraftment than identically prepared cells from This work was funded by NIH grants DK53037, HL 63357, WT control animals (Cheng et al., 2000a). One possible and HK 69207.

References

Becker PS, Nilsson SK, Li Z, Berrios VM, Dooner MS, Jensen UB, Lowell S and Watt FM. (1999). Development, Cooper CL, Hsieh CC and Quesenberry PJ. (1999). Exp. 126, 2409 ± 2418. Hematol., 27, 533 ± 541. Karanu FN, Murdoch B, Gallacher L, Wu DM, Koremoto Bigas A, Martin DIK and Milner LA. (1998). Mol. Cell. M, Sakano S and Bhatia M. (2000). J. Exp. Med., 192, Biol., 18, 2324 ± 2333. 1365 ± 1372. Billia F, Barbara M, McEwen J, Trevisan M and Iscove NN. Kouzarides T. (2000). EMBO J., 19, 1176 ± 1179. (2001). Blood, 97, 2257 ± 2268. Koyano-Nakagawa N, Wettstein D and Kintner C. (1999). Bond HM, Bonelli P, Mesuraca M, Agosti V, Masone C, Mol. Cell Neurosci., 14, 327 ± 339. CuomoC,NisticoA,TassoneP,TuccilloF,CeccoL, Lam LT, Ronchini C, Norton J, Capobianco AJ and Iacopino L, Barbieri V, Cerra M, Costanzo FS, Morrone Bresnick EH. (2000). J. Biol. Chem., 275, 19676 ± 19684. G and Venuta S. (1998). Stem Cells, 16, 136 ± 143. Lehming N, Thanos D, Brickman JM, Ma J, Maniatis T and Broccoli D, Young JW and De Lange T. (1995). Proc. Natl. Ptashne M. (1994). Nature, 371, 175 ± 179. Acad. Sci. USA, 92, 9082 ± 9086. Levesque JP and Simmons PJ. (1999). Exp. Hematol., 27, Brouard N, Chapel A, Neildez-Nguyen TMA, Granotier C, 579 ± 586. Khazaal I, Pearlt B and Thierry D. (1998). Leukemia, 12, Liu X, Sun Y, Constantinescu SN, Karam E, Weinberg RA 1128 ± 1135. and Lodish HF. (1997). Proc. Natl. Acad. Sci. USA, 94, Cheng T, Rodrigues N, Dombkowski D, Stier S and Scadden 10669 ± 10674. DT. (2000a). Nat. Med., 6, 1235 ± 1240. Lutterbach B and Hiebert SW. (2000). Gene, 245, 223 ± 235. Cheng T, Rodrigues N, Shen H, Yang Y, Dombkowski D, Lyman S, James L, Vanden Bos T, de Vries P, Brasel K, Sykes M and Scadden DT. (2000b). Science, 287, 1804 ± Glinial B, Hollingsworth L, Picha K, McKenna H, Splett 1808. R, Fletcher F, Marakovsky E, Farrah T, Foxworthe D, Cheung WL, Briggs SD and Allis CD. (2000). Curr. Opin. Williams D and Beckmann M. (1993). Cell, 75, 1157. Cell Biol., 12, 326 ± 333. McNiece I and Briddell R. (2001). Exp. Hematol., 29, 3±11. Coulombel L, Auray I, Gaugler M and Rosenblatt M. Milner LA and Bigas A. (1999). Blood, 93, 2431 ± 2448. (1997). Acta Haematologica, 97, 13 ± 21. Milner LA, Kopan R, Martin DI and Bernstein ID. (1994). CraddockCF,NakamotoB,AndrewsRG,PriestleyGVand Blood, 83, 2057 ± 2062. Papayannopoulou T. (1997). Blood, 90, 4779 ± 4788. Montecino-Rodriguez E and Dorshkind K. (2001). J. Dao MA and Nolta JA. (1999). Leukemia, 13, 1473 ± 1480. Leukoc. Biol., 70, 341 ± 347. Detmer K, Walker AN, Jenkins TM, Steele TA and Dannawi Nerlov C, Querfurth E, Kulessa H and Graf T. (2000). Blood, H. (2000). Blood Cells Mol. Dis., 26, 360 ± 372. 95, 2543 ± 2551. Dorshkind K. (1990). Annu.Rev.Immunol.,8, 111 ± 137. Notaro R, Cimmino A, Tabarini D, Rotoli B and Luzzato L. Eaves CJ, Cashman JD, Kay RJ, Dougherty GJ, Otsuka T, (1997). Proc. Natl. Acad. Sci. USA, 94, 13782 ± 13785. Gaboury LA, Hogge DE, Lansdorp PM, Eaves AC and Okuda T, van Deursen J, Hiebert SW, Grosveld G and Humphries RK. (1991). Blood, 78, 110 ± 117. Downing JR. (1996). Cell, 84, 321 ± 330. Florence B and McGinnis W. (1998). Genetics, 150, 1497 ± Papayannopoulou T and Nakamoto B. (1993). Proc. Natl. 1511. Acad.Sci.USA,90, 9374 ± 9378. Fortunel N, Hatzfeld J, Kisselev S, Monier MN, Ducos K, Perkins A, Mercer J, Jenkins N and Copeland N. (1991). Cardoso A, Batard P and Hatzfeld A. (2000). Stem Cells, Development, 111, 479. 18, 102 ± 111. Phillips RL, Ernst RE, Brunk B, Ivanova N, Mahan MA, Goodman RH and Smolik S. (2000). Genes Dev., 14, 1553 ± Deanehan JK, Moore KA, Overton GC and Lemischka 1577. IR. (2000). Science, 288, 1635 ± 1640. Hay¯ick L. (1998). Exp. Gerontol., 33, 639 ± 653. Pilpel Y, Sudarsanam P and Church GM. (2001). Nat. Henschler R, Brugger W, Luft T, Frey R, Mertelsmann R Genet., 29, 153 ± 159. and Kanz L. (1994). Blood, 84, 2898 ± 2903. PorcherC,SwatW,RockwellK,FujiwaraY,AltFWand Hu M, Krause D, Greaves M, Sharkis S, Dexter M, Orkin SH. (1996). Cell, 86, 47 ± 57. Heyworth C and Enver T. (1997). Genes Dev., 11, 774 ± Pui JC, Allman D, Xu L, DeRocco S, Karnell FG, Bakkour 785. S, Lee JY, Kadesch T, Hardy RR, Aster JC and Pear WS. Jacobsen FW, Stokke T and Jacobsen SEW. (1995). Blood, (1999). Immunity, 11, 299 ± 308. 86, 2957 ± 2966. Redner RL, Wang J and Liu JM. (1999). Blood, 94, 417 ± 428.

Oncogene Regulation of HSC fate DS Krause 3269 Rosendaal M and Krenacs TT. (2000). Pathol. Oncol. Res., 6, Varnum-Finney B, Xu L, Brashem-Stein C, Nourigat C, 243 ± 249. Flowers D, Bakkour S, Pear WS and Bernstein ID. (2000). Sharma AK, Nelson MC, Brandt JE, Wessman M, Mahmud Nat. Med., 6, 1278 ± 1281. N, Weller KP and Homan R. (2001). Blood, 97, 426 ± 434. Walker L, Lynch M, Silverman S, Fraser J, Boulter J, Shivdasani RA and Orkin SH. (1996). Blood, 87, 4025 ± 4039. Weinmaster G and Gasson JC. (1999). Stem Cells, 17, Sitnicka E, Ruscetti FW, Priestly GV, Wolf NS and 162 ± 171. Bartelmez SH. (1996). Blood, 88, 82 ± 88. Washburn T, Schweighoer E, Gridley T, Chang D, Fowlkes Small D, Levenstein M, Kim E, Carow C, Amin S, Rockwell BJ, Cado D and Robey E. (1997). Cell, 88, 833 ± 843. P, Witte L, Burrow C, Ratajczak MZ, Gewirtz AM and Xie T and Spradling AC. (2000). Science, 290, 328 ± 330. Civin CI. (1994). Proc. Natl. Acad. Sci. USA, 91, 459 ± Yamada Y, Pannell R, Forster A and Rabbitts TH. (2000). 463. Proc Natl Acad Sci USA, 97, 320 ± 324. TakahashiT,SuwabeN,DaiP,YamamotoM,IshiiSand Yamada Y, Warren AJ, Dobson C, Forster A, Pannell R and Nakano T. (2000). Oncogene, 19, 134 ± 140. Rabbitts TH. (1998). Proc Natl Acad Sci USA, 95, 3890 ± Teixido J, Hemler ME, Greenberger JS and Anklesaria P. 3895. (1992). J. Clin. Invest., 90, 358 ± 367. Zandstra PW, Conneally E, Petzer AL, Piret JM and Eaves Tsai FY, Keller G, Kuo FC, Weiss M, Chen J, Rosenblatt M, CJ. (1997). PNAS, 94, 4698 ± 4703. Alt FW and Orkin SH. (1994). Nature, 371, 221 ± 226. Zhang P, Behre G, Pan J, Iwama A, Wara-Aswapati N, Tsai S, Fero J and Bartelmez S. (2000). Blood, 96, 950 ± 957. Radomska HS, Auron PE, Tenen DG and Sun Z. (1999). Tsukiyama T and Wu C. (1997). Curr.Opin.Genet.Dev.,7, Proc Natl Acad Sci USA, 96, 8705 ± 8710. 182 ± 191. ZhangP,ZhangX,IwamaA,YuC,SmithKA,MuellerBU, Valdez P and Robey E. (1999). Cold Spring Harb. Symp. Narravula S, Torbett BE, Orkin SH and Tenen DG. Quant. Biol., 64, 27 ± 31. (2000). Blood, 96, 2641 ± 2648. Varnum-Finney B, Purton LE, Yu M, Brashem-Stein C, Zhu AJ, Haase I and Watt FM. (1999). Proc Natl Acad Sci Flowers D, Staats S, Moore KA, LeRoux I, Mann R, Gray USA, 96, 6728 ± 6733. G, Artavanis-Tsakonas S and Bernstein ID. (1998). Blood, 91, 4084 ± 4091.

Oncogene