Hemangioblastic Origins of the Adult Endothelial Progenitor Cell

Material Information

Hemangioblastic Origins of the Adult Endothelial Progenitor Cell
Copyright Date:


Subjects / Keywords:
Angiogenesis ( jstor )
Blood ( jstor )
Bone marrow ( jstor )
Cells ( jstor )
Endothelial cells ( jstor )
Hemangioblasts ( jstor )
Hematopoietic stem cells ( jstor )
Progenitor cells ( jstor )
Stem cells ( jstor )
Tumors ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright the author. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:


This item is only available as the following downloads:

Full Text




Copyright 2005 by Christopher Lawrence Bray


To my precious


iv ACKNOWLEDGMENTS I would like to first thank my mentor, Ed ward Scott, for providing invaluable encouragement and guidance during my gradua te training. I would also like to thank Robert Fisher, Bill Slayton, and my co mmittee members (Bryon Peterson, Naohiro Terada, Terence Flotte, and Henry Baker) for their precious time, immeasurable patience, and engaging dialog. I would like to thank a nd congratulate the many members of our lab for their commitment to scientific research th at often inspired me. I would like to also thank my parents for giving me the capacity to choose such an in teresting and dynamic career. Lastly, I would like to thank my wi fe, Carmen, for reminding me of life’s most important elements and for shar ing with me her unqualified love.


v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... ..x CHAPTER 1 THE HEMANGIOBLAST AND ITS PROGENY.......................................................1 Introductory Comments................................................................................................1 Physical Evidence.........................................................................................................2 Mesodermal Origins of Embryonic Hemangioblasts............................................2 Yolk-sac Blood Island Formation.........................................................................4 Embryonic Vascular Development........................................................................5 Embryonic Hematopoietic Development............................................................12 Yolk-sac.......................................................................................................12 Aorta-gonad-mesonepheros.........................................................................13 Placenta........................................................................................................15 Liver, spleen, and bone marrow...................................................................16 Genetic Hemangioblast Evidence...............................................................................17 Vascular Endothelial Growth Factor...................................................................18 Tie1 and Tie2.......................................................................................................21 In-Vitro Hemangioblast Evidence..............................................................................22 Adult Hemangioblast..................................................................................................24 Adult Hematopoietic Development.....................................................................24 Adult Vascular Development..............................................................................26 Adult Hemangioblast Activity.............................................................................30 Adult Hemangioblast Medical Therapy..............................................................33 Tumor vascularization..................................................................................33 Nontumor vascularization............................................................................37 2 GENETIC MARKING TO TRACK CLONALITY..................................................41 Non-Retroviral Marking Studies................................................................................41 Early Retroviral Marking Studies...............................................................................42 Retroviral Clonality of Huma n Hematopoiesis in Mice.............................................43


vi Retroviral Clonality of Non-Human Primate Hematopoiesis.....................................44 Thesis Statement.........................................................................................................46 3 MATERIALS, METHODS, AND STATISTICS......................................................48 Creating the Retrovi ral Tag Library...........................................................................48 Constructing the Recombinant Retroviral Tagging Plasmid...............................49 Retroviral Packaging Cells..................................................................................52 Cell Culture.........................................................................................................54 Packaging of Recombinant Retrovirus................................................................54 Determination of Retroviral Titer........................................................................55 Introduction to FACS..........................................................................................56 Demonstrating Random Labeling of Cells with the MIGRN Library........................57 Retroviral Tagging of NIH3T3 Cells..................................................................57 Isolation and Sequencing of NIH3T3 Retroviral Tags........................................58 Bone Marrow Harvesting....................................................................................58 Bone Marrow Culture and Retroviral Tagging...................................................59 Isolation and Sequencing of Retroviral Tags from Cultured Bone Marrow Cells.................................................................................................................60 Generating Chimeric Re trovirally Tagged Mice........................................................63 Secondary Transplantation of Retrovirally Tagged HSCs..................................63 Screening of the Transplant Recipients...............................................................64 Clonal Analysis of Stem Cell Progeny of Mice Transplanted with Tagged Marrow.65 Statistical Analysis......................................................................................................66 4 USING RANDOM RETROVIRAL LI BRARY AS CELL LABELS.......................69 Packaging High-Titer Retrovirus................................................................................69 Constructing the Recombinant Retroviral Tagging Plasmid......................................72 Specificity and Sensitivity of the Retroviral Tag Technique......................................73 Complexity of MIGRN Labeling................................................................................75 5 USING TAGGED BONE MARROW TO IDENTIFY THE CLONAL ORIGIN OF ENDOHTHELIAL PROGENITOR CELLS........................................................78 Cofirmation That Stem Cells Were Transduced.........................................................78 Primary Transplants....................................................................................................82 Secondary Transplants................................................................................................85 6 DISCUSSION.............................................................................................................88 Enrichments for the Endothelial Progenito r Cell and the Hematopoietic Stem Cell.88 Thesis Questions.........................................................................................................89 Studying Clonality In-Vivo........................................................................................90 Identification of Hemangioblast Activity...................................................................92 General Conclusion....................................................................................................94 Future Studies.............................................................................................................94


vii LIST OF REFERENCES...................................................................................................99 BIOGRAPHICAL SKETCH...........................................................................................123


viii LIST OF FIGURES Figure page 1-1 Locations of putative hemangioblast activity.............................................................3 1-2 Angiogenesis and vasculogenesis..............................................................................7 1-3 Phenotypes of bone marrow stem cells and progenitors..........................................28 3-1 Map of MIGRN........................................................................................................52 4-1 Phoenix-A and GP293 packaged retrovirus.............................................................70 4-2 Phoenix-E packaged retrovirus................................................................................71 4-3 Ecotropic BOSC23 packaged retrovirus..................................................................72 4-4 Retroviral transfection and harvest optimization.....................................................72 4-5 Random insert used to construct retr oviral library...................................................73 4-6 Detection of the retroviral tag..................................................................................74 4-7 Sensitivity of retr oviral tag detection.......................................................................75 4-8 Clonal tags of NIH3T3 cel ls demonstrate complexity.............................................76 4-9 Enriched bone marrow tags demonstrate complexity..............................................77 5-1 Phenotype of cultured bone marrow cells................................................................79 5-2 Long-term engraftment of tagged hematopoietic stem cells....................................80 5-3 Multilineage engraftment of tagged hematopoietic stem cells.................................81 5-4 Tumor in MIGRN transplanted mouse.....................................................................81 5-5 Primary transplant endot helial progenitor cell sort..................................................83 5-6 Primary transplant hemat opoietic stem cell enrichment..........................................84 5-7 Secondary transplant endo thelial progenitor cell sort..............................................86


ix 5-8 Photographs of endothelial progenitor enriched cells..............................................86 5-9 Clonal analysis of transplanted mice........................................................................87 6-1 Overlap in oligoclonal pools de monstrating hemangioblast activity.......................92


x Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy HEMANGIOBLASTIC ORIGINS OF THE ADULT ENDOTHELIAL PROGENITOR CELL By Christopher Lawrence Bray August 2005 Chair: Edward Scott Major Department: Molecular Genetics and Microbiology The existence of an adult endothelial progenitor cell (EPC) has augmented our understanding of the process of neovascularizat ion in both health and disease. Similar to embryonic vascular development, adult vascular repair involves bot h angiogenesis and an EPC-driven vasculogenesis. In earlier work, our lab showed that a single highly purified hematopoietic stem cell (HSC) has the capacity to form not only hematopoietic cells, but also retinal endothelial cells in the adult mouse. In the current study, transplants were performed from both whole bone marrow and enri ched stem cells that had been uniquely labeled with a novel retroviral tagging system. This novel tagg ing system used a library of random retroviral tags that has distin ct advantages over other clonal tracking techniques. This technique was used to s how that in transplantation of whole bone marrow and enriched stem cells, the same oli goclonal pool of stem cells that generates mature hematopoietic cells, also generates EP Cs. This finding suggest s that gene therapy


xi of the transplantable HSCs may someday be used to treat a myriad of EPC-related vascular diseases.


1 CHAPTER 1 THE HEMANGIOBLAST AND ITS PROGENY Introductory Comments Efficient and dynamic transportation and communication are vital for any kind of growth. Attesting to their importance in vert ebrate embryogenesis, th e circulatory system is the first functional organ system to deve lop. Survival of the embryo requires a rapid interface with the maternal blood supply. Th e circulatory system (in development and adulthood) facilitates transpor t of gases, nutrients, waste, and hormones throughout the body. Additionally, it is a conduit for the trafficking of many cells, including hematopoietic progenitor cells. Without the e fficient vascular dist ribution system, for example, hematopoiesis would quickly ove rwhelm the bone marrow of adults without providing the body with needed cells. The vascular syst em, like the hematopoietic system, can, with remarkable accuracy, dyna mically respond to the ever-changing needs of the body, thus maintaining homeostasis. Research on development, differentiati on, and regulation in these two systems paved the way for a wide range of current and future medical therapies. At the intersection of these two re search paths is a somewhat elusive stem cell (the hemangioblast) which has both hematopoietic and endothelial developmental potential. The hemangioblast, a term first coined by Murray (Murray 1932), was proposed by Sabin in her seminal work observing the close rela tionship between vascular endothelial cells and appearance of hematopoietic cells in the developing chicken (Sabin 1920). Reportedly, she directly observed the buddi ng of an erythroblastic cell from the


2 angioblasts of embryonic st ructures called blood isla nds. Since then, the close relationship between hematopoietic and e ndothelial cells has unf olded on many levels. Chapter 1 discusses our curre nt understanding of the clos e relationship between blood and blood vessels in both physiologic and pa thologic states; and extrapolates (from present trends) future direct ions and medical therapies. Physical Evidence Mesodermal Origins of Embryonic Hemangioblasts The embryonic development of blood a nd blood vessels from mesodermal stem cells is consistent with the existence of a common hemangioblast. Along with endoderm and ectoderm, the mesoderm is one of three pr imary germ layers in early development. Mesoderm forms a variety of tissues incl uding the bones, kidneys, and muscle; in addition heart, blood vessels, and blood cells. Mesodermally derived cells with hematopoietic and endothelial characteristics ar e near each other in two or more separate stages of early development. Amphibian embr yos have at least two distinct 32-cell stage blastomeres (individual cells of the preimpla ntation embryo) that give rise to these separate mesodermal origins (Ciau-Uitz et al. 2000). However, in chicks, mice, and humans, it is unclear whether these mesoderm al regions are the result of independent mesodermal origination or due to seeding by a unique mesodermal stem cell. Some studies focused on the extraembryonic (embryo-de rived tissues that fo rm structures such as the placenta and the umbilical cord) me soderm, which forms blood islands in an embryonic structure called the yolk sac (YS, Figure 1-1). Blood islands are composed of cells with both angioblast (embryonic va scular progenitor) and erythroid progenitor activity (Sabin 1920; Moore and Metcalf 1970; Haar a nd Ackerman 1971; Weissman et al. 1978; Yoder et al. 1997). Other work ha s focused on the intr aembryonic paraaortic


3 splanchnopleural mesoderm which contri butes to the deve loping aorta-gonadmesonephros (AGM, Figure 1-1) region of the embryo and has hemangioblasts (also called “hemogenic” endothelial cells because of their endothelial phenotype) (Muller et al. 1994; de Bruijn et al. 2000b; Tavian et al. 2001; Peault and Tavian 2003). In addition to the YS and AGM, recent evidence suggests that sites with hemangioblastic potential may be more widespread than previously t hought, including the avia n allantois (Caprioli et al. 1998), mouse vitelline and umbilical arteries (de Br uijn et al. 2000a), and mouse placenta (Gekas et al. 2005; Ottersbach and Dzierzak 2005). Figure 1-1. Locations of putative hemangi oblast activity. The extraembryonic yolk-sac blood islands, the ventral wall of th e developing dorsal aorta, and the postnatal bone marrow are the three main areas in which patterns of hematopoietic and endothelial cell de velopment suggest a hemangioblast.


4 Yolk-sac Blood Island Formation In mammals, the inner, embryoblast portion of the blastocyst gives rise to the embryo-proper. The outer, trophoblast portion fo rms most of the extra-embryonic tissues. The embryoblast forms a bilaminar embryonic di sk composed of two layers: the epiblast and hypoblast. Hypoblast cells migrate to line the inner surface of the cytotrophoblast, and form the primitive yolk sac, an e xo-coelomic cavity. Between the hypoblast endoderm and the cytotrophoblast, extraembryoni c mesoderm—derived from the epiblast (mice) or hypoblast (primates)—invades and creates an extra-embryonic coelom. Eventually, extraembryonic somatic mesoderm lines the trophoblast si de of the chorionic cavity and extraembryonic visceral mesoderm lin es the yolk sac side of this chorionic cavity. The tight association of endoderm and mesoderm germ layers is termed splanchnopleure. The mesoderm likely recei ves inductive signals, such as Indian hedgehog, smoothened (Smo), bone morphogeni c protein-4 (BMP-4), and vascular endothelial growth factor A (VEGFA) from the adjacent visceral endoderm (Haar and Ackerman 1971; Kimelman a nd Griffin 2000; Byrd and Grab el 2004; Vokes et al. 2004). In response to these signals, yolk-sac visceral mesoderm differentiates into the earliest angioblasts (mouse embryonic post-fertilizati on, Day E7.5 to E8; human, Day E17-21). The first hematopoietic cells also emerge from the splac hnopleure of the yolk sac in specific structures called blood islands (mouse, Day E7.5 to 8; human, Day E18 to 21). These blood islands form between two angioblas t-derived vascular plexuses. Shortly after forming (mouse, Day E8-9), the cells around the perimeter of the blood islands flatten into spindle-shaped endothelial cells (H aar and Ackerman 1971). The central cells become primarily primitive erythroblasts. This close spatial and temporal association


5 between hematopoietic progenito rs and a portion of the yolksac angioblasts led to the early hypothesis of a common hemangioblast. Embryonic Vascular Development The association between hematopoietic a nd vascular cells extends beyond the early steps of YS blood island commitment. In fact , these two distinct tissues show a strong interdependence throughout many stages of pr enatal development, suggesting underlying genetic programs that facil itate direct communication and interaction. Such coordination would certainly be facilitated by overlapping developmental programs, as would be the case if they originated from a common hemangi oblast. Because much of the research on hematopoietic and vascular development wa s performed assuming that these tissues developed independently, much of the phys ical hemangioblast evidence is merely circumstantial. Since this assumption is proba bly not entirely true, recent efforts are underway to unify these developmental models. The vascular system of the developing embryo forms in intraembryonic and extraembryonic tissues by mechanisms of vasculogenesis and angiogenesis (Vokes et al. 2004). These processes ultimately create vascular structures involving endothelial cells, mural cells, and extracellular matrix (Figure 1-2). Embryonic vasculogenesis generates the early vascular network from angi oblast progenitor cells (CD31+, CD34+, VEGFR2+). Intrinsic angioblasts (p rogenitors that emerge in s itu) contribute to a vascular network by vasculogenesis, a process of de novo vessel formation (Drake 2003). Intraembryonic vasculogenesis is initiated (mouse, Day E7.5; human, Gestational Week 3 to 4) along with the emergence of endocar dial progenitor cells (Drake and Fleming 2000). Vasculogenesis is limited to early em bryogenesis and creates a honeycomb-like primary capillary plexus in developing intern al organs, such as the spleen and pancreas.


6 Vasculogenesis also creates the framework fo r the primitive aortic and vitelline arteries (Pardanaud et al. 1989). The mesodermally de rived angioblasts prol iferate and coalesce to form cords and continuous chains of e ndothelial cells. Vascul ogenesis continues as spaces between angioblasts form, enlarge, and join with other spaces to create uniform tubular structures. Only later is this early vascular netw ork of tubes stabilized and remodeled by angiogenesis into an inte rconnecting functional vascular system. A comparison of intraembryonic and extrae mbryonic angioblasts suggests similar, but nonidentical cellular mechanisms. In bot h cases, the first blood vessels form from mesodermally derived angioblasts in clos e association with endoderm (Vokes et al. 2004). Endodermal induction via hedgehog or fibr oblast growth factor (FGF) signaling is likely to be involved in angioblast vascular endothelial tube formation (Poole et al. 2001; Vokes and Krieg 2002; Byrd and Grabel 2004) . In addition to a ngioblast vascular development, endodermal induction may be invo lved in preserving the hemogenic fate of certain angioblasts. Pardanaud and colleagues showed that the endode rmal growth factors FGF-2, vascular endothelial gr owth factor (VEGF)A, and transforming growth factor (TGF) maintain the hemangi opoietic potential of mesoderm (Pardanaud and DieterlenLievre 1999). Furthermore, ec todermal growth factor TGF , and epidermal growth factor (EGF) abrogate the hemogenic potential and produce strictly angi opoietic angioblasts. Although the similarities suggest this, the differences show the incompleteness of our model for mesoderm commitment a nd vascular development. Extraembryonic angioblasts contribute to the endothelium of the avian allantois, mouse placental labyrinth, vitelline, and yolk sac vessels. In th e yolk sac, layers of angioblasts are often directly associated w ith hematopoietic cells in the blood island structures. In contrast, intraembryonic angioblasts are rarely (with certain excep tions) found in layers or


7 associated with hematopoiesis, but are usua lly found as isolated cells in mesodermal tissues (Vokes et al. 2004). Figure 1-2. Angiogenesis and vasculogenesis. In the developing embryo, the vascular structure is initially created by angioblasts through a process of vasculogenesis and is followed by angi ogenesis, a process of vascular remodeling. In the adult, vascular ization was thought to be due to angiogenesis. However, a type of postn atal vasculogenesis has recently been described. The lack of an abundant association between intraembryonic angioblasts and hematopoietic progenitors during most of em bryonic vasculogensis has been used to


8 argue against the existence of the hemangioblast in late r stages of development. However, since hematopoiesis would clea rly be ineffective without a vascular distribution system, the timing of commitment and expansion of these sibling progenitors is likely weighted toward earlier vascular development. Therefore the relative number and migratory status of vascular progenitors and hematopoietic progenitors is a delicate balancing act that dynamically adjusts in the developing embryo. Although differences may exist in the physical arrangement of these progenitors at various points in development, a common hemangioblast stem cell cannot be excluded. The term angiogenesis is broadly defined as the growth and extension of vessels from a preexisting vascular network. Angiogene sis, in contrast to vasculogenesis, is known to occur throughout life. More recently, the definition of embryonic angiogenesis has expanded to also include the processes of growth, remodeling, and maturation that transform the primitive vascular plexus into a complex vascular network. The typical definition does not specify the cellular origin s of the contributing endothelial cells, but simply that it be an elaboration of an ex isting structure. Therefore, extrinsic embryonic angioblasts (or adult endothelial progenitor cells) contributing to an existing vascular network is also considered angiogenesis (Pardanaud and Dieterlen-Lievre 2000). Both sprouting and nonsprouting angiogenesis have been observed. Sprouting angiogenesis, perhaps mediated by hypoxia, occurs during organogenesis and is prevalent in organs vascularized primarily by extr insic endothelial progenitor activ ity, such as the limb buds, brain, and retina (Pardanaud et al. 1989). Sprouting angiogenesi s is also believed to be responsible for most new postnatal vasculoge nesis because it seems to be the default response to cytokine stimulated endothelia l proliferation. Nonspr outing angiogenesis, including bridging and intussuscep tion (transluminal pillar formation) (Burri et al. 2004),


9 are significant in embryonic vascular deve lopment as well as postnatal lung and endometrial vascular development. The nonspr outing types may be more of a response to large metabolic changes creating hemodyna mic disturbances rather than the cytokine-mediated sprouting type. Embr yonic vasculogenesis followed by angiogenic remodeling creates the vascular structures of the initially avascular embryonic organs (Risau 1997). Three vascular circuits are formed in th e embryo at about the same time that the heart begins to beat (huma n, Day E21; mouse, Day E7.5) (Gourdie et al. 2003). The embryonic circuit consists of pa ired dorsal aortae that emerge from the endocardial heart tube and bifurcate into embryonic capillary be ds. The vascular return path is completed with anterior and posterior cardinal veins that merge and drain back into the endocardial heart tube. The remaining two circuits are both extraembryonic. The vitelline (omphalomesenteric, yolk sac) circuit dive rts blood from the dorsal aortae into the vitelline arteries to s upply the yolk sac and is returned back to the heart through vitelline veins. The umbilical (allantoic, placental) circ uit diverts blood from the dorsal aortae into the umbilical arteries to be oxygenated in the placenta and returned via umbilical veins. Within the developing aorta, certain endot helial cells function in a manner that defies a strict endothelial or hematopoietic delineation. Jaffredo and colleagues showed that the aorta is lined with CD45(a panhematopoietic surface antigen) VEGFR2+ (an endothelial surface antigen) flat tened endothelial cells (Jaffr edo et al. 1998). A wave of CD45+ monocytes is initially observed and followed by a wave of CD45+ hematopoietic progenitors, distinguishable because of the intense CD45 staining and scattered locations of monocytes. Acetylated low-density li poprotein conjugated to a fluorescent dye (Ac-LDL) was used to label ao rtic endothelial cells before the emergence of CD45+ cells


10 showing that these endothelia cells generate clusters of CD45+ VEGFR2hematopoietic cells localized to the floor of the aorta. In later work, using retroviral LacZ (a gene that produces the -galatosidase enzyme, which converts a specific carbohydrate substrate into a blue dye) labeling, they also showed that some of these non-monocytic endothelial-derived hematopoie tic cells migrate to para-aor tic foci (which underlies the developing aort a) from which they also tran siently contribute to he matopoiesis (Jaffredo et al. 2000). Splanchnopleuric (those e xposed to endoderm) and not somitic mesoderm-derived angioblasts have this hematopoietic capacity (Pardanaud and Dieterlen-Lievre 2000; Dieterlen-Lievre et al . 2002). Peault and co lleagues (Labastie et al. 1998; Peault and Tavian 2003; Zambidis et al . 2005) identified a similar process in the human embryo, where CD34+ CD45+ hemogeni c cells develop from aortic CD34+ CD45VEGFR2+ endothelial cells. As the vascular system matures, mura l cells (consisting of both pericytes and smooth muscle cells) take resi dence at the interface between the endothelial cells and the surrounding tissue. The specific mesodermal orig ins of mural cells is uncertain, although recent evidence suggests that a VEGFR2 (F lk1)+ hemangioblast-forming cell may also generate at least one type of mural cell, the vascular smooth muscle cell (Ema et al. 2003). Mural cells may use different environmen tal cues to either promote or inhibit endothelial proliferation and migration (D arland and D'Amore 1999). Mural cells also stimulate the production of ex tracellular matrix. Arteries and veins are surrounded by one or more layers of vascular smooth muscle cells (vSMC). Capillaries are surrounded by varying degrees of solitary pericytes. In termediate-sized vessels have distinctive combinations of both types of supportive mural cells (Gerhardt and Betsholtz 2003; Betsholtz et al. 2005). The Notch and Eph/Ephr in signaling pathways participate in the


11 specification of the various vascular struct ures (Wang et al. 19 98; Krebs et al. 2000). Because pericytes share struct ural properties with vSMC, pericytes may be a type of vSMC progenitor (D'Amore 1992) or perhap s a more multipotent stem cell (Sims 2000). A major role of vSMC in the arterioles is to regulate blood pressure . Interestingly, in the metabolically active retina, where blood flow must be carefully regulated, pericyte density is quite high (Sims 2000). Histamin e and angiotensin II responsive pericytes cause endothelial rel ease of endothelin-1 and downregul ation of iNOS which ultimately causes vasoconstriction (Sims 2000; Gerhardt and Betsholtz 2003). Pericytes are known to also be phagocytic and may contribute to antigen presentation in certain inflammatory situations. Because of their location and dens ity in neural and retinal vascular beds, pericytes may also participate in the bloodbrain barrier. Perivascular macrophages are also phagocytic cells located in approximate ly the same position as pericytes (Guillemin and Brew 2004). Under certain conditions lik e inflammation, it is unclear whether perivascular macrophages and pericytes act in a complementary or supplementary role. Pericytes play an integral role in both prenatal and postnatal angiogenesis, although the complete mechanisms also remain uncl ear. Production of platel et-derived growth factor (PDGF)-B by endothelial cells has emerge d as a key player in pericyte recruitment and hyperplasia (Gerhardt and Betsholtz 2003). Ge ne ablation of PDGF-B or its receptor produces mice with frequent capillary micr oaneurysms, a phenotype similar to the pericyte-reduced diabetic retina (Lindahl et al. 1997). In addition to providing inhibitory control, the pericyte may be involved in promoting early stages of certain types of angiogenesis. In the vascular beds of normal retina and tumors, approximately 1% of the capillaries are pericyte tubes without e ndothelial contribution (O zerdem and Stallcup 2003). In angiogenesis of the corpus luteum , pericytes, induced by endothelial-derived


12 nitric oxide (NO), invade and become th e major source of VEGFA, which initiates angiogenesis (Amselgruber et al. 1999). The individual and combined effects of VEGF, PDGF-B, and other major factors (includi ng TGF, Ang-1, Ang-2, and ephrin-B2) are actively being scrutinized for their role in physiologic and pathologic angiogenesis (Zhang et al. 2001; Hammes et al. 2004; Winkler et al. 2004). Embryonic Hematopoietic Development Yolk-sac Concurrent with the emergence of embr yonic angioblasts is the emergence of extraembryonic yolk-sac (YS) blood islands that contain primitive hematopoietic cells (mouse, Day E7.5; human, Day E16) (Palis and Yoder 2001; Peault and Tavian 2003). Current evidence suggests that these primitive he matopoietic cells originate from a subset of YS angioblasts (Li et al. 2005). These blood islands produce primitive erythrocytes that contain embryonic globin and are characteri stically large, nucleated, and mitotically active (Barker 1968; Bethlenfalvay and Block 1970; Brotherton et al. 1979; Russel 1979). The YS-derived primitive erythrocytes circulate in the developing vascular network to carry maternally derived oxygen to embryonic tissues. By Day E16, these primitive nucleated erythrocytes no l onger circulate in the mouse em bryo (Bethlenfalvay and Block 1970). Enucleated megalocytic erythroblasts fi rst appear in the mouse at Day E13. These megalocytes have a morphology consistent with primitive erythrocytes, although enucleated, and persist until Day E18. In hum ans, yolk-sac hematopoiesis is present from Gestational Week (gw) 3 to 6 (Charbord et al. 1996). Besides primitive erythrocytes, there is some evidence that the YS generates small numbers of definitive erythrocyte, macrophage, microglial, and myeloid progenito r cells (Moore and Metcalf 1970; Alliot et al. 1999; Palis et al. 1999). Transplants of ear ly (mouse, Day E8-9) YS stem cells fail to


13 have long-term (LT) adult repopulating activity. Studies usin g fetal conditioning techniques, however, suggest that these YS stem cells may be fetal repopulating (Fleischman et al. 1982; Tole s et al. 1989). Transplants wi th later (mouse, Day E10.5 to 11) YS stem cells do have LT adult repopul ating activity (Muller et al. 1994). It is unclear whether these repopulati ng cells are derived from the YS or are the result of migration through the newly ge nerated vascular system. Aorta-gonad-mesonepheros Although evidence suggests that embryonic and neonatal repopulating HSCs arise from the mouse yolk sac between Days E8 and 10 (Weissman et al. 1978; Yoder and Hiatt 1997; Yoder et al. 1997; Palis and Yoder 2001), the pers istence and migration of these cells to or from intraembryonic site s remains uncertain (Godin and Cumano 2002). Regardless, two other sites, the aortagonad-mesonephros (AGM) and the placenta, gained recent interest. The first fully comp etent long-term adult engrafting HSCs arise independent of the YS in the para-aortic sp lanchnopleura of the developing AGM region (mouse, Day E10.5; human, Day E27) (Garcia-Po rrero et al. 1995; Dz ierzak et al. 1998; de Bruijn et al. 2002; Peault and Tavian 2003) . Hematopoietic cells appear to emerge from mesodermally-derived “hemogenic” endoth elial cells located in the ventral wall of the developing dorsal aorta and the adjacent vi telline and umbilical arteries. In humans, this second wave peaks, consists of thousands of cells, begins at Day E35, and disappears by Day E40. In mice, hemogenic endothelial ce lls are phenotypically defined as Ly-6A/E (Sca-1+), CD31+, CD34+, c-Kit+, VE-cadheri n+ (de Bruijn et al. 2002). In humans, CD34+ / CD31+, CD45aortic wall endothe lial cells generate CD31+, CD34+, CD45+, and lineage marker negative hematopoietic cells that also express Tal1/SCL, c-myb, GATA2, GATA3, c-Kit, and VEGFR2 (Flk1/KDR ) (Labastie et al. 1998; Peault and


14 Tavian 2003). In culture, the hemogeni c endothelial cells have lympho-myeloid hematopoietic differentiation potential (Ober lin et al. 2002), thus demonstrating clonal hemangioblast activity (Ham aguchi et al. 1999). An alternative interpretation of HS C emergence from the AGM region was proposed by Bertrand and colleagues (Ber trand et al. 2005). They found mouse Day E10.5 CD45+ cells to be excl usively monocyte progenitors. This is consistent with Jaffredo and colleagues’ observa tion of an initial wave of CD45+ monocyte progenitors (Jaffredo et al. 2000). The CD45-/low cKit+ fraction was found to contain lymphomyeloid progenitors and AA4.1+ CD41+ HSCs. The suggestion is that these cells are the same as the CD45Flk1+ cells obser ved by Jaffredo. It is unclear, however, why the CD45-/low c-Kit+ cells seen by Be rtrand cannot overlap with the CD45+ hematopoietic progenitors seen by Jaffredo. These prehepatic HSCs also expressed CD31, GATA3, and GATA2, but not Flk1 and were located in both intraaortic clusters (HIACs) and subaortic patches (SAPs). Because of the lack of Flk1 on these CD45-/low cells, the existence of true endothelium that is also hemogenic was questioned. Based on their data, Bertrand and colleagues argued that neither a lack of CD 45 nor the uptake of Ac-LDL is sufficient to identify endothelial ce lls (EC) and thus it is impossible to assume that cells with these charact eristics are hemogenic endothe lium. They propose that the HSCs and aortic floor ECs are non overlappi ng populations and that HSCs are generated in the SAP, migrate to the aorta, contribut e to the HIAC, release through the endothelium into circulation, and colonize the fetal liver. If true, this view would not exclude the existence of a common hemangioblast, but woul d merely move the fate divergence back one step by removing hemogenic endothelium from the equation.


15 The theory that some of the endothelial cells found in the AGM are hemogenic and some are not is similar to the observation that in the YS, some angi oblasts participate in the blood islands and some are strictly endothe lial. These observations are consistent, in fact, with the proposed idea that two dis tinct subsets of mesoderm (somatic and splanchnopleuric) generate e ndothelial progenitors with different potential in the chick/quail system (Pardanaud and DieterlenLievre 1999; DieterlenLievre et al. 2002). Only the splanchnopleuric mesoderm ge nerates both blood and endothelial cells (Pardanaud et al. 1996). In chic ks, the cells of the ventral wa ll of the dorsal aorta arise from the endoderm-associated splanchnopleura l (ventral) mesoderm. Endoderm induction involves VEGF, bFGF, and TGF, whereas ectodermal inhibition involves EGF and TGF (Pardanaud and Dieterlen-Lievre 1999) . Other non-hemogenic regions of the aortic wall arise from paraxial (lateral) mesoderm. Transient dye and permanent retroviral fate mapping have been used to confir m clonal hemangioblast activity in aortic endothelial cells (chick Flk1+ DiI-Ac-LDL+; mouse VE-cadherin+ CD45Flk1+) derived from splanchnopleural mesoderm (N ishikawa et al. 1998a; Nishikawa et al. 1998b; Jaffredo et al. 2000; Dieter len-Lievre et al. 2002). Limita tions in the specificity of the labeling and/or detection procedures coul d be used to support the alternative view suggested by Bertrand and colleagues. Howeve r, the act of Flk1+ cells generating CD45+ cells was not captured in the Bertrand study s uggesting that the picture is not completely clear. Transplant of a single (or uniquely), permanently labe led cell(s) would be the most convincing in-vivo experiment to show hemangioblast activity. Placenta Gekas et al. and Ottersbach et al. identif ied the placenta as another HSC niche in the mouse conceptus (Gekas et al. 2005; Otte rsbach and Dzierzak 2005). Similar to the


16 AGM, the mouse placenta first harbors pluri potent, adult BM rec onstituting HSCs at E10.5-E11. Also like mouse AGM HSCs, the surface phenotype of these cells is Ly6A+, CD34+, c-Kit (high) and their observed distri bution is consistent with the theory of hemogenic endothelium. Unlike the AGM, howev er, the pool of placental HSCs (which peaks between Days E12 and E13.5) is quantita tively more robust. Gekas and colleagues speculate that the origin of these placental HSCs is likely to be allantoic mesoderm, but further experimentation is required. Liver, spleen, and bone marrow As the AGM and placenta provide for a transient HSC niche, hematopoiesis (the maturation of hematopoietic progenitors) shif ts from the YS to the fetal liver (FL) (Muller et al. 1994). In the mouse, multi lineage hematopoiesis (erythropoiesis, myelopoiesis, lymphopoiesis) takes place in th e liver from approximately E11 until the end of gestation (Keller et al. 1999). In the hum an conceptus, the liver is the major site of hematopoiesis from 6 to 22 gw (Charbord et al. 1996). In addition to being a niche for hematopoietic maturation, the FL is the major destination of migratory HSCs, which accumulate in levels that exceed all other feta l tissues (Christensen et al. 2004). The only clear association between hemat opoietic and endothelial cells in the liver is the fact that the (mouse, Day E11) sinusoids (which are lined with endothelial cells) are associated with YS-derived scavenger macrophages (S asaki and Iwatsuki 1997). Around Day E14, these macrophages migrate to the hepatic cords and become surrounded by a ring of erythroblasts in an erythroblastic blood island. The bone marrow is another major site of hematopoiesis in the fetus. The mouse bone marrow begins to be colonized at E15 (Jordan et al. 1995; Sanchez et al. 1996), although the liver retains most of the HSC activity at birth (Harrison and Astle 1997).


17 Human bone marrow development progresses thr ough a series of five stages before the final organization of the long bones (Charbor d et al. 1996). These include Stage II, the formation of the bone marrow cavity by CD68+ osteoclasts (8.5 to 9.0 gw); and Stage III, the formation of sinuses and “primary logettes ”, vascular structures composed of inner, flattened CD34+ endothelial intimal cells surrounded by SM actin+ smooth muscle medial cells and connective tissue (9.0 to 10.5 gw). After these prep aratory stages, round, CD45+, glycophorin A+ erythroblasts are first observed in the human fetal bone marrow, Stage IV (10.5 to 15 gw); and the final orga nization of long bones, Stage V (over 16 gw), begins. Stage IV is the first stage in which rare (in contrast to YS, AGM, FL) population of round, hematopoietic, CD34+ cells can be observed. Bone marrow erythropoiesis, as in the liver, takes place in migratory eryt hroblastic blood island s (Bernard 1991). Around the time of transitioning from FL to BM in the mouse and human (which remains controversial), the spleen also contributes to hematopoietic maturation and as an HSC niche (Wolber et al. 2002; Lim et al. 2005). Ho wever, this organ is dispensable for fetal mouse survival and normal hematopoietic devel opment and does not seem to be a critical “stepping stone” to the bone marrow. These major embryonic hemangioblast and fe tal HSC migrations are known to be mediated by the chemokines VEGFA and SDF-1 (CLC12) binding to their cognate receptors VEGFR2 (Hiratsuka et al. 2005) and CXCR4 (Ara et al. 2003), and the cytokine stem cell factor (SF, steel fact or) binding to its receptor c-Kit (CD117) (Christensen et al. 2004). Genetic Hemangioblast Evidence The close proximity of hematopoietic cells to endothelial cells in extraembryonic and intraembryonic tissues could be due to so mething other than arising from a common


18 hemangioblast. The main support for hema ngioblast theory comes from the gene expression similarities among hematopoietic cells, angioblasts, and endothelial cells. Overlapping expression has been reported in surface antigens VEGFR2 (Flk1, KDR), VEGFR1 (Flt1), Tie1, Tie2, CD31 and CD34 and transcription factors GATA1, GATA2, SCL/Tal1, and Runx1. Disruption experiments of many of these gene s cause defects in both endothelial and hematopoietic development. Vascular Endothelial Growth Factor Vascular endothelial growth factor (VEGF) and its receptor family play a critical role in both normal and pathol ogical angiogenesis. In mamm als, there are five VEGF members (VEGFA, VEGFB, VEGFC, VEGFD, and PlGF) and three VEGF tyrosine kinase receptors (VEGFR1, VEGFR2, VEGFR3). In addition, alternative exon splicing accounts for at least four forms of VEGFA (human: VEGF121, VEGF165, VEGF189, VEGF206, mouse: VEGF120, VEGF164, VEGF188, VEGF205), each with nonidentical biological properties. In response to hypoxia, endothelial ce lls immediately respond by accumulating hypoxia-inducible factors (HIFs). Norma lly, in the presence of oxygen, HIF-1 and HIF-2 are targeted for proteasomal degradati on by an E3 ubiquitin ligase containing the VHL tumor suppressor protein (K rieg et al. 2000). The HIF-1 and HIF-2 form dimers with the constitutively expressed HIF-1. Thes e dimers (specific to various types of endothelium) activate a number of genes via hypoxia response elements (HREs), including VEGFA, HGF, PlGF, VEGFR1, VEGFR2, Tie2, SDF-1, PDGF-B, iNOS, MCP-1, IL-6, IL-8, MMP-2, MMP-13, angiopoi etin, erythropoietin, and glucose transport / anaerobic glycolytic enzymes. Acting in both a paracrine and autocrine manner, VEGF promotes a number of phys iologic changes, including enhanced


19 endothelial growth and surv ival, bone marrow-derived m onocyte mobilization (with the involvement of monocyte chemoattractant prot ein-1/MCP-1), and vascular permeability. Dysfunctional HIF-1 regulation, from inac tivation of the von Hippel-Lindau (VHL) tumor suppressor gene, is believed to be re sponsible for the capillary hemangioblastomas observed in the retina and cereb ellum of affected patients (Mole et al. 2001). Although VEGF signaling is critical in embryonic develo pment, it may not be HIF dependent in the early embryo where hypoxic conditions are rare and oxygen readily diffuses through tissue (Hiratsuka et al. 2005). VEGFR1 (Flt1) is expressed on the angi oblast-generating embryonic mesodermal cells. Later in development, VEGFR1 is f ound on endothelial cells, hematopoietic stem cells (HSC), and some monocytes. VEGFR1 may have multiple roles depending specific cues from the environment, such as specifi c ligand binding combin ations. VEGFR1 binds the angiogenic factor VEGFA as well as PlGF . The relay of the VEGFA binding signal to the intracellular mechanics may not be th at significant since VE GFR1 has only weak tyrosine autophosphorylation in response to VEGF. This is consistent with the VEGFR1-/mice. VEGR1-/mice die in uter o between days 8.5 and 9.5 from vascular overgrowth and hematopoietic blood islands co nsisting solely of angioblasts (Fong et al. 1995; Fong et al. 1999). A VEGFR1 mutant lacking the tyrosine kinase domain, however, does not result in lethality or any ove rt vascular abnormalities (Hiratsuka et al. 1998). Thus, VEGFR1 is critically important in early hematopoietic and vascular development, without depending on direct in tracellular signaling. A dditional evidence supporting this is that VEGFR1 is cleaved in to a soluble form (s Flt1) that negatively regulates the angiogenic respons e of VEGFR2 to VEGF (Fon g et al. 1995; Hiratsuka et


20 al. 1998; Fong et al. 1999). These findings are consistent with the theory that VEGFR1 acts as a ‘decoy’ receptor for VEGF. Another role of VEGFR1 involves monocyt e and HSC recruitment (Barleon et al. 1996). Monocyte migration in response to VEGF does require an intact VEGFR1 tyrosine kinase domain and was affected in VEGFR1-/mice (Barleon et al. 1996; Hiratsuka et al. 1998). Monocyt es help to facilitate a nd organize endothelial tube formation. A defect in monocyte recruitment to areas of active vascular development would therefore cause characteristic vascular de fects. If monocyte trafficking is affected in VEGFR1-/mice, it is unclear why the ty rosine kinase mutant would have no overt vascular abnormalities. In addition to m onocyte recruitment, HSC recruitment may depend on VEGFR1. Placental growth factor, PlGF, plays an important role in this recruitment. PlGF binds exclusively to VE GFR1, in contrast to VEGFA, which binds both VEGFR1 and VEGFR2. In the bone marrow, PlGF ultimately causes release and mobilization of the normally quiescent hema topoietic stem cell by upregulating matrix metalloproteinase-9 (MMP-9) causing th e cleavage of soluble Kit ligand. VEGFR2 (Flk1 / KDR) is expressed on all endothelial cells and on cells in the embryo with hemangioblast potential. VEGFR2 is considered the primary mediator of the angiogenic effect of VEGFA, involving endot helial cell proliferation, migration, and survival. Activated VEGFR2 induces the phosphorylation of phospholipase C, PI-3 kinase, Ras GTPase-activing pr otein, and the Src family. VEGFR2-/mice also die in utero between embryonic days 8.5-9.5 (Shalaby et al. 1995; Shalaby et al. 1997). In this case, embryos die showing a lack of blood islands (hematolymphopoiesis) and blood vessels (vasculogenesis). Embryonic VEGFR2-/ cells fail to contribute to primitive or definitive hematopoiesis following transplant. They also fail to is migrate from the


21 posterior primitive streak to extraembryoni c and intraembryonic sites hematopoiesis (Kabrun et al. 1997; Hi ratsuka et al. 2005). Tie1 and Tie2 Tie1 and Tie2 are also receptor tyrosine ki nases with expression patterns consistent with the hemangioblast. Tie1 and Tie2 are expressed on endothelial cells and some hematopoietic cells, including HSCs of embryonic AGM, fetal liver, and adult bone marrow as well as some monocytes. The bi nding of angiopoietin-1 (Ang-1) and its antagonist (Ang-2) to Tie2 is a vital part of embryoni c vascular development and postnatal vascular maintenance. Overexpr ession of Ang-1 has been shown to produce hypervascularity. Tie1-/mice s how deficiencies with vasc ular structural integrity resulting in edema and locali zed hemorrhage (Suri et al. 1998). Ang-1 and Tie2 mutant mice show normal early vascular development, but severe perturbation of later vascular remodeling of the primitive vascular plexus es (Sato et al. 1995; Suri et al. 1996). Similarly, in hematopoietic development, Tie1 and Tie2 seem to be dispensable in the embryonic environment, but required for comp etitive survival / e xpansion in the adult environment. Work by Arai and collea gues suggests that the quiescent 5-FU (fluorouracil) resistant SP-defined HSC is Tie2+ and that the Ang-1/Tie2 signaling pathway contributes to the quiescence and adhe rence of HSCs in their osteoblastic niche (Arai et al. 2004). Kopp and colleagues have established a model for chemotherapeutic bone marrow recovery in adults that highlight s the critical role of Tie2+ BM neovessel (Kopp et al. 2005). They show that chemothera peautic damage causes a decrease in BM vascularity and an increase in circulating VEGFA. Further evidence suggests that this VEGFA stimulates Tie2 expression on the regenerating blood vessels and promotes hematopoietic reconstitution.


22 In-Vitro Hemangioblast Evidence Some of the strongest evidence for the existence of the embryonic hemangioblast has emerged from the use of cultured em bryonic stem (ES) cells in single-cell differentiation studies (Choi et al. 1998; Nishikawa et al. 1998b; Nishikawa et al. 2000; Zambidis et al. 2005). Differentiating mouse and human ES cells can recapitulate the initial stages of the hematopoi etic program including the primi tive to definitive transition (Keller et al. 1993; Kennedy et al . 1997; Robertson et al. 1999; Zambidis et al. 2005). It has become evident that mouse ES-deriv ed brachyury positive primitive mesoderm generates hemangioblastic cells, termed BL-C FCs, which express Flk1, SCL/Tal1, CD34, and Gata1 (Robertson et al. 1999). In human ES cells, as Oct-4 is downregulated, SCL, Gata1, Gata2, and CDX4 are upregulated in what Zambidis and colleagues call the hemangioblast/primitive hematopoiesis phase of human EB differentiation, which peaks around day 10 (Zambidis et al. 2005). They demonstrate that CD 31 and CD34 increase expression at the transition into the definitive/erythro-myeloid phase, where CD45 first begins to be expressed at approximately hEB differentiation day 16. Under specific culture conditions, both hemat opoietic and adherent cells ca n be clonally derived from BL-CFCs. The adherent cells, derived fr om VEGF treatment, are phenotypically identified as endothelial based on makers for CD31 (PECAM-1), Flk1 (VEGFR2), Tie2, Flt1 (VEGFR1), and Ac-LDL uptake (Choi et al. 1998). Nishikawa and colleagues have also observed in-vitro hematopoietic activ ity from ES-derived Flk1+ VE-Cad+ CD45endothelial cells. BL-CFCs have also been should to generate vascular smooth muscle cells, in response to PDGF (Yamashita et al. 2000; Ema et al. 2003). The additional potential of these hemangioblastic cells is not surprising considering the close interdependence that endothelium and mural cel ls have with one another, in vivo. Flk1


23 can be used enrich for BL-CFCs in ES-deriv ed cultures (Faloon et al. 2000), however not all Flk1+ cells are BL-CFCs. In fact, Flk1+ cel ls have also been shown to give rise to cardiomyocytes and skeletal muscle (Motoike et al. 2003). The poten tial of Flk1+ cells in these systems is very suggestive of a multi potent stem cell with significant plasticity. Characterization of the genetic program that underlies the lineage commitment steps of the hemangioblast (and more primitive stem cells) is paramount in advancing our understanding of initial phenotypic observati ons. SCL/Tal1, Runx1/AML1, Gata1 and Lmo2 have emerged as important genes in BL-CFC commitment and differentiation. SCL/Tal1 expression in Brachyury+ Flk1+ cells correlates with a transitional commitment of the BL-CFCs towards a hematopoietic fate (Chung et al. 2002; D'Souza et al. 2005). Indeed, targeted disruption of SC L in mice revealed that it was necessary for blood development (Shivdasani et al. 1995). SCL (Tal1) appears to have a second role in endothelial maturation and proliferation and perhaps not necessarily one in the commitment decision (D'Souza et al. 2005; Dooley et al. 2005). Although Tal1-/ES cells fail to generate hematopoi etic or endothelial cells in vitro, endothelial cells are not entirely absent from Tal1-/embryos (Ema et al. 2003). Lmo2 seems to function in a manner similar to SCL. Lmo2 mutants are E10.5 lethal due to an initial failure of the hematopoietic program (Warren et al. 1994; Gering et al. 2003). Not surprisingly, SCL functions in a transcriptional complex w ith Lmo2 (Wadman et al. 1997). Gering and colleagues found that in the ab sence of specific hematopoietic inducers, such as Gata1, SCL & Lmo2 induced hemangioblasts were dir ected towards endothelial differentiation (Gering et al. 2003). Runx1 is a transcription f actor expressed in a pa ttern that correlates with commitment to BL-CFCs. Runx1 muta nts generate substantially fewer BL-CFCs and die between E11 to E12 from a complete block in definitive hematopoiesis (Okuda et


24 al. 1996; Lacaud et al. 2002). Runx1 expres sion identifies three pools of embryonic HSCs: CD45+ type, CD31+ CD45endotheli al type, and CD31/ VE-CadCD45mesenchymal type (North et al. 2002). Runx1 has also recently been identified in adult HSCs (North et al. 2004). Adult Hemangioblast Adult Hematopoietic Development Adult hematopoiesis begins in the bone ma rrow cavities. This process is generally illustrated as a progressive lineage commitment from primitive stem and progenitor cells (Kluger et al. 2004). One of the most well ch aracterized stem cells is the adult HSC, which primarily resides in the bone marrow ni che. Highly purified populations of HSCs have been obtained using a variety of techniques. The technique of multicolor flow cytometry now allows for a nearly homogenous enrichment for the HSC. The surface phenotype of the HSC, in the mouse, is lineage marker-negative, c-Kit(CD117)-positive, Sca-1(Ly6A/E)-positive, and Thy-1-low (KTS L) (Shizuru et al. 2005). Use of a lineage cocktail ensures the exclusi on of granulocytes (Gr-1), macrophages (Mac-1), B-cells (B220,CD5), T-cells (CD3, CD4, CD5, CD8), a nd erythrocytes (Ter119). Enrichment for the human HSC uses a similar scheme except th at the mouse Sca-1 is generally replaced with human CD34 or CD133. The use of CD34-positivity to enrich for the human long-term repopulating (LTR or LT) HSC remain s somewhat controvers ial. In the mouse, however, the LTR-HSC is found in the CD 34-/low fraction of adult bone marrow, although at other points in development th e HSC expresses CD34 (Gekas et al. 2005; Zhong et al. 2005). CD133/Prominin-1 has also emerged as a popular human stem cell antigen (Shmelkov et al. 2005). Specific dye (Ho echst 33342) efflux char acteristics, or a


25 so-called side-population FACS profile, have also been used to enrich for both human and mouse HSCs with a comparable purity to KTSL enrichment (Goodell et al. 1996). In mice, the transplantable, functiona lly defined, HSC falls in the pool of VEGFR1-positive, VEGFR2-negative cells (Hatto ri et al. 2002). This is supported by the fact that PlGF, which selectively inter acts with VEGFR1, improves post-radiation hematopoietic recovery in comparison to VEGFA, which intracellularly signals via VEGFR2 (Rafii et al. 2003). Similarly, a blocking antibody to VEGFR1 hinders postradiation recovery (H attori et al. 2002). In humans, VEGFRs may have different functions than in mice. Specifically, VE GFR2 was found to be expressed on human CD34+ NON/SCID mouse repopula ting cells (Ziegler et al. 1999). Lastly, Tie2 also defines a subset of cells containing the l ong term repopulating HSC (Arai et al. 2004). Traditionally VEGFR1, VEGFR2, and Tie2 have been consid ered endothelial specific. Identification of these receptors on HSCs cer tainly suggests overlapping hematopoietic and endothelial programs, even in the adult. The Tie2+ adult bone marrow HSC is 5-FU resistant, quiescent, and found in the side-population gate. This Tie2+ HSC is found tightly adherent to the microenvironment created by osteoblasts that are seated on the endosteal surface of the marrow cavity (Calvi et al. 2003; Zhang et al. 2003). The adhesion to fibronectin and collagen is mediated, in part by Ang-1 secreted by the osteoblasts (Sat o et al. 1998; Arai et al. 2002). From their quiescent osteoblastic niche, HSCs can be mobili zed into a vascular niche. In this niche, Tie2+ bone marrow endothelial cells have been ob served to be in close proximity to these CD45+ Tie2+ HSCs (Arai et al. 2004). One theory is that mob ilization occurs in response to the chemokine signals VEGFA and PlGF , which interact with VEGFR1 on the VEGFR1+ c-Kit+ HSC causing upregulation of MMP-9. As mentioned earlier, MMP-9


26 cleaves membrane-bound Kit ligand into sol uble Kit ligand, which induces cycling and increases HSC (along with endothelial progenitor cell) motility toward the BM sinusoids, the vascular niche (Hattori et al. 2002; Heissig et al. 2002) . Once in circulation certain vascular beds may provide a temporary suppor tive niche for the ci rculating HSC (Ohneda et al. 1998; Li et al. 2004). Adult Vascular Development Following in-utero formation and conti nuing throughout postnatal life, tissues continue to require repair and remodeli ng. Vascular tissue, similar to the dynamic production and mobilization found in hematopoietic tissue, must response to a variable environment, in order to ensure survival . The vascular system responses to both physiologic conditions, such as the fluctu ating oxygen demand of menstruation and exercise-induced skeletal muscle hypert rophy, and pathologic conditions, such as revascularization of ischemic tissue. Postnatal as well as prenatal vessel growth is a tightly regulated event and involves a complex interplay between proangiogenic and antiangiogenic factors. Howeve r, relative to prenatal e ndothelium, nonstressed, postnatal, especially adult, endothelium is much mo re quiescent. For many years, it was believed that adult vascular development was exclus ively angiogenic, where preexisting vessels either enlarged or produced new sprouts from existing mature endothelial cells. Although earlier evidence existed, the dogmatic shift re ally began when the source of fallout endothelialization of bypass graf ts were identified as non-local (Shi et al. 1994; Wu et al. 1995). No longer were questions regarding the origins of adult endot helial cells limited to an obscure realm of developmental biology, but were recognized as fundamental to the advancement of the medical therapies for a myriad of vascular diseases. Additionally, it became evident that tumors often require a complex vascular system, which, in some


27 cases, also have a non-local endothelial contri bution (Maniotis et al. 1999; Hendrix et al. 2001; Hendrix et al. 2002). The dogma that angiogenesis was the only process active in a dult vascularization, led to conjecture that circ ulating endothelial cells (CEC s) which had emerged from existing endothelial structures were cont ributing to distant vascular structures. Identification of an angioblast-like endotheli al progenitor cells (EPCs) and circulating EPCs (CEPs) in the adult, however, confirme d that the dogma must be revised to include the role of these newly identified progenito rs (Asahara et al. 1997; Rafii and Lyden 2003). This process is now identified as pos tnatal vasculogenesis, a type of adult neovascularization, by bone marrow derived EP Cs. In humans, EPCs (the generic term EPC will be used for bone marrow and circulat ing endothelial progenitor cells) have been characterized as CD133+, CD34+, and VEGF R2 (KDR)+ (Rafii and Lyden 2003; Iwami et al. 2004). In the mouse, at least one type of EPC is VEGFR2+ and, like the HSC, Sca-1+, c-Kit+, lineage-negative. Exact phenotyp ic identification of the EPC is difficult due to the shared expression of VEGF R2, Tie2, VE-cadherin, CD34, CD146, and E-selectin between various stages of EPCs and mature endothelial cells. Also, various hematopoietic cells share phenot ypic characteristic with EP Cs, including expression of CD34, CD31 (PECAM), Sca-1, Tie2, vWF, VEGFR1, and uptake of Ac-LDL (Figure 1-3). In-vitro culture conditions have also been used to distinguish EPCs from hematopoietic and mature endothelial cells. EP Cs can be culture expanded in endothelial media containing VEGF, FGF-B, and IGF-1 whic h identifies an early and late type of EPC (Asahara et al. 1997; Hu r et al. 2004). Short-term cultu re of mononuclear fractions results in adherent cells that resemble a population of CD45+/l ow CD14+ monocytic EPC, which has been observed by several gr oups as having endothe lial-like properties


28 (Harraz et al. 2001; Schmeisser et al. 2001; Gulati et al. 2003; Kuwana et al. 2003; Rumpold et al. 2004). There has been specu lation that, in-vivo, this cell-type may temporarily contribute to a microenvironment that promotes the development of a more stable endothelial structure. Long-term cu lture of mononuclear fractions results in adherent, late-EPCs or so-called outgrowth-e ndothelial cells. These cells phenotypically resemble human umbilical vein endothel ial cells (HUVEC) with a lack of CD45 expression and a high expression of VE -cadherin, CD31, VEGFR2 (Flk1/KDR), VEGFR1(Flt1), eNOS, and vWF (Hur et al. 2004). Figure 1-3. Phenotypes of bone marrow st em cells and progenitors. This is a progression illustrating the phenotyp ic fate of the bone marrow hemangioblast. The pluripotent stem cell is only hypothetical. The lists of phenotypic surface antigen are not mean t to specific one exact phenotype, but rather list the surface antigens that have been identified in isolation or combination on these cell populations from a multitude of sources. Furthermore, the arrows represent a hierarchy and are not meant to exclude any bidirectionality.


29 Both culture expanded EPCs and freshly isolated EPCs have been shown to incorporate into in-vivo va scular structures, assume an endothelial phenotype, and improve vascular outcome of ischemic conditions (Kawamoto et al. 2002; Khakoo and Finkel 2005). The medical releva nce of circulating EPCs may be quite broad in light of studies correlating the numbers in circulation with a variety of factors. When evaluating these studies on an individual basis, the phenotype of the detected EPC must be considered. In the more rigorous huma n studies, the CD34+ AC133+ VEGFR2+ EPC phenotype is used. Even this rigorous phe notype does overlap w ith the human HSC phenotype. Use of statin drugs, HMG-CoA re ductase inhibitors used to control hypercholesterolemia, positively correlates with increased cEPC number (Dimmeler et al. 2001; Llevadot et al. 2001; Urbich et al. 2002; Walter et al. 2002; Urbich and Dimmeler 2005). Angiotensin II receptor antagonists (Bahlmann et al. 2005) and antidiabetic peroxisome proliferator-activat ed receptor-gamma agonists ha ve a similar effect on EPCs (Wang et al. 2004). Both exercise (Adams et al. 2004; Laufs et al. 2004b; Rehman et al. 2004; Ciulla et al. 2005) and es trogen (Iwakura et al. 2003; Strehlow et al. 2003) also correlate with cEPC number. Changes in certain hematopoietic cytokines may be involved, at least in part, in these observations. VEGF (Asahara et al. 1999; Kalka et al. 2000a; Kalka et al. 2000b), SDF-1 (Yamaguchi et al. 2003; Butler et al. 2005), Ang-1 (Hattori et al. 2001; Moore et al. 2001), G-CSF (Peichev et al. 2000; Kong et al. 2004; Powell et al. 2005), GM-CSF (Takahashi et al. 1999), and EPO (Bahlmann et al. 2003; Heeschen et al. 2003; Bahlmann et al. 2004) al l have the capacity to mobilize cEPCs. The previous examples demonstrate that many beneficial physiologic states correlate with elevated numbers of cEPCs. Cardiovascular damage from myocardial infarction (Shintani et al. 2001; Massa et al. 2005), unstable angina (George et al. 2004),


30 congestive heart failure classes I-II (Hrist ov and Weber 2004), and vascular trauma (Gill et al. 2001; Nakatani et al. 2003; Hunting et al. 2005) also correlates with circulating progenitor number. Cardiovascular (CV) risk factors, however , tend to inversely correlate with cEPC number and function. This includes both a cumulative assessment (Hill et al. 2003) as well as the individual risk factors of CRP levels (Verma et al. 2004), diabetes (Tepper et al. 2002; Fadini et al. 2005), sm oking, family history, and hypertension (Vasa et al. 2001). Although the correlation between EPC mobilization and protective/repairing CV states or harmful CV states seems we ll delineated, it may depend on dose or degree of injury. Very high levels of statin drugs , for example, block angiogenesis and induce endothelial cell death (Urbich et al. 2002). Also, in congestive heart failure cl asses III-IV, cEPC levels are decreased in contrast to classes I-II (Hristov and Weber 2004). Ultimately, as the genetic programs su rrounding EPC and HSC commitment and differentiation become deciphe red, a clearer, more reliab le genetic fingerprint will emerge. Fundamental to our understandi ng of the adult hemangioblast is an understanding of the exact physio logic role of the EPC or EP C-like cells and the factors regulating their mobilization in both heath and disease. Adult Hemangioblast Activity The co-emergence of hematopoietic and e ndothelial cells from the YS, AGM, and placenta, which at various points in devel opment share phenotypic characteristics and genetic programs, certainly points to a co mmon embryonic hemangioblast. Clearly, this theory gained support from the finding that a mouse ES-cell derive d Flk1+ BL-CFC has both endothelial and hematopoietic activity (Choi et al. 1998) . But until recently, there was little to support that this hemangioblastic cell existed in the adult. With the discovery of a circulating, bone marrow-derived adult e ndothelial progenitor cel l, the existence of


31 an adult hemangioblast could no longer be immediately rejected. Using CD34+ KDR(VEGFR2)+ enrichment, whic h is also used to identify the EPC, Pelosi and colleagues observed clonal hemangioblast act ivity in human cord blood and adult bone marrow (Pelosi et al. 2002) . Alternatively, Loges a nd colleagues have observed approximately 2% of mobilized AC133+ cells with an in-vitro he mangioblast capacity (Loges et al. 2004). Because of the possibili ty that certain rare cells may have a hemangioblast capacity when exposed to poten tially non-physiologi c culture conditions, an in-vivo demonstration of this clonal activity was still required. This was first shown, in the mouse, using HSC transplantation in a retinal neovascularization model (Grant et al. 2002). In this study HSCs from a highly enriched Sca-1+, c-Kit+, lin(SKL) bone marrow clona l (single-cell transplant) and self-renewing (serial transplant) population provided l ong-term multilineage reconstitution of the hematopoietic system and contributed to the endothelial network of a functional, perfusable retinal vascular bed. Th e overexpression of VEGFA along with photocoagulative injury is required to mobilize bone marrow derived Flk1+ EPCs, which based on FACs ploidy analysis are not produc ts of fusion. Although variable contribution is observed in this model, these HSC derived cEPCs appear to contribute to up to 20% of the observed preretinal vascul ar tufts. Bailey and collea gues also observed low-level CD31+, vWF+, Ac-LDL uptake+ endothelial co ntribution in various adult mouse organs from single transplanted SKL-enriched HSCs (Bailey et al. 2004) . Jiang and colleagues also observed a 2% contribution in gut and sk in of sex-mismatched transplanted human patients (Jiang et al. 2004). Transplanted bone marrow cells, although not necessarily clonally pure, have also been observed contributing to neovasculoge nesis in limb ischemia, postmyocardial


32 infarction, vascular grafts, atherosclerosi s, and tumor growth (Rafii and Lyden 2003; Iwami et al. 2004; Hunting et al. 2005; Liu et al. 2005; Peters et al. 2005). In the uninjured and unmobilized state, cEPC levels are quire low (Aiche r et al. 2005). It is interesting to conjecture that the origins of vascular turnover and repair depend on the degree of injury. In the uninjured or mild ly injured states local angiogenesis may predominate. With more severe injury, incl uding some types of pro-angiogenic tumors, neovascularization from bone marrow derived cEPCs may be engaged as a supplement. Similar reasoning has been used in observati ons of HSC to muscle plasticity (Sherwood et al. 2004). In addition to single-cell transplants, which are technica lly quite challenging, another technique that has been used to demonstrate the clonal origin of cells is permanent retroviral tagging of stem and proge nitor cells (Nolta et al. 1996; Dick et al. 2001; Kiem et al. 2004). Nolta and colleague s used this technique to confirm the existence of a common progenitor for lymphoid and myeloid cells. This technique is used quite frequently, along with temporary dye marking and Cre-LoxP recombination in developmental studies to track lineage commitment. Lineage tracking studies are necessary to identify well characterized, bone marrow derived, mature endothelial cell contribution in which postnatal va sculogenesis has been stimulated. In all of the examples discussed in th is section, hematopoietic cells are seen contributing to endothelium after transplant, but, remarkably, the i nverse experiment has also been done and has produced an interes ting result. As little as 10mg of endothelium purified from adult mouse aorta and vena cava, and transpla nted under the kidney capsule, has been shown to be radioprotectiv e and to contribute to the tissue surrounding post-radiation spleen CFUs (Montfort et al. 2002). Although long-term hematopoiesis


33 was not derived from endothelial cell transp lants, the radioprotection afforded to the hosts is remarkable and reminiscent of the endothelial support of hematopoiesis in the fetal YS and AGM. Adult Hemangioblast Medical Therapy Tumor vascularization Aberrant vascular maturation is character istic of a number of human diseases. The vascular network that develops to nouris h solid tumors is, un fortunately, one common example. In solid tumor vasculature, br anching patterns, vessel diameter, vessel permeability, and mural cell organization are characteristically abnormal (Jain 2005). Whereas normal vascular beds develop with a hierarchical pattern of branching, tumor vascular beds develop with an apparently random pattern of branching (Brown et al. 2001). The rapid and irregular growth of solid tumors, at least pa rtially, leads to the unbalanced release of proangioge nic factors and contributes to the disruption of normal vascular growth. Also, within this chaotic microenvironmen t, endothelial cells tend to form vessels with uneven diameters and cel l-cell junctions. The lack of normal endothelilal cell-cell junctions allows gaps to form in the vessel, allowing perturbations in capillary fluid retention. One of the net consequences of al l these vascular deficiencies is the creation of aci dotic, hypoxia patches of tumor tissu e. Paradoxically, theses regions of hypoxia that are so damaging to most cells, including the pa tient’s immune cells, often prove beneficial for the survival of the tumo r cells during medical therapy. Tumor cells in these regions have a low metabolic activity and are therefore more resistant to radiation and chemotherapy. Furthermore, the environm ent creates selective pressure for the survival of genetically unstable tumor cells , which are better able to metastasize and evade the immune system.


34 The relevance of tumor vascular development to those interested in the adult hemangioblast lies in the origin of the cells that are recruited (or not) to form this vascular network. In some cases, the normal preexisting vascular network extends into the growing tumor with bone marrow derive d endothelial cells (Lyden et al. 2001; Murayama et al. 2002; Stoll et al. 2003; Jode le et al. 2005). The expl anations that have been used to describe these observations have mainly centered on tumor endothelial contribution from circulating bone marrow derived EPCs that are responding to angiogenic signals such as VEGFA and Ang1. Peters and colleagues have recently screened multiple types of cancers at various time points in human patients that had received sex-mismatched bone marrow tran splants. Their assessment was that bone marrow derived cells contributed to a pproximately 5% of CD45vWF+ tumor vasculature. How significant is the contribution of bone marro w derived cells and what is the exact function of those cells are two very important questions. Satisfying answers will likely emerge as the physiol ogic roles of vascular-relate d cells (mature endothelium, smooth muscle, pericytes, adherent leukocytes ) are well delineated. Another question that lingers is one revolving around the transplantabil ity of the EPC. In order to differentiate bone marrow cells from local tumor cells, most studies have harnessed permanent differences in transplanted cells, for exampl e sex chromosome mismatches or transgenic GFP expression. The only cells that s how long-term self-renewing bone marrow engraftment following myeloablation are HS Cs. The observation that EPC activity is observed from a distinguishable population of tr ansplanted cells at time points generally considered beyond the point of progenitor die of f, is suggestive of hemangioblast activity. Also suggestive of adult he mangioblast activity is recent work that identifies a specific BCR-ABL fusion gene in endotheli al cells (Gunsilius 2003) and Flk1+ bone


35 marrow hemangioblastic cells (Fang et al. 2005 ) of CML patients. More generally, this illustrates that a genetic mutation can occur in a stem cell that is far removed from the point in lineage commitment where that mutation causes a noticeable phenotype. The presence of a cancer stem cell generates a number of confounding problems in realm of cancer therapeutics (Marx 2003). Fi rst, without eliminating a mutant cancer stem cell, cytotoxic treatments that destroy progen itor populations would only temporarily remedy the malignant condition. Second, the fact that an endothelial cell, wh ich is not typically targeted by treatments for hematological mali gnancies, can serve as a reservoir for an unstable genetic mutation has been largely i gnored. Third, because ge netically unstable cancer cells have demonstrated the stem celllike capacity to self-renew (Jamieson et al. 2004), other cryptic and potentially quies cent phenotypes may serve as cancer “hideouts”. In contrast to the darker side of can cer stem cells, medical therapies that specifically target tumor vascular growth have had promising results. In a recent review, Jain illustrates a widely held view that tumors have a microenvironment that is improperly skewed towards a proangiogeni c state (Jain 2005). His proposal that “normalization” or a rebalancing the proa nd anti-angiogenic influences within tumors would allow better therapeutic targeting of can cer cells is supported by work with human tumor xenografts (Winkler et al. 2004). In th is study, Winkler, Jain, and colleagues, using DC101, a VEGFR2-specific blocking antibody, created a normalization window, during which time human glioblastoma into mouse xe nografts were more ra diosensitive than controls. Remarkably, after blocking VEGFR 2, they observed an upregulation of Ang-1, which recruited Tie2+ pericytes. These peri cytes seem to play a role in basement membrane debulking via matrix metalloproteinase activity. Pa radoxically, it is likely the


36 decreased tumor cell hypoxia that renders these cells more sensitive to radiation. Pharmaceutical companies have recognized the medical potential of targeting tumor angiogenesis via the VEGF/VEG FR interaction and develope d at least 16 inhibitors which are currently working their way through clinical trials (Hicklin and Ellis 2005). The results from a phase III trial of addi ng bevacizumab, a monoclonal antibody against VEGF, to standard therapy for metastatic colorectal cancer has resulted in a significant improvement of patient survival (Hurwitz et al. 2004). After release of this data, the FDA rapidly approved bevacizumab for this use. Future strategies for targ eting tumor angiogenesis may involve modulation other angiogenic signals, like stromal cell derived factor-1 (SDF-1), angiopoietin-1 (Ang-1), fibroblast growth factor-2 (F GF-2), platlet derived growth factor (PDGF), or monocyte chemoattractant protein-1 (MCP-1). SDF-1 (CXCL12) is found upregulated in a number of cancers including ovarian, breast, and glial carcinomas (Bachelder et al. 2002; Salmaggi et al. 2004; Kryczek et al. 2005). CXCR4, the SDF-1 receptor, has been found on circulating EPCs, in addition to other he matopoietic cells (Yamaguchi et al. 2003). Furthermore, it has been demonstrated th at the SDF-1/CXCR4 signal is critical for recruitment of the EPC into the vascular st ructures of a retinal angiogenesis model (Butler et al. 2005) and may also be involve d in recruitment for tumor angiogenesis. Upregulation of FGF-2 has also been obs erved in a number of cancers, including prostate, lung, skin, and astrocytic carci nomas (Kurtz et al. 2004; Kwabi-Addo et al. 2004; Stiver 2004; Van der Auwera et al . 2004). Compounds interrupting the normal FGF-2 signaling pathway may also prove to be viable anti-tumor treatments (Bossard et al. 2004; Tong et al. 2005). MCP-1 is upregulat ed in squamous cell, glioma, melanocytic, and breast carcinomas (Ueno et al. 2000; Nesb it et al. 2001; Ohta et al. 2003; Platten et


37 al. 2003; Koide et al. 2004). MCP-1 is beli eved to contribute to tumor angiogenesis primarily by the release of angiogenic factors from r ecruited monocytes. However, endothelial cells also ex press the MCP-1 receptor, CCR2, and respond to MCP-1 suggesting a more elaborate effect (Sal cedo et al. 2000). The PDGF/PDGFR signaling pathway has been implicated in several hum an tumors and blockade has shown some early successes in reducing ovarian, colore ctal, and glioblastoma multiforme tumor growth (Apte et al. 2004 ; Roberts et al. 2005). Another future strategy for modulating tu mor angiogenesis involves using the EPC as a “trojan horse” by poisoning th em with a suicidal program prior to transplant (Wei et al. 2004). Due to the paucity of well-character ized adult endothelial progenitor cells, this group chose to use murine embryonic EPCs, which have a robust proliferative capacity as well as the ability to integr ate into adult vasculature. Embryonic EPCs (eEPCs) were given a gene to convert the harmless pro-drug 5-fluorocytosine (5-FC) into the cytotoxic 5-fluorouracil (5-FU). Transplanted eEPCs inte grate into the vascul ature of the poorly vascularized metastases of lung carcinomas and, in the pres ence of 5-FC, produced 5-FU, which was cytotoxic to surrounding tumor cells through a bystander e ffect. Mice treated in this manner showed a decrease in the number of lung metastases and an increased survival time. Although the ther apy was significantly less po tent on highly vascularized metastases, this technique, along with the others mentioned in this section, clearly demonstrates the creative potential of EPC therapies for treating tumors and other diseases involving aberrant vascularization. Nontumor vascularization One of the major driving forces behind the efforts to identify and characterize the adult hemangioblast is to use knowledge about commitm ent, proliferation, and


38 mobilization in medical therapies. Postnatal vasculogenesis can be both physiologic and pathologic. Unfortunately, sometimes the tran sition from physiologic to pathologic is difficult to pinpoint. Such is the case with the body’s response to vascular damage caused by the high blood glucose of diabetes. Chroni c exposure to high glucose levels causes damage to endothelial cells throughout the body, in particular the microvasculature of the retina. This damage results in fluid leakag e and tissue swelling, which eventually causes vision distortions due to macula r damage. Over time, the ischemia of retinal tissue causes new fragile blood vessels to grow. These fragil e vessels sometimes aberrantly grow into the vitreous where they are prone to hemorrh age. These types of vessel are the pathology behind proliferative diabetic retinopathy and require laser pho tocoagulation to control. Although attempting to remedy the ischemic s ituation, these aberrant new blood vessels also become damaged by the high glucose. This ongoing cycle of tissue damage very often leads to complete blindness. A mouse mo del of the preretinal “angiogenesis” (the fragile vessels growing into the vitreous) that is characterist ic of diabetic retinopathy has been used by our lab to identify and modulate factors that affect th is aberrant vascular growth (Guthrie et al. 2004; Butler et al. 2005). Butler and colleagues showed that vitreous levels of SDF-1 correlate with the severity of diabetic retinopathy in humans. Furthermore, they used blocking antibodies to SDF-1 to inhibit the formation of the diseased preretinal vascular tufts in th e mouse model. Guthrie and colleagues have demonstrated bone marrow derived EPC-genera ted vascular branching defects in the retina correlate with abnormalities of nitric oxi de. It is clear that VEGF, SDF-1, and nitric oxide are critical for mobilization of bone marrow EPCs to the retina to form these vessels.


39 Diabetic retinopathy is just one of ma ny potential targets for the therapeutic manipulation of adult vasculogenesis. Physic ians grapple with the complications of ischemic disease quite commonly. Research in vascular biology has led us to a crossroads that requires an exacting integr ation of the hematopoietic cont ribution into adult vascular development, maintenance, and remodeling. In the long-term, this involves assessing the vasculogenic significance and defining the comm itment steps of the now identified adult hemangioblast. In the short-term, this i nvolves continuing to decipher the complex cellular interactions and signaling events that control the endothelial outcomes of the transplantable, bone marrow derived EPC. M yocardial infarction, stroke, and chronic limb ischemia would benefit from therapies that could enhance EP C-derived collateral vascular growth. Diabetic retinopathy, cancer , and rheumatoid arthritis would benefit from therapies that prevent and cause regr ession of inappropriate vascular growth by modulation of the anti-angiogenic/anti-EPC gr owth factors. EPC enhanced vascular stability of transplanted organs may eventu ally improve long-term outcomes. Transplant of EPCs with genetic modificat ions may be able to correct vascular diseases with both genetic and environmental etiology. Hemat opoietic recovery and expansion after myeloablative procedures may be improve d by modifying therapies to protect the supportive vascular networks. The potential medical therapies are many. There is great beauty in this potential for mankind. There is also great beauty in simply deciphering the complex systems in place to regulate the development and later recruitment of the hemangioblast and its closely related progeny, the HSC and the EPC. The only obstacles are those unsightly void s, the unanswered questions that must be answered. Do fetal hemangioblasts emerge from distinct mesode rmal stem cells? Do adult hemangioblasts emerge from mesodermal or some pluripot ent stem cells? Is the bone marrow the only


40 location of these stem cells? Can EPCs emer ge from other niches? Are HSCs the only transplantable cells with hemangioblast activity ? Is there a direct pa th of differentiation between EPCs and monocytes? What are the di stinct roles of inflammatory cells, EPCs, pericytes, and smooth muscle cells in angi ogenesis and postnatal vasculogenesis? How does the vascular bed make the choi ce between angiogenesis and postnatal vasculogenesis? The questions are many, as well. It is by answering these questions that the beauty of the hemangioblast will be revealed.


41 CHAPTER 2 GENETIC MARKING TO TRACK CLONALITY After the describing the hemangioblast, I wi ll in Chapter 2 examine the history of a technique commonly used to characterize th e lineage potential of stem and progenitor cells, in vivo. Various transitory, usually dye -based, marking strategies suffer from the disadvantage of being diluted with each cycl e of cell division. Also, because the dye can sometime leak out of cells, there is concern about the sensitivity (if the dye is lost) and specificity (if the dye is phagocytosed by another cell) of the labeling procedure. Especially in longer-term studies, genetically marking cells in some manner offers a permanent, non-diluting way of tracking its pr ogeny throughout the lif e of the organism. Since it was not until recently that gene therapeutic appr oaches were developed, early studies used preexisting geneti c variations divided a populati on into two or more subsets of “labeled” cells. These techniques have been particularly useful in characterizing the potential of the hematopoietic stem cell. Non-Retroviral Marking Studies Genetic marking studies have been used to study the kinetics and lineage potential of hematopoietic stem cells for nearly thirty years. Many of the pi oneering studies relied on differences in the expression of polymo rphic genes, such as glucose-6-phosphate dehydrogenase (G6PD) (Fialkow et al. 1977). Women that are heterozygous at the G6PD locus (located on the X chromosome) have cel ls marked with a specific G6PD isoenzyme (due to X inactivation), creati ng a type of binary tag. Fial kow et al. noticed that in patients with chronic myelocytic leukemia, a ll granulocytes, erythroc ytes, platelets, and


42 monocytes expressed a single isoenzyme vers us mixed expression in normal patients. They concluded that CML is caused by a muta tion in a cell capable of forming all these lineages. The fact that certain cancers originat e in mutations in proge nitors or stem cells is critical for selecting effective therapeutic options. Other pioneering studies involved transplantation of cells with allelic variation, such as Ly5.1/Ly-5.2 (Smith et al. 1991). In the work of Smith et al., they used mi xed Ly-5.1/Ly-5.2 cell transplants in limiting dilutions to calculate that 1 in 13 Thy-1lo Sca-1+ Linenriched bone marrow cells contributed to the blood, in excess of 1%, following transplant. By increasing the cell dose, a corresponding increase in the number of clones contributing to hematopoiesis was observed. After 9 weeks, only a third of the clones were found to be active. From their data, they calculated the LT-HSC frequenc y in bone marrow at 1 in 0.5 to 1.3 x 105 cells. This frequency has been confirmed and is now widely accepted in literature. Both of these studies highlight the remarkable observa tions that arise from a clonal analysis of stem cell progeny. Early Retroviral Marking Studies The use of recombinant retroviruses capable of introducing a gene into hematopoietic stem cells (Williams et al. 1984) paved the way for using them to study clonality in the hematopoietic system (D ick et al. 1985). Because the location of retroviral genomic integra tion is considered somewh at random, the number of identifiable tags increased enormously from the relatively simplistic binomial tags of earlier work. Because of this, the resolution of clonal marking was dramatically increased. Also, since the early use of binomial tags relied heavily on statistics to analyze the results, the multiple tags of “random” retroviral integration site analysis allowed for a more obvious argument. Lemischka et al. firs t used the retroviral system to mark


43 hematopoietic cells. They showed that specif ic retroviral tags we re identified in both multilineage (virus infected stem cells) and rest ricted lineage (virus infected progenitors) hematopoiesis (Lemischka et al. 1986). Furthe rmore, their study showed that in mice, generally 1 or 2 clones are ac tive at any one time (oligoclona lity) and that those clones change over time (clonal succession). Later, Jordan and Lemischka demonstrated fluctuating clones in the first 4-6 months after transplant, followed by quad-lineage (myelopoesis, lymphopoiesis, erythropoiesis , and megakaryocytopoesis) hematopoiesis from a small number of totipotent clones (Jordan and Lemischka 1990). Using statistical modeling, the initial peri od of clonal variability was determined to be consistent with a mix of stem and progenitor cel l tagging. Since PCR was not r eadily available prior to the 1990’s, these early works used Southern blot s to probe for the retrovirus in uniquely sized bands of dige sted genomic DNA. Retroviral Clonality of Human Hematopoiesis in Mice The ability to amplify specific segments of DNA by PCR changed the way that retroviral integration sites were analyzed. In 1994, Rill et al. developed a strategy of circularizing genomic DNA digests in orde r to PCR amplify, with outward extending primers, a portion of the genomic DNA located adjacent to the inserti on site (Rill et al. 1994). Using this inverse PCR technique, th ey demonstrated that autologous bone marrow transplantation for treatment of neuroblastoma, a solid tumor, resulted in a reinfusion of tumorigenic cells. Nolta et al. adapted this inverse PCR strategy to examine lymphoid and myeloid clones from immune-def icient mice that had been transplanted with CD34+ human bone marrow. Oligoclonal ity was observed in this study with an average of 1.3 clones contribu ting to T-cells and 3.6 clones c ontributing to myeloid cells.


44 However, overlapping insertion sites between these two lineages confirmed the existence of a common progenitor. The study by Nolta et al. was one of the fi rst to examine human hematopoiesis at a clonal level. One of the major limitations of th is and other early retr oviral tagging studies was the limited efficiency of retroviral tran sduction of stem cells (particularly human stem cells). On the other hand, an improve ment made by study was the switch to PCR. Since PCR-based approaches for detecting inte gration sites had sens itivities of between 100 and 1000 cells, they were advantageous over Southern blot detection, which required a larger number of cells (Hui et al. 1998; Schmidt et al. 2001). However, the Southern blot technique is still used for retroviral insertion site detection and is widely accepted. Guenechea et al. used retrovira l clonal tagging of enriched human cord blood cells whose progeny after transplant were characterized by Southern blot (Guenechea et al. 2001). Their work suggested the existence of huma n stem cells with va riable self-renewal potential. It is now know th at the bone marrow contains a mix of long-term HSCs, short-term HSCs, and intermediateand short-term progenitors. They observed an average of 3.6 clones present in whole bone marrow of transplanted (oligoclonal). Retroviral Clonality of NonHuman Primate Hematopoiesis Studies using non-human primates originally suffered from the same difficulties of those using human stem cells. Retroviral mark ing levels were below the limit of detection for both PCR and Southern blot (Van Be usechem and Valerio 1996). Although various cytokine stimulatory regimes were tried during in-vitro culture and retroviral transduction of primate CD34+ bone marrow, it was not unt il stem cell mobilization with SCF and G-CSF that engraftment was sufficient for analys is. A similar effect is seen in mice that have been treated with fluorouracil (5-FU), since both of these treatments have been


45 shown to be beneficial in mouse HSC tran sduction (Van Beusechem and Valerio 1996). In addition to inverse PCR, another PCR de tection of retroviral integration sites was recently developed, called LAM-PCR, which uses extension primer tag selection, ligation, and solid-phase nested PCR (Sch midt et al. 2001). Using SCF and G-CSF mobilization with LAM-PCR detection, D unbar et al. have demonstrated, that hematopoiesis is polyclonal in larger anim als (Schmidt et al. 2002; Kiem et al. 2004; Laukkanen et al. 2005). Mean clone numbers were reported as 17, 15, and 6 in rhesus macaques, baboons, and dogs, respectively. It is unclear what factor contributes to this higher clonality. One feasible expl anation is that as animal size increases, the number of stem cells needed to supply hematopoi esis for the larger volume of blood, may correspondingly increase. Anot her explanation may relate to the specific culturing conditions in IL-3, IL-6, megakaryocyte gr owth and development factor (MGDF), and Flt-3 ligand, which in some way promote more polyclonality. A third explanation may be that the increased sensitivity of LAM-PCR si mply detects more of the rare clones not detected with less sensitive techniques. Dunbar et al. have also investigated the contribution of adult marrow-derived cells to the circulating EPCs and large vessel endo thelial cell (Hu et al. 2002). In this study, rhesus macaques were used, which had been transplanted with retrovirally-transduced, mobilized CD34+ peripheral blood. EPCs were mobilized with 10 g/kg G-CSF per day for 5 days, enriched with CD34+ selecti on, and grown in endothelial cell basal medium(EBM)-2 with FBS, VEGF-1, FGF2, EGF, IGF-1, and ascorbic acid. The number of vWF+ Ac-LDL+ CD31+ CD14endot helial colonies on Day 14 was used to assess the number of EPCs in circulation. Also, large vessel slices from transplanted macaques were grown in endothelial growth media 2 (EGM-2). Cells growing at Day 14


46 were sorted for CD31 and grown for another week. Neither population of cells showed any evidence of retroviral marking. The dete ction sensitivity was estimated at 1 in 105 vector-containing cells. In orde r to assess the negative result, they compared transduction efficiencies between hematopoietic cells and EPCs. They found that their culture conditions were not favorable for EPC tran sduction, neither was the addition of VEGF. EPC transduction was dramatically enhanced using EGM-2. Another factor contributing to the negative result, particularly in the la rge vessels, could have been that bone marrow derived EPCs are quite rare. Finding them is often like finding a needle in an arena-sized haystack. Without the ability to quickly sort through a large nu mber of cells to select out those that are clonally marked (via GFP or some phenotypic marker), a random sampling would very likely show negative results. Genetic clonal marking and detection t echniques have underg one a progression of changes over the past thirty years. Some we re born out of a new technology (recombinant retrovirus, PCR), some have emerged from a clever technical twist on existing techniques (LAM-PCR), and some are emerging based on changes in our understanding of the underlying physiology (bone ma rrow derived EPC culture). Thesis Statement The historical context surrounding th e controversial adult hemangioblast (Chapter 1) suggests that it is likely to be the origin of the adult endoth elial progenitor cell, a cell with abundant th erapeutic potential. However, it has been difficult to definitively show that hemangioblast activity is present in the adult due to poorly characterized phenotypic and genetic even ts surrounding its commitment decisions. Single-cell transplant is one technique that ha s been used by our lab to definitively show that the HSC can, under extreme selective pre ssure, act like a hemangioblast. However, it


47 was unknown whether the HSC is the only such bone marrow cell that has this capacity. Furthermore, it was unknown whether, under less stringent conditions, the same clonal pool that generates hematopoie tic cells also generates EPCs . I have created a modified version of the classic retrovira l tagging techniques (Chapter 2) that contains a unique identifier, which can be used as an alternat e to or in conjunction with the traditional technique (Chapter 4). I have used this technique to demons trate that the clonal pool of cells that generates blood also generates e ndothelial progenitor cel ls (Chapter 5). The details of the methods and the rationale fo r choosing those methods are presented next (Chapter 3). I end this thesis with a discussion of the resu lts and of future directions (Chapter 6).


48 CHAPTER 3 MATERIALS, METHODS, AND STATISTICS The methods that I have included in this section are, for the most part, the exact ones used for the research presented in this thesis. In a few cases, it was more appropriate to present a minor technical variation along with the pr esentation of results. These methods are clustered into four major object ives. First is the process of creation the retroviral tag library. Also included at the end of this section is an introductory description of fluorescence-ac tivated cell sorting (FACS), since it was used throughout this research. Second is the process of demonstr ating that this library can uniquely label a population of cells. Third is the process of generating a chimeric animal from tagged stem cells. Fourth is the clonal analysis of stem cell progeny from that chimeric animal. At the end of this chapter, I have also incl uded some information re garding the statistics of sampling, which is relevant to the analysis of a subset of labeled cells from a larger population. Creating the Retroviral Tag Library The process of generating a retrovirus capable of delivering genetic material to target cells involves a number of steps. Plasmids, small circ ular double-stranded pieces of DNA, are altered to carry a genetic package flanked by retroviral long terminal repeat (LTR) sequences. This plasmid, when transfecte d inside of cells, cr eates copies of the retroviral genome that are de fective, since they no longe r carry the genes required for proliferation, but are still be packaged into infectious viral particles. The process, however, requires specific tissue culture cells ca lled packaging cells that contain the extra


49 genes used for retroviral construction. Th e packaging cells, once transfected, release recombinant retrovirus into the surrounding medium. This medium supernatant is, in some cases, purified or concentrated. In or der to quantitatively assess viral production (titer), other tissue culture cells are exposed to various concentrations of the viral supernatant. An infected cell is determin ed based on an observable phenotypic change, such as the expression of a new enzyme or a fluorescent protein. Constructing the Recombinan t Retroviral Tagging Plasmid Since the retrovirus would be used to in fect stem cells, it was decided that the retroviral backbone of the ta gging construct would be the enhanced green fluorescent protein (eGFP) expressing Murine Stem Ce ll Virus (MSCV)-based retroviral vector MIGR1. MSCV-based vectors have been op timized for the introduction and expression of genetic material to pluripotent murine and human hematopoietic and embryonic stem cells (Pear et al. 1993). This optimization includes Porcine Cytomegalovirus (PCMV) LTRs, which have been shown to drive high-le vel constitutive expressi on in target stem cells. Warren Pear (University of Pennsyl vania, Philadelphia) generously provided MIGR1. The retroviral cassette of MIGR1 has the ability to simultaneously express two genes because of its bicistronic packaging construct. The second gene, in the case of MIGR1, was eGFP, abbreviated in this work as just GFP. Transcripti on of the bicistronic cassette was regulated by the promoter, enhanc er, and polyadenylation signals of the viral LTR. However, translation of the GFP was from an internal ribosome entry site (IRES) of encephalomyocarditis virus (EMCV) positioned just upstream of the GFP. MIGR1 has a retroviral packaging signal between the 5’ LTR and the first gene position. The full MIGR1 plasmid also contains the gene encoding -lactamase, an enzyme that can be used for bacterial cloning of the plasmid.


50 Multiple attempts were made to create a double-stranded DNA insert (for insertion into MIGR1) with enough randomness to eventu ally be used to uniquely label a large pool of cells. Initial attempts involved mutu ally primed synthesis (MPS), a process in which single-stranded primers can form a dime r with themselves, which can be used to prime the synthesis of a double stranded pr oduct. The reason for this was that the single-stranded primer contained multiple random nucleotides, which, at least theoretically, made it difficult to predict correct primer annealing in order to generate the double-stranded product needed for ligation in to MIGR1. Either MPS or the purification of the small product following the enzy matic reaction was problematic. Agarose electrophoresis, polyacrylamide electrophoresis, and size exclusion columns resulted in a product that failed to efficiently ligate into MIGR1. In addition to an MPS synthesized random nucleotide insert, the creation of a library of sheared (300-500 base pair fragments) salmon sperm DNA bluntly ligated into a cloning plasmid also proved not useful. The resulting librar y was not sufficiently random. Ultimately, a random nucleotide insert wa s created by annealing together the two single-stranded primers, RPFor (5’-p-TCGAGGGG-nnn-CACACACACA-nnnGGAAGGAAGG-nnnCCCG-3’) and RPRev (5’-p-AATTCGGG-nnn-CCTTCCTTCCnnn-GTGTGTGTG-nnn-CCCC-3’). The assumption that these random single-stranded primers would fail to anneal turned out to be wrong. Using an equimolar reaction of 233 ng RPFor and 238 ng RPRev were annealed in a volume 20 L, incubated at 68oC for 30 minutes followed by incubation at 4oC for 5 minutes. Since no enzymes were required due to the 5’-phosphorylated primers and restriction sites that were immediately available, no purification step was performed.


51 The first position of the expression cassett e of MIGR1 contains a multiple cloning site with sequences specific for the rest riction enzymes XhoI and EcoRI. The MIGR1 backbone was prepared to receive the insert by restriction enzyme digestion with these two enzymes using manufacturer specified me thods (New England Biolabs). Despite the fact that the two restriction sites were inco mpatible with plasmid self-reclosure, calf intestinal phosphatase (CIP) was used to rem ove the 5’ phosphate from the EcoRI end of the linearized plasmid and reduce the lik elihood of improper plasmid reclosure. An overnight 16oC ligation was performed using a 1:10 molar ratio between the MigR1 digest and the annealed RPFor/RP Rev primers in 15 Ls containing ligation buffer and T4 ligase (New England Biolabs) . The ligation product was transformed into chemically competent Escherichia coli DH5 (Invitrogen) using the manufacturer recommended 37oC heat shock method with 3 Ls of ligation product. Bacteria was grown in SOC medium, supplied with the bacteria, for 1 hour at 37oC prior to plating at a clonal density on plates of LB (Luria-Berta nia) agar containing 100 g/mL ampicillin (Sigma) in order to isolat e positive transformants. Initially, the plates were washed and grow n in LB broth containing ampicillin in order to harvest a large pool of transformant s containing the retrovir al expression plasmid with the random nucleotide insert. Unfortunate ly, this method generated an imbalance in the distribution of clones, which tended toward domination by a limited few based on clonal sequence analysis. Therefore, 200 coloni es were individually picked and grown up in 5 mLs of standard LB medium (1% trypt one, 0.5% yeast extract, 1% NaCl) containing 50 g/mL ampicillin. The Quantum Prep Plas mid Miniprep Kit (Bio-Rad) was used to lyse the bacteria and isolate plasmid DNA us ing a proprietary DNA bi nding silica matrix.


52 DNA from the separate clones was pooled in eq uimolar ratios. This library of retroviral clones was called MIGRN (Figure 3-1). Figure 3-1. Map of MIGRN. MIGR1, an eGFP expressing Murine Stem Cell Virus (MSCV) with Porcine Cytomegalovirus (PCMV) LTRs was used as the plasmid to construct the MIGRN retrovi ral tagging library. A short segment of DNA with 9 random nucleotides wa s ligated into the XhoI and EcoRI site. Retroviral Packaging Cells Several packaging cell lines were analyzed for their ability to produce transiently transfected high-titer vi rus capable of infecting hematopoietic stem cells. These included Phoenix-A, Phoenix-E, GP293/ VSV-G, and BOSC23, which cons truct retroviruses with a variety of tropisms. Amphotropic viruses (from Phoenix-A) infect both mouse and other species by binding to the rPit2 receptor, a sodium-dependent phosphate transporter. Ecotropic viruses (from Phoenix-E and BOSC 23) infect only mous e cells by binding to the mCAT-1 receptor, a cationic amino acid transporter. Polytropic viruses such as vesicular stomatitis virus G (VSV-G) pseudot yped viruses (from GP 293/VSV-G) infect a


53 very wide range of species by binding to phospholipids present on all cell types. VSV-G also has the advantage that it adds stability to the re trovirus, allowing it to be concentrated up to 2000-fold or more by ultracentrifugation. BOSC23 was used for the production of a ll batches of MIGRN because it showed a superior ability to produce virus with that plasmid. BOSC23, derived from Ad5-transformed human embryonic kidney 293 cel ls, was obtained as a gift from Warren Pear. BOSC23 cells were created by tran sfection with plasmids, pCRIPenvand pCRIPgag-2. The pCRIPenvplasmid was an envelope mutant that expressed only the retroviral gag and pol genes. The pCRIPgag-2 has mutations that caused expression only of ecotropic envelope gene. The packaging signals were re moved from these plasmids and the 3’ LTRs were replaced with a polya denylation signal to prevent recombination and formation of replication-competent viru s. These packaging plasmids were stably transfected in consecutive steps to also avoi d this problem. The stab le expression of gag, pol, and env eliminated the need for co-trans fection with helper pl asmids. The lack of replication-competence in BOSC23 packaged re troviruses had been previously verified. Although selection for the expression of retr oviral packaging proteins was possible using 1 mg/mL neomycin, 400 g/mL hygromycin, a nd 50 g/mL mycophenolic acid, it offers limited benefit for BOSC23 cells. The original screening was performed by selecting for high expression of reverse transcriptase a nd the envelope protein by RT-PCR, followed by a secondary selection for high titer inf ectious viral production. For this work, a BOSC23 clone was used no more than 2 pa ssages after verification that it produced high-titer (>106 IU/mL) virus.


54 Cell Culture The cell lines BOSC23 and NIH3T3 (embr yonic Swiss mouse fibroblast cell line, ATCC) were grown on 10 cm tissue culture treated plates in 12 mLs of Dulbecco’s Modified Eagle’s Medium (DMEM, Gi bco BRL) containing 10% FBS (Summit Technologies, lot FA1208), 100 U/mL of peni cillin/streptomycin (Gibco BRL), and 2 mM Glutamax (Gibco BRL), referred to in this work as DMEM complete. Culturing was performed at 37oC in a humidified 5% CO2 atmosphere. The cell lines, BOSC23 and NIH3T3, required passaging (1:8) every 3 to 4 days. Passaging of adherent cells required 5 to 10 minutes of trypsinization (0.05%, I nvitrogen), followed by neutralization with medium containing FBS. Centri fugation (300xg) was used to pe llet the cells, so that the medium could be changed. Aliquots of 107 cells were stored in 1 mL of 90% FBS and 10% dimethylsulfoxide (DMSO, Sigm a) in liquid nitrogen at -195oC. Packaging of Recombinant Retrovirus The GFP expressing plasmid MIGR1 a nd the plasmid library MIGRN were transfected into the BOSC23 cells using a CaPO4 transfection procedure. Although other transfection reagents were tried, includ ing Lipofectamine (Inv itrogen) and Fugene (Roche), this method proved to be both co st effective and produce better results. Transfections were performed on 10 cm tissue culture treated plates. BOSC23 plates that had reached approximately 60% confluency (7.7x106 cells per plate) had their medium reduced to 5.6 mLs to which 5.6 Ls of ch loroquine was added (25 uM) at least 5 minutes prior to transfection. HBS buffer wa s made as a 2x stock consisting of 50 mM HEPES, 10 mM KCl, 12 mM Dext rose, 280 mM NaCl, and 1.5 mM Na2HPO4 (Sigma). This solution was filter sterilized with a 50 ml conical tube with a 20 m filter (Millipore). It was critical that the pH of this solution be adjusted to as close to 7.05 as


55 possible in order to maximize transfection effi ciency. Furthermore, since frozen aliquots of HBS no longer performed optimally after one month of storage, solutions were made on a monthly basis. A CaCl2 stock solution was made with 2 M CaCl2 (Sigma). 40 g of DNA (unless otherwise specified) was added to sterile water, to a total volume of 1.4 mLs, containing 173.6 Ls of the CaCl2 stock. To this, 1.4 mLs of the stock HBS solution was added. Immediately after the HB S was added, a mechani cal pipette-aid was used to aerate the mix by injecting air for 60 seconds. This step formed a visible crystalline precipitate containing DNA. These 2.8 mLs were added to the tissue culture plate containing BOSC23 cells in 8 mLs of medium and chloroquine. Medium was replaced with 12 mLs of DMEM comple te after 8 hours of incubation at 37oC. Longer incubation times were harmful to the pack aging cells. After another 16 hours, the medium was changed using a volume of 7 mL s of DMEM complete per plate. After another 24 hours, the medium was removed a nd the packaging cells were destroyed. The medium was centrifuged at 300xg, isolated, and ce ntrifuged again in order to remove any contaminating packaging cells. The viral supernat ant was used either fresh or from frozen 1-milliliter aliquots stored at -80oC. The act of freezing supernatants reduced the titer by approximately 50%. Determination of Retroviral Titer The retroviral titer was determined by infecting NIH3T3 cells with viral supernatants. Three wells of 1x105 NIH3T3 cells per 35 mm well of a 6 well tissue culture treated plate were infected viral d ilutions. A useful single dilution was 35 Ls of virus diluted into 2 mLs of DMEM complete. Infectious media was supplemented with 4 g/mL of polybrene (Sigma). After 24 hours, the medium was changed to fresh DMEM complete. After another 24 hours, cells were trypsinized and quantitatively analyzed for


56 retroviral GFP expression by fluorescence ac tivated cell sorting (FACS), a technique described in the next section. This percenta ge was used to calcul ate the titer (IU/mL) based on the number of cells present in the well at the time of infection. Although titers were occasionally 105-106 IU/mL, only viral preparations with titers > 106 IU/mL were used for further experiments. Introduction to FACS Since FACS was used heavily in this rese arch, it was important to consider the details of this technique. Indi vidual cells can be labeled with a fluorescent color in a number of ways. In some cases, fluorescen t proteins can be synthesized endogenously from a gene that was introduced into a cell. This was the case with the GFP expressing retrovirus after it was allowed to infect the target cells. In other cases, antibodies that are attached to a fluorescent dye are allowed to bind to cell surface antigens (this technique was used for sorting experiments discussed late r in the methods section). Since antibodies are extremely specific, the cells expressing only that antigen are labeled. Simultaneous detection of multiple antigen and/or endoge nous fluorescent identifiers is possible because of the existence of multiple non-overlapping fluorescent dyes. The first step of the mechanized sort pr ocess mixes the fluorescently labeled cells with sheath fluid. The cells enter the detecti on apparatus and are pass ed in a single-file through a laser (light amplifica tion by stimulated emission of radiation) light beam. As it travels through this excitation laser, the emission from the fluorescent molecules on each and every single cell are detected. This informati on can be used to determine if that cell is positive or negative for the presence of a specific fluorescently labeled phenotypic characteristic. At the same time, the light scattered by each cell is measured. The light scattering is reported as a forward scatter and a side scatter. Forward scatter correlates


57 with the size of a cell. Side scatter correlates with the granul arity of a cell. The stream is then forced through a nozzle to make droplet s containing a single cell. The droplet is given a charge based on the various sort crit eria set up on an interfaced computer system. An electric field is used to separate the dr oplets based on this char ge and distribute the cells into distinct collection tubes. Because each droplet can be sorted in a matter of milliseconds, several million cells can be indi vidually sorted in one hour. For this research a FACS Vantage SE Turbosort a nd a FACScan (BD Biosciences) were used with their CellQuest software for sorting and analysis, respectively. Demonstrating Random Labeling of Cells with the MIGRN Library The pool of ecotropic retroviruses genera ted from MIGRN transfection of BOSC23 cells would have had a complexity of 200 different tags, assuming there was 100% efficiency at the library construction stage. The complexity of this pool needed to be experimentally verified. To accomplish this, a pool of NIH3T3 cells was retrovirally tagged with MIGRN and interrogated at a clonal level by sequencing. Subsequently, a pool of BM cells was tagged and interrogated at a clonal level usi ng a highly sensitive amplification, cloning, and sequencing technique. Retroviral Tagging of NIH3T3 Cells NIH3T3 cells were infected as previously described. Transduced cells were FACS sorted twice for retroviral GFP expression to produce a popul ation that was greater than 95% GFP+. Sorted cells were seeded into the wells of 96 well dishes at a density of 0.5 cells per well. Visual confirmation of the pr esence of a single a dherent GFP+ cell was made 4 hours after plating. Single cells were allowed to grow to a colony consisting of 1.5 to 5.0x104 cells. Cells were trypsinized, rinsed in PBS, and pelleted. DNA extraction and retroviral tag analysis of these ce lls is described in the next section.


58 In another experiment, retrovir ally transduced NIH3T3 cells were diluted in a series of 10-fold dilutions from 1x107 cells to single-cell con centrations. DNA extraction was performed as described in the next section, but tag analysis was pe rformed using the PCR pre-amplification procedure, which will be discussed later in this chapter. Isolation and Sequencing of NIH3T3 Retroviral Tags NIH3T3 pellets were resuspended in 100 L s of a lysis buffer consisting of 10 mM Tris-HCL (pH 8.5), 50 mM KCl, 0.01% ge latin, 0.45% NP-40, 0.45% Tween-20, and 100 g/mL DNase-free proteinase-K (Sigma ). The incubation was performed at 55oC for 3 hours. The proteinase-K was inactivated with a subsequent 10-minute incubation at 95oC. An aliquot of this genomic DNA ( 200 to 800 ng) was used in a sequencing reaction, which was performed as direct ed by the Big Dye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems). Labe led DNA was analyzed on an ABI Prism 377XL or 373XP Automated DNA Sequencer. Bone Marrow Harvesting Bone marrow was obtained from 2 to 4 month-old male C57BL/6 mice (Jackson Laboratories). These mice were housed in our specific pathogen free (SPF) animal facility and were handled at all times in a manner approved by the University of Florida Institutional Animal Care a nd Use Committee. Mice were anes thetized using isoflurane (Aerrane, Baxter) and sacrificed by cervical dislocation. Femu rs, tibias, and humerii were isolated from the mice using aseptic surgic al technique. Muscle was removed from the isolated bones. The joint surfaces were sli ghtly trimmed off to reveal the bone marrow cavity. Marrow cavities were evacuated by inse rting a 27-gauge needle into one end of the hollow bone and flushing with 1-2 milli liters of Dulbecco’s Phosphate-Buffered Solution (PBS, Gibco BRL). Bone marrow was triturated into a single-cell suspension


59 using a 25-gauge needle. Cells not adhering to TC-treated plates after 2 hours were collected and rinsed in DMEM or PBS. This step removed adherent stromal cells that may have been dislodged. The t ypical yield was approximately 1x108 cells per animal. Bone Marrow Culture and Retroviral Tagging Multiple experiments were conducted to obtain the optimal conditions for retroviral infection of bone marrow cells. The use of co -culturing packaging ce lls and target cells was examined using a Transwell dish (C orning). Also, direct exposure to viral supernatants in standard cu lture conditions was examined. And lastly, spinoculation, a technique of centrifuging the cells and virus together to increase contact time, was examined using between one and three rounds . Two rounds of spinoculation separated by 24 hours in 4 g/mL of polybrene were dete rmined to be optimal. It also became apparent that the more the stem cells cycl ed, the better the retrovi ral infection. Clearly this makes sense when considering the retrovi ral life cycle. Combinations of cytokines, which promoted the survival of the HSC, were examined, including various concentrations of interleukin-3 (IL-3), inte rleukin-6 (IL-6), and st em cell factor (SCF, steel factor, mast cell growth factor, c-Kit ligand). It was de termined that 20 ng/mL IL-3, 20 ng/mL IL-6, and 50 ng/mL SCF provided the necessary signals for HSC survival, as determined by phenotypic charact erization and transplantatio n studies. Furthermore, it was determined that a 5-FU pretreatment us ing an intravenous (retro-orbital) 150 mg/kg dose administered 5-days prior to bone marro w harvest offered an important advantage in generating transplanted animals with highe r levels of engraftment with the GFP+ retrovirally infected cells. This drug is a chemotherapeutic agent used to deplete progenitors and promote hema topoietic stem cell cycling.


60 Therefore, bone marrow cells used fo r retroviral tagging were obtained from animals that had been pretreated 5-FU. Im mediately after bone marrow harvest, cells were plated at a density of 2x106 cells/mL and cultured in Iscove’s Modified Dulbecco’s Media (IMDM, Gibco BRL), supplemented with 10% FBS, 100 U/mL penicillin/streptomycin, 2mM Glutamax, 20 ng/ mL murine IL-3, 20 ng/mL murine IL-6, and 50 ng/mL murine SCF (R&D System). This media was referred to as IMDM complete. After 48 hours of cytokine stimul ation, bone marrow cells were spinoculated (770xg for 60 minutes at room temperature) in medium containing virus and 4 g/mL polybrene. Spinoculation was performed at a multiplicity of inf ection (MOI) of 0.8-1.0 and a cell density of 1x107 cells/mL diluted in IMDM, where necessary. Fresh IMDM complete was added two hours after spinocul ation. A second round of spinoculation was performed 24 hours later. Nonadherent cells we re used for transpla nt 24 later (96 hours post harvest) or for analysis after 48 hours later (120 hours post harvest). Isolation and Sequencing of Retroviral Tags from Cultured Bone Marrow Cells The isolation of genomic DNA from retrovirally transduc ed bone marrow cells was performed in a manner identical to that us ed for NIH3T3 cells. Briefly, lysis buffer containing proteinase-K was used to lyse and digest the cells in 100 Ls. At this point, the procedures diverged. In th e case of the clonally pure colonies of NIH3T3 cells, DNA could be directly sequenced. In the case of a clonally mixed population of tagged bone marrow cells, a strategy of isolating the signa ls from single clones needed to be used. This was accomplished by polymerase chain reaction (PCR) amplification of the tag, agarose purification of the PCR product, li gation of the PCR product into a cloning vector, cloning in TOP10 E. coli , plasmid isolation of multiple clones, and sequencing of those multiple clones.


61 Although most scientists thoroughly unders tand the process of PCR, I have included a description here for non-scientis ts. PCR amplification is a strategy of amplifying a specific region of DNA using a DNA polymerase. A common polymerase is Taq polymerase, which was originally isolat ed in 1976 from the thermophilic bacteria Thermus aquaticus. Taq and other thermo stable DNA polymerases (some with a substantially improved proofread ing ability) facilitated the automation of PCR. Although the first PCR reaction was performed in 1985, the technique has only been widely used for the past 13 years. Currently the pro cess involves an automated cycler, which incubates the reaction at various temperatures for a specified amount of time for between 30 to 40 cycles. The reaction contains th e original DNA template, nucleotides, DNA polymerase, a buffer solution, and primers. The primers are short DNA sequences required for DNA polymerase initiation or pr iming. DNA is first denatured at a high temperature, such as 94oC. Primers are then annealed at a temperature that promotes specific primer-to-template binding, such as 54oC. Lastly, extension of the primer segments to create double-stranded DNA occu rs at a temperature optimal for the DNA polymerase, such as 72oC. Each round of PCR results in a 2-fold amplification of the specific region of DNA that is defined by th e primers. Therefore after 30 rounds, with 100% efficiency, the amplification would be 2^30, or 1x109-fold. Retroviral tags were usually not plentiful enough to dir ectly clone; therefore PCR amplification was used to pre-amplify the re troviral tags from genomic DNA. It is important to recognize that PCR, although exponentially amplifying the gene product, maintains the ratios of gene products present in original mixed populations. The region of genomic DNA containing the retroviral tags was amplified using 35-40 cycles of PCR. Cycle conditions included an initial 94oC for 90 s, followed by cycles of 94oC for 30 s,


62 63oC for 60 s, and 72oC for 60 s, followed by a final extension of 4 minutes at 72oC. The 10x PCR buffer contained 100 mM Tris-H Cl pH 8.3, 500 mM KCl, 15 mM MgCl2, 0.01% gelatin. PCR reactions also include d 100 M dNTPs, 1 U Taq, and 200 nM of each primer. The primers, 5’-TTG AAC CTC CTC GTT CG A CCC C-3’ and 5’-CAC CGG CCT TAT TCC AAG CG-3’, anneal sp ecifically to MIGRN DNA just up and downstream of the random nucleotid e insertion site, respectively. The PCR reaction was electrophoresed on an agarose gel so that the amplification product could be verified, purified, and isolate d. In general, agarose gels were 1.2% with 0.5 g/mL ethidium bromide, the load dye wa s a 10x stock solution of 40% sucrose, and the ladder was 1kb+ (Invitrogen). Appropr iately sized bands (a pproximately 200 base pairs) were excised with a scalpel or razor blade, taking precautions to not cross-contaminate the bands. Agarose was disrupted using an 18-gauge needle and resuspended in water. A standard phenol extraction with ethanol precipitation was performed to recover the PCR product. Purified PCR products were ligated into the pCR4-TOPO plasmid (Invitrogen) for cloning. Ligations were performed using the recommended 10 Ls in the tube containing the dehydrated PCR product. The TOPO-TA fo r cloning kit (Invitrogen) is a kit that allows for rapid ligation at room te mperature due to a covalently bound DNA topoisomerase I, which act as both a restrict ion enzyme and a ligase. The ligation product was transformed into chemically competent TOP10 E. coli and plated at clonal densities on LB-Agar plates with 100 g /mL ampicillin. The pCR-TOPO contains the ccdB, which is toxic to E. coli. Positive ligation results in disruption of ccdB, and survival of the E.coli. Positive transformants were grow n in 5 mLs of LB broth. Plasmid DNA was isolated using the Quantum Prep Plasmid Mi niprep Kit. Plasmid DNA was screened for


63 the correct insert size by digesting 1 g of plasmid DNA with XhoI and EcoRI restriction enzymes and visualization on an agarose gel with ethidium bromide. The sequencing reaction was primed with T7 or T3 using 200 to 800 ng of DNA. At least 10 and up to 50 colonies, when available, were sequenced for each population of analyzed cells. Generating Chimeric Retrovirally Tagged Mice Female C57BL/6 mice, 2 to 4 months old, housed in our SPF animal facility, were used as recipients in transplantation experime nts. The purpose for using male into female transplants was two-fold. First, female mice are easier to group hous e for long periods of time. Second, the Y chromosome could, if need ed, be used as a s econdary marker of transplanted cells. Recipient mice were expos ed to 950 rads (cGy), a lethal dose, of irradiation by exposure to a Cs137 source in a Gamma Cell 40 irradiator. Mice were isolated in a PVC (polyvinyl ch loride) container during irradi ation. Four hours later, mice were anesthetized with isoflu rane. Bone marrow cells, either freshly tagged in culture or isolated and enriched from mice that were engr afted with tagged stem cells, were injected into the retro-orbital sinus, a capillary bed just behind the eye, in a volume of 100 Ls. Primary transplant re cipients received 5x105 or 2x106 cultured bone marrow cells. Following transplant, mice were housed in th e SPF animal facility and given food and non-medicated water ad libitum. Secondary Transplantation of Retrovirally Tagged HSCs Bone marrow cells were harvested fro m the primary transplants in manner identical to the harvesting of bone marrow from unmanipulated C57BL/6 mice, except that no 5-FU was used prior to harvest. B one marrow cells were stained in a volume of 0.5 mL of PBS containing 1% bovine serum albumin (BSA) and 0.1% sodium azide, abbreviated PBA. Sodium azide was used to prevent endocytosis of the antibody bound


64 surface protein. Fluorescently la beled rat anti-mouse antibodies Sca-1-PE and c-Kit-APC or Scal-1-biotin (labeled with a seco ndary goat anti-rat streptaviden-APC/Cy7), c-Kit-APC, and lineage-PE were used to stain the bone marrow (Pharmingen). Staining was performed in 500 L at 4oC for 20 minutes using antibody titers determined to be optimal for the specific lot (appr oximately 1 g for between 106-107 cells). When lineage markers were used in this manner, th ey included CD11b, CD3, CD4, CD8, B220, Gr-1, and Ter119. Various combinations were unsuc cessfully attempted, including a dose of 2500 Sca-1+ c-Kit+ GFP+ cells. Secondary tran splants were successfully generated using the large dose of 25000 Sca-1+ c-Kit+ GFP+ cells obtained from the bone marrow of primary transplanted mice that showed si gnificant engraftment (>20%) 3 months after transplantation. Screening of the Transplant Recipients Peripheral blood was collected from a sma ll incision made in the tail vein just above the anus. Blood was collected in 1 mL of PBS containing 10 mM EDTA (Sigma) to prevent coagulation. Mononuclear cells were enriched from whole blood using Ficol-Hypaque (Amersham) density gradient separation, as specified by the manufacturer. Cells were rins ed in PBS and resuspended af ter centrifugation in 0.5 mL of PBA. Cells were fluorescently labeled using rat anti-mouse CD45-APC, and CD11b, B220, CD4/8, or Flk1-PE (Pharmingen) at dilu tions determined to be optimal for the specific lot (approximately 1g for between 106 and 107 cells). Engraftment was considered positive for animals with great er than 10% GFP+ contribution to the lymphocyte gated (lower complexity/side scat ter, variable size/f orward scatter) CD45+ blood compartment by FACS analysis using 20,000 events or more. The positive control for GFP was peripheral blood from the Nagy GF P transgenic mouse (Jackson Labs). The


65 negative control was peripheral blood from an unmanipulated C57BL/6 mouse. IgG controls for APC and PE were also used to verify a lack of nonspecific binding, which is particularly important when staining for very rare populations such as Flk1. Clonal Analysis of Stem Cell Progeny of Mice Transplanted with Tagged Marrow Both blood and bone marrow needed to be collected from primary and secondary transplanted animals at the time of sacrifice, which generally occurr ed at 3 months after transplantation. In some cases, mice were kept to longer time points to verify that the LT-HSC could be retrovirally transduced usi ng the techniques described in this thesis. Since mice were being sacrificed, a large sa mple of blood could be obtained by either bleeding the retro-orbital sinus or by a cardiac puncture. In either case, animals were given a lethal dose of Avertin (2-2-2 Tribromo ethanol, Aldrich) or is oflurane prior to the bleed. Blood was collected in 1.5 mLs of PBS containing 10 mM EDTA. Mononuclear cells were isolated as previously descri bed using density grad ient separation. Bone marrow was harvested from all the marrowcontaining long bones, as previously described. Bone marrow and blood samples were fluorescently labele d in 0.5 mL of PBA with rat anti-mouse Flk1-PE, CD45-APC, a nd Sca-1-biotin (secondary goat anti-rat streptavidin-APC/Cy7). From the blood and bone marrow, GFP+ Flk1+ Sca-1+ (endothelial progenitor populat ion) and GFP+ CD45+ Flk1(hematopoietic population) cells were separately collected by FACS. In the case of bone marrow sorting, if a sufficient (>500,000 cells) was obtained in the GFP+ CD45+ Flk1population, the population was split into three and re-label ed with rat anti-mouse CD11b-PE, B220-PE, and CD3-PE for another round of sorting in order to enrich for monocytes, B-cells, and T-cells, respectively. Th e final sorted populations were placed in separate 1.5-mL eppendorf tubes and centrifuged. DNA was isolated using the previously described lysis


66 buffer containing proteinase-K. PCR was used to amplify the retroviral tags from the genomic DNA. Gel electrophoresis was used to purify the PCR product, which was cloned into the pCR4-TOPO cloning plasmid. Between 10 and 50 separate positively transformed TOP10 bacterial clones were isol ated were isolated from each ligation. Each bacterial clone contained a tag from one single cell in the original so rted populations. The plasmids from those clones were isolated and sequenced in order to identify the retroviral tags. These sequences determined the clona l stem cell origins of the endothelial progenitor and hematopoietic cells. Statistical Analysis In situations where a population is de termined from only two entities, the distribution is considered binomial. For ex ample, in Smith and colleagues work to identify the frequency of LT-HSCs in enri ched bone marrow populations, it was feasible to use the binomial equation to calculate limited dilution probab ilities (Smith et al. 1991). Marked Sca-1+ LinThy-1lo cells had two possibilities: to contribute or not to an observable clone. This two state system is quite manageable, mathematically. Here n1+n2=N and p1+p2=1.0, where N represents the tota l number of observations, which are composed of the two different outcomes (n1 and n2) each with their respective probability distribution with in the overall population (p1 and p2). P(n1, n2; p1, p2) = N!(p1 n1p2 n2)/(n1!n2!) In situations where a popul ation is determined from more than two entities and sampled, the distribution is multiple hypergeometric. However, with the assumption that the population is large with re spect to the number of sampled items, the distribution can


67 be simplified to multinomial. A example of this, in the biologic system encountered in everyday genetics, is the screen ing bacterial colonies (Singer et al. 2000). In the case of the multinomial distributions, the ma thematical expression becomes: P(n1, , nj; p1, , pj) = N! (p1 n1p2 n2pj nj)/(n1!n2!..nj!) This expression allows you to calculate the pr obability that in N trials or samples, you would see some specific multinomial distribution {n1, , nj} when you have j items to choose from with each of those j items having probabilities {p1, , pj}. This can be further simplified with the assumption that th ese j items are evenly distributed in the population, thus reducing {p1, , pj} to {1/j, , 1/j}. In order to determine what the probability that some actual number items is pr esent in this population given that some number of items is observed in a sample d group, the individual probabilities for each combination of {n1, , nj} meeting the observe d criteria, need to be determined and added. For example, the probability that 3 dis tinct items (j=3), each with a frequency of 1/3 (pi=1/3) are present in a large population when 4 items are sampled (N=4) and 2 distinct items are found is e xpressed as follows (the func tion P is described above): P=P(2,2,0)+P(2,0,2)+P(0,2,2)+P(3,0,1)+P( 3,1,0)+P(1,3,0)+P(1,0,3)+P(0,1,3)+P(0,3,1) This equation becomes progressively more complex as the size of the numbers involved get larger. It is ther efore helpful to have a comput er perform this calculation. For this thesis, a computer program was written in C++ to make these multinomial probability calculations based on the estimated number of transplanted HSCs in order to


68 determine the number of distin ct HSC / EPC clones that must be sequenced or based on using the observed clonal number and the total number sampled to determine the complexity of the tagging population.


69 CHAPTER 4 USING RANDOM RETROVIRAL LI BRARY AS CELL LABELS The results are organized into two chapters primarily to maintain the distinction between the results from testing a new system for tracking clonality and the results from the use of this system to demonstrate in-v ivo hemangioblast activity. In this chapter, I will discuss the general results from early re troviral studies. These results were used to select the optimal retroviral manufacturing cond itions for construction of a vector capable of stem cell transduction. In the second part of this chapter, results are presented which relate to the construction of the retrovir al tagging library and assessment of the complexity of this library. Packaging High-Titer Retrovirus The specifics of the retrovirus, the pack aging, and infection conditions turned out to be critical to transduction of hematopoi etic stem cells. Stable packaging cell lines contain the gag, pol, and env viral genes, which are needed for the production of recombinant retrovirus. Since each of these gene s can be variably expressed or lost over time, the need for a selection or screening of high-titer clones is necessary. In the preliminary work for this thesis, a wide va riation in retroviral production capacities was observed between types of packaging cells , specific packagi ng cell clones, and expression plasmid transfection techniques. After examining retrovirus produced by GP293 (Figure 4-1), Phoenix-A (Figure 41), Phoenix-E (Figure 4-2), and BOSC23 (Figure 4-3) using Lipofec tamine, Fugene, and CaPO4 transfection methods with or without VSV-G pseudotyping, a choice was made to use a specific BOSC23 clone with


70 the CaPO4 transfection method based on the fact that it produced viruses with the highest consistent hematopoietic cell infectiv ity (>30% when further optimized). Figure 4-1. Phoenix-A and GP293 packag ed retrovirus. Early EGFP retroviral optimization studies used a variety of packaging cells and transfection conditions. None of these options proved useful for manufacturing retrovirus capable of robust HSC inf ection. A) Fugene with GP293 cells. B) Lipofectamine with GP293 cells. C) Fugene with Phoenix-A cells, indirectly showing poor titers. D) Lipofectamine with Phoenix-A cells, indirectly showing poor titers. E) Fugene with Phoenix-A cells cotransfected with VSV-G. VSV-G cotransfection caused a decreas in titer. F) Lipofectamine with Phoenix-A cells cotransfected with VSV-G. Between Lipofectamine, Fugene, and CaPO4, CaPO4 produced slightly higher titer virus, although none of the thre e were statistically distingu ishable. This is surprising considering the advertised claims and the difference in price between the commercial


71 products and the CaPO4. Although VSV-G pseudotyping offers the advantages of more stability and an ability to concentrate, co -transfection with a VSV-G envelope plasmid caused a significant drop in viral titers a nd transduction of hematopoietic cells with undiluted virus (Figure 4-2). Figure 4-2. Phoenix-E packaged retrovirus. Phoenix-E packaging cells were also tested. VSV-G pseudotyping dramatically decrea sed titer. Viruses packaged with Phoenix-E were the first packagi ng cells capable of transducing hematopoietic cells, but only with >5% efficiency. A) Lipofectamine with cotransfection of VSV-G. B) Fugene with cotransfection of VSV-G. C) Lipofectamine without cotransfection. D) Fugene without cotransfection. Once the selection of the packaging ce ll line and transfection method was made, based on titers (using NIH3T3 cells) and tran sduction efficiency (h ematopoietic cells), there was still variability in the production of hi gh titer virus. In an effort to increase the consistency of producing high-titer prepara tions, other transfection conditions were optimized, including the amount of DNA used fo r transfection (Figure 4-5A) and the time of retroviral harvest (F igure 4-5B). In these figures, perc ent GFP was quite often used to assess the level of infection. Si nce this is only a relative m easure, viral titer (IU/mL) was calculated by multiplying the cell population size at the time of viral transduction by the percent GFP detected by FACS after 2 days of retroviral expression. GFP percentages


72 used in this calculation were obt ained from viral dilutions that lie on the linear part of the dilution curve (Figure 4-5C). Figure 4-3. Ecotropic BOSC23 packaged re trovirus. BOSC23 packaging cells were transfected using a standard CaPO4 transfection protocol and produced retroviral titers > 1x106 IU/mL which were capable of consistently transducing hematopoietic ce lls with >30% efficiency. Figure 4-4. Retroviral tran sfection and harvest optimiza tion. Further optimization experiments were performed usi ng the eGFP expressing retroviral expression plasmid, MIGR1, from BOSC 23 packaging cells. A) The amount of the retroviral expression plasmi d added to a 10 cm plate with 7.7x106 packaging cells was determined by us ing a range of between 0 and 80 g. Transfection amounts of 40 and 80 ug produced the highest titer virus, although since the difference between the two was negligible, 40 g was chosen as optimal. B) The amount of virus produced by transfected cells was found to dramatically decrease over ti me. C) A series of viral dilutions were used to accurately asse ss viral titer on NIH3T3 cells. Constructing the Recombinan t Retroviral Tagging Plasmid Several variations on the choice of the random insert eventually led to the use of a double stranded insert of 37 base pairs, 9 of which were random and 28 of which had


73 homology with the opposite strand. A 65oC incubation step was sufficient to add enough energy to the system to promote single strande d annealing to a homologous (or nearly so) pair (Figure 4-1). Due to the potential mi smatches between random bases in a double stranded product, anchor regions proved to be critical to promote dsDNA stability (based on difficulties with an earlier design without those anchors) . Furthermore, the primers, designed to minimize secondary structure, did not show evidence of st ructures interfering with proper annealing. These features theref ore proved to be adequate to generate a random nucleotide linker-like insert capable of subsequent ligation. This insert was ligated into MIGR1, as described, and pr oduced several hundred colonies after transformation. 200 of these colonies we re used to produce the MIGRN library. Figure 4-5. Random insert used to construct retroviral libra ry. The random insert was created by annealing two primers c ontaining 9 random nucleotides and 28 bp of homology. This 3% agarose gel was used to separate the primers (loaded in lanes 2 and 3 for size comparison) from the double stranded product (loaded in lanes 4 and 5). Lane 1 shows a 100 bp band. Specificity and Sensitivity of the Retroviral Tag Technique The MIGRN library of retroviral tags wa s used to infect a population of cells. Theoretically, that population of cells would each receive a unique tag, which could then be detected using sequencing. Pr ior to determining the complexi ty of the pool of tags, the sensitivity and specificity of the techniques used to detect the tag were assessed. The


74 specificity was assessed using the PCR amp lification process, which was used to pre-amplify tags prior to cloning. In ge neral, if a band was visible on agarose electrophoresis, it could be isolated a nd successfully cloned. When a band was not visible, the cloning was sufficiently ineffici ent, that no transformants were observed (greater negative controls). Thus, specificit y was assessed based on the absence of a band in the negative control PCR reaction, which was loaded with only ly sis buffer instead of lysis buffer containing genomic DNA (Figure 4-6) . When using 40 cycles or less of PCR, the negative control was fals ely positive only once out of 15 experiments. Out of 4 experiments using 45 cycles, 2 had negative controls that were faintly positive. Figure 4-6. Detection of the re troviral tag. In or der to detect the re trovirus and the tag that it carries, the regi on of genomic DNA containing the tag was amplified using retroviral specific primers. Lane 1 is the 1kb+ ladder. Lane 2 is the amplification of the empty retroviral expression plasmid MIGR1, a positive control. Lanes 3-7 are amplifications from 2x104 tagged NIH3T3 clones. Lane 8 is amplification of only lysis bu ffer, a negative control. 35 cycles of PCR were used. Controls reactions, such as these, were used to assess the specificity of this technique. Next, the sensitivity of this technique was assessed using dilutions (in lysis buffer) of tagged NIH3T3 cells. Although 10 cells we re sufficient to elicit a band at 40-45 cycles, the judgment was made that no more than 40 cycles would be used based on a lack of specificity at higher detection levels (Figure 4-7) . The band present at 10 cells using 40 cycles was not sufficient to produce colonies for the ligation and cloning with the correct insert. A band from 100 cells wa s, however, able to produce 20-25 colonies when similarly processed. Therefore, for this study, the minimum sensitivity was set to


75 DNA from 100 positive cells. As a comparis on, chemiluminescent detection of singlecopy genes using the Southern blot tec hnique requires DNA from approximately 1x104 to 1x105 cells. Figure 4-7. Sensitivity of retroviral tag detec tion. The sensitivity of the PCR detection was examined by generating dilutions of retrovirally tagged NIH3T3 cells from 1x106 cells to a single cell. Genomi c DNA was isolated and retroviral tags were amplified with between 25 and 45 cycles. No advantage was seen with using the more sensitive, but less specific 45 cycles. Complexity of MIGRN Labeling In order to assess whether or not this ne w clonal tracking method would be able to uniquely label a large pool of cells and that those labels could be subsequently detected, NIH3T3 cells were infected with the libr ary of MIGRN retrovira l tags. Two rounds of FACS were used to enrich (>95%) for the retrovirally tagged GFP+ NIH3T3 cells two days after infection. 480 cells we re seeded into 96-well dishes at a ratio of 0.5 cells per well. After 4 hours, 69 wells were identified as containing a single vi able adherent GFP+ cell. After 2 weeks, these colonies had grown to 1.5-5.0x104 cells. DNA (200-800 ng) was isolated and sequenced as discussed in the methods using 30 cycl es of PCR. Out of 30 sequenced clones, 29 distinct clones were observed (Figure 4-8) . Using statistical


76 modeling based on a multinomial distribution, th is is consistent with greater than 150 distinct tags. Figure 4-8. Clonal tags of NI H3T3 cells demonstrate complex ity. A pool of retrovirally tagged NIH3T3 cells was plated at si ngle-cell density. Single clones were verified visually. Genomic DNA was isol ated from colonies of between 1.5 to 5x104 cells. Retroviral tags were amp lified, cloned, and sequenced. Each clone represents one cell in the original mixed population. There was one repeat C13 and C19 out of the 30 clones that were sequenced. Statistically, this was consistent with a pool of greater than 150 retroviral tags. Verification that this level of complex ity is maintained when switching from NIH3T3 cells to HSCs was achieved by inf ecting a pool of cultured bone marrow cells. 1.0x105 bone marrow cells, which had undergone in-vitro retrovira l tagging with MIGRN, were sorted for the Sca-1+ c-K it+ linHSC phenotype a nd were interrogated using the pre-amplified clonal sequencing a pproach. 30 of 36 tags were identified as unique (Figure 4-9) with no repeats greater than 2. Again, using statistical modeling based on a multinomial distribution, this is consiste nt with greater than 150 distinct tags.


77 Figure 4-9. Enriched bone marrow tags de monstrate complexity. Whole bone marrow cells were harvested from 5-FU pretr eated C57BL/6 mice, exposed to IL-3, IL-6, and SCF, and tagged with the MI GRN pool of retroviruses. After 5 days of total culture, FACS was used to enrich for 1x104 GFP+ Sca-1+ c-Kit+ Lincells. Genomic DNA was isolated, retroviral tags were amplified, cloned, and sequenced. Each cl one represents a single cell. There were 3 repeats found out of 36 clones analyzed. This is also statistically consistent with a pool of great er than 150 retroviral tags.


78 CHAPTER 5 USING TAGGED BONE MARROW TO ID ENTIFY THE CLONAL ORIGIN OF ENDOHTHELIAL PROGENITOR CELLS In this chapter, I present the results from using the retrovi ral tag library to individually label mouse HSCs, which were transplanted in a primary, unenriched, manner and in a secondary, enriched, manner into myeloablated recipi ents. The first part of this chapter describes the steps needed to confirm that stem cells were, in fact, retrovirally tagged. Also it was important to confirm that retrovir al GFP expression is maintained since silencing has been problema tic in several GFP expr essing retroviruses (Klug et al. 2000; Swindle et al. 2004). The GFP signal was used to re-isolate cells derived from retrovirally tagged stem cells. Al so, it was essential to show that the clonal tags could be identified in the re-isolated ce lls. The second part of this chapter presents the results from a clonal analys is of hematopoietic and endot helial progenitor cells in the primary and secondary transplants. This anal ysis was performed in order to assess the clonal origins of these two populations. Cofirmation That Stem Cells Were Transduced Several studies have report ed different requirements for hematopoietic stem cell maintenance in short-term in-vitro culture s (Sitnicka et al. 1996; Emery et al. 2000; Nakauchi et al. 2001; Zhang a nd Lodish 2005). Since there was need, in this portion of the project, to maintain the HSC in cultu re to allow for retroviral transduction, preliminary nonquantitative studies were pe rformed to compare combinations of cytokines (SCF, IL-3, IL-6, GM-CSF, TPO, FLT3 L, and IL-11) at various concentrations (5 to 100 ng/mL) for their ab ility to stimulate prolifera tion (morphologically consistent


79 with progenitors and stem cells) without significant different iation (heterogeneous morphologies). These studies resulted in a confirmation that commonly used triple cocktail of SCF (50 ng/mL), IL-6 (20 ng/mL ), and IL-3 (20 ng/mL) added to FBS containing growth media was worth pursuing. Therefore, bone marrow from three mice in three separate experiments was used to assess the phenotypic changes induced by culturing on the hematopoietic stem cell (Sca-1+ c-Kit+) a nd the endothelial progenitor cell (Flk1+) populations. After 96 hours of culture, both groups demonstrated proliferation. Data is presented as normalized by the input cell number. Figure 5-1. Phenotype of cultured bone ma rrow cells. The phenotype of cultured bone marrow cells was examined in order to assess the HSC & EPC maintenance / expansion capacity of th e culturing conditions. Flk1+ is a marker for EPCs. The Flk1+ population expanded 4-fold. Mo st of the Flk1+ cells present in culture after 4 days were Flk1+ CD 45+. This population expanded 3-fold. The Sca-1+ c-Kit+ group, which contai ns the HSC also increased 3-fold. The Sca-1+ c-Kit+ number of cells increased approximately 3-fold from a mean of 3.97 to 11.59% (p<0.05). This corresponded to 28.55% of the final population. The Flk1+ number of cells increased approximately 4fold from a mean of 2.90 to 11.47% (p<0.05). This corresponded to 26.09% of the final population. Surprisingly, most of those Flk1+ cells were also CD45+. The Flk1+/CD45+ fraction increased 4-fold from 2.50% to 10.30% (p<0.05). This corresponded to 23.39% of the final population (Figure 5-1). The Flk1+/CD45fraction increased 3-fold from 0.41 to 1.17% (p<0.05). This corresponded


80 to 2.7% of the final population. Although the prevailing opinion is that EPCs should be CD45-, this data suggests there may be an interesting story behind these CD45+ / Flk1+ cells. In order to further confirm that the in vitro conditions and the retrovir al tagging did not diminish the in-vivo engraftment and stem cell properties of the LT-HSC, 25 mice were transplanted following lethal irradiation with 2x106 retrovirally tagged bone marrow cells. Out of those 25, 24 survived to one month and had a mean engraftment of 27%. Engraftment was assessed as CD45+ GFP+ c ontribution to the mononuclear cells from tail vein bleeds. Since transgenic GFP animal s have approximately 30 % as this detection value, 27% is sometimes reported as 90% adjust ed engraftment. This is confusing, so my numbers will represent observed FACS pe rcentages and not adjusted engraftment percentages (Figure 5-2). Trilineage engr aftment was verified using CD11b, B220, and CD4/8 after two months (Figure 5-3). Figure 5-2. Long-term engraftment of tagge d hematopoietic stem cells. Bone marrow cells that had been retrovirally infected in-vitro and transplanted into lethally irradiated mice contributed to hematopoiesis past 12 months. These FACS plots demonstrate the GFP+ ce lls in the lymphocyte gated, CD45+ fraction of peripheral blood 13 months following transplant. Only the HSC is known to produce blood for that le ngth of time, sugge sting retroviral infection of the HSC.


81 Figure 5-3. Multilineage engraftment of tagge d hematopoietic stem cells. Retrovirally infected bone marrow that had been tran splanted into lethally irradiated mice participated in multilineage hematopoiesis, including CD11b+ monocytes, B220+ B-cells, and CD4/8+ T-cells 3 months after transplant. Of these 24 surviving mice, 1 had a solid tumor growing on its leg and was sacrificed (the tumor was cultured and fr ozen sections made for later analysis, Figure 5-4), 10 were used for experiments not related to this project (retinal neovascularization studies) and 13 were kept to the 6-month time point. These 6-month time point animals showed 32% engraftment. After this time point, 1 died at 8 months from an unknown hepatic carcinoma (carca ss was unavailable), 10 were used an unrelated project (retinal ne ovascularization), and 2 were kept past 12 months. Figure 5-4. Tumor in MIGRN transplanted mouse. A) A GFP+ tumor was identified growing near the femur of this mouse that had been transplanted with MIGRN tagged bone marrow cells after 5 months. B) Tumor tissue was cryogenetically preserved. Although adherent cells grew for a brief period of time in culture, they eventually died before analysis. These two mice showed 30% engraftment (Figure 5-2). None of the mice that were bled over the time course demonstrat ed GFP inactivation once they had passed the


82 2 month time point with positive engraftmen t. Since only HSCs are known to produce blood for this length of time, these results demonstrate that the HSC was in fact retrovirally transduced and tagged. The defini tion of a stem cell is one that robustly contributes to multilineage progeny (F igure 5-3) and is self-renewing. Primary Transplants In order to assess the clonal origins of EPCs with resp ect to hematopoietic cells, without imparting a phenotypic bias, whole bone marrow cells were used for transplant. A limited number of cells, however, were used in order to remain well within the complexity limits of the available clonal tracking system. Since the complexity of the MIGRN pool was 200 and the chance of uniquely labeling n HSCs is 200!/((200-n)!*200^n), the probabil ity of uniquely labeling 5 HS Cs is greater than 95%. Five tagged HSCs were transplanted for e xperiments assessing clonal overlap between EPCs and HSCs (based on the assumption of 1 HSC per 105 cells mentioned in Chapter 2). Tagged bone marrow was transplanted into 55 mice, with each mouse receiving 5x105 retrovirally tagged bone marrow cells. Of th ese 55 mice, 15 survived to 1 month and were considered positively e ngrafted. 3 of these mice were used for primary sequence analysis 3 months after transp lant, 5 were used for subseque nt secondary transplants, 5 died between 1 month and 3 months, and 2 failed to show robust engraftment at 3 months. The analysis was performed on FACS sorted hematopoietic and endothelial progenitor cells (Figure 5-5). A separate cohort of mice was used to assess the clonal complexity of bone marrow pool. In this c ohort, 10 mice were each transplanted with 2x106 retrovirally tagged bone marrow cells. Of the 10, 2 mice were analyzed at 2 months following transplant, 2 died, and 6 are awaiting analysis. The clonal analysis of


83 those 2 animals was performed using HSCs enriched by Sca-1+ c-Kit+ LinFACS sorting (Figure 5-6). Figure 5-5. Primary transpla nt endothelial progen itor cell sort. Bone marrow and blood from primary transplants were sorted for the tagged EPC phenotype, GFP+ Sca-1+ Flk1+, and for tagged hematopoi etic cells. Primary transplants had been transplanted with cultured, retr ovirally tagged marrow cells (primary transplant) and were analyzed 3 m onths after transplantation. Tagged hematopoietic cells were first so rted using the GFP+ CD45+ Flk1phenotype, then, when sufficient cells we re available, further divided into CD11b monocytes, B220 B-cells, and CD3 T-cells.


84 Figure 5-6. Primary transplant hematopoietic stem cell enrichment. Bone marrow from animals that had undergone transplantat ion with retrovirally tagged marrow cells was isolated and enriched for th e HSC phenotype, Sca-1+ c-Kit+ Lin-, and the expression of the retroviral G FP. These sorted cells were used to assess the size of the clonal pool of enriched HSCs, 2 to 3 months after transplant. Bone marrow and peripheral blood from th ese primary transplanted mice were sorted, collected, and analyzed as discussed in the Chapter 2. The resu lts used to clonally assess EPC origins in primary whole bone marrow transplanted mice were generated from an analysis of 94 clones from three pr imary transplanted mi ce (Figure 5-9). These data show monoclonal and oli goclonal activity in the three mice transplanted with 5x105 cells. Clonal tags were either identical or overlapping in th e hematopoietic (CD45+ Flk1or CD45+ Flk1Lin+) and EPC (Flk1+ S ca-1+) populations. Since monoclonality and oligoclonality in the hematopoietic pr ogenitor and mature hematopoietic cell compartments of mice is common, the complexity of only the HSC compartment was also assessed. Thus, Sca-1+ c-Kit+ LinG FP+ enriched bone marrow cells were sorted and analyzed using 21 clones from the two mice transplanted with the 2x106 dose. Clonal


85 tags revealed an oligoclona l pool that was more complex than that observed in blood from the other cohort (Figure 5-9). Secondary Transplants Because most of the mature blood cells, th ree months after lethal irradiation and transplant, should be derived from transplanted HSCs, the contribution of transplanted progenitors should be minimal after this ti me. Nevertheless, HSC enriched secondary transplants were also performed to elimin ate the slight chance that transplanted long-lived progenitors had contaminated pr imary transplants. Five of the primary transplanted mice (5x105 cell cohort) were used for s econdary bone marrow transplants. Bone marrow from these primary transplants was harvested 3 months after the initial transplant. Reconstitution doses less than 1x107 whole bone marrow cells or 5000 Sca-1+/c-Kit+ cells were not viable. Two mice were generate d from transplantation of 25000 Sca-1+/c-Kit+ cells. Both of those mice originated from the same primary transplant mouse. FACS Sorting was used to enrich for hematopoietic (CD45+ Flk1-) and endothelial progenitor cells (Fl k1+ Sca-1+), as before (Fig ure 5-7). Photographs of the two populations, following FACS sort, visibly demonstrate the rare nature of the EPC when compared to hematopoietic cells (Fi gure 5-8). Sequence analysis of 52 clones from these two mice (Figure 5-9) was performed. Th ese data show monoclonality in both mice with the exception of two clones from the bone marrow of one animal. The tags observed in both hematopoietic and EPC populations were nearly identical within each mouse, but distinct between mice, suggesting that differe nt stem cells were transplanted, engrafted, or activated from the bone marrow of primary transplanted mice.


86 Figure 5-7. Secondary transplant endothel ial progenitor cell sort. Bone marrow and blood from secondary transplants were sorted for the tagged EPC phenotype, GFP+ Sca-1+ Flk1+, and for tagged hematopoietic cells. Secondary transplants had been transp lanted with 25000 Sca-1+ c-Kit+ cells which were obtained from the bone marro w of primary transplants and were analyzed three months after transpla ntation. Tagged hematopoietic cells were first sorted using the GFP+ CD45+ Flk1phenotype, then, when sufficient cells were available, furt her divided into CD11b monocytes, B220 B-cells, and CD3 T-cells. Figure 5-8. Photographs of e ndothelial progenitor enriched cells. Photographs of GFP+ CD45+ Flk1(hematopoietic) cells (p anels A and B) and GFP+ Sca-1+ Flk1+ (endothelial progenitor) cells (panels C and D) sorted from bone marrow of a mouse transplanted with retrovirally ta gged cultured bone marrow cells. The relative numbers demonstrate that the EPC is an extremely rare cell.


87 Figure 5-9. Clonal analysis of transpla nted mice. Blood and bone marrow EPC and hematopoietic populations were clona lly analyzed from both primary and secondary transplanted mice. Each co lor represents a unique clone, which was detected by retroviral tag amplif ication, cloning, and sequencing. These results show oligoclonality, tending to monoclonality in both primary and secondary transplants. Furthermore, b ecause the same clones generate blood and EPCs, clonal hemangioblast activity was observed. Lastly, an increased complexity in the enriched HSC compartment of primary transplants suggests that although multiple clones are present, they may not be participating in hematopoiesis and hemangioblastic activity.


88 CHAPTER 6 DISCUSSION Enrichments for the Endothelial Progenitor Cell and the Hematopoietic Stem Cell With the abundance of evidence suggesting that there is a bone marrow derived EPC that, in some manner, promotes and physically contributes to new vascular structures in the adult, de bate regarding the origin of this cell has grown. Several phenotypic characteristics have been used to attempt to purify the EPC population to homogeneity. Use of one of the most promising markers, Flk1 (KDR / VEGFR2), very often results in such a small population of cells that a more elaborate and selective enrichment scheme is challenging. To my knowledge, no reports have shown that Flk1+ cells from adults have hematopoietic activ ity. The murine long-term reconstituting cell, in fact, does not express Fl k1 and is not incapacitated by VEGFR2 blocking antibodies (Haruta et al. 2001; Rafii et al. 2003). Surprisingly, human KDR+ cells have shown HSC activity in one report (Ziegler et al. 1999) . Without multiple markers, EPC enrichment often leads to heterogeneous populations that can ultimately cause imprecise conclusions. In humans, the availability of AC133 to de tect a specific epitope of Prominin-1 has proven useful in distinguishing Flk1+ circul ating endothelial cells from Flk1+ cEPCs (Shmelkov et al. 2005). Unfortunately AC133 is not yet available in mice. In mice, Sca-1 has been shown to enrich for EPCs (Iwakura et al. 2003; Shaw et al. 2004; Chavakis et al. 2005). Sca-1, like AC133, is expressed on HSCs as well as EPCs. For this study, the clonal origins of murine hematopoietic and endothelial progenitor cells were compared using the EPC phenotypic expression of Sca1 and Flk1. This phenotype isolates a very

PAGE 100

89 rare population of cells from the bone marro w and peripheral blood demonstrating EPC activity both in vivo and in vitro. One of the more intriguing characteristics of EPCs, a least a subset of them, is that they are derived from bone marrow stem cells. Transplantation experiments using genetically marked bone marrow cells show EPC activity originating from these transplanted cells. Arguments have been made that even highly enriched HSC populations, such as Sca-1+ c-Ki t+ Lin(SKL), may contain stem cells other than HSCs, which generate these bone marrow derived EPCs . However, our lab has used a single-cell transplant approach to confir m the hemangioblast activity of an HSC with this phenotype. This approach definitely show ed that it is possible for a SKL cell to have hemangioblast activity, at least under the extr eme selective pressure of single-cell transplant. This approach does not exclude the possibility that SKL repopulating HSCs are also a heterogeneous group. One might speculate that some HSCs may have a hematopoietic tendency, some may have an endothelial te ndency, some may have a more primitive hemangioblastic capacity, and some may even have a more plastic potential endowed with or without engraftment capacities. Thesis Questions Several immediate questions arose from the observations of HSC hemangioblast activity. First, is ther e another bone marrow cell that pref erentially generates EPCs when the stringency of single-cell tr ansplantation is relaxed? Si nce no other transplantable adult engrafting stem cell has demons trated endothelial potential, except culture-generated MAPCs (Rey es et al. 2002), the hypothesis was that the HSC is unique in this capacity. Second, if the HSC is the onl y transplantable stem cell with endothelial potential, is the hemangioblast activity clonal? The clonal succession theory of

PAGE 101

90 hematopoiesis suggests that a limited number of hematopoietic stem cell clones are active at any one time and that this clonal pool may fluctuate over time or with certain types of injury. It is therefore possi ble that distinct sets of HSC clones are actively producing blood and endothelial progenitor cells. Because of the interdependence of hematopoiesis and vascular structures in embryonic developm ent and adult vascular repair and because of the evidence supporting hemangioblastic st em cells, the favored hypothesis was that hematopoietic and endothelial progenitor cel ls are produced from the same clonally active bone marrow stem cell. Studying Clonality In-Vivo Various options were considered for studyi ng clonality. Consideration was given to the fact that there was a significant body of work that utilized the randomness of retroviral integration as a unique way of marking populations of cells. However, there were a couple of assumptions about this technique that I wanted to avoid. One assumption, that the retrovirus integrates randomly, has been challenged by work that demonstrates higher integration frequencies in specific regions of transcriptionally active genes (Laufs et al. 2003; Wu et al. 2003; Laufs et al. 2004a). However, although perhaps nonrandom, integration sites may be sufficien tly variable for this assumption to be adequate in many circumstances. The signi ficant overlap in gene expression and phenotype of early EPCs, HSCs, and hema ngioblasts may have complicated this assumption. Another assumption, that the integrati on site is stable for the life of the stem cell and all of its progeny, is weakened by evidence that re troviruses like MLV, HTLV-1, and HIV-1 may act like LTR retrotransposons (Kazazian 2004; Brandt et al. 2005) and have genetic mobility. If this were the case, conclusions based on fluctuating integration

PAGE 102

91 sites over time would certainly be weakened as well as conclusions that sampled progeny were from different stem cells. In this study, I designed an approach that used the permanency of genetic marking but did not rely on a random and stable integration event to demonstrate clonal hemangioblast activity in transp lanted adult bone marrow cells. I validated this technique by verifying that the complexity of the retr oviral pool was reflected in the identified tagging complexity of both NIH3T3 a nd mouse bone marrow cells. A pool of 200 retroviral tags was used, which although small, can readily be scaled up to a much more complex tagging pool. By knowing the degree of co mplexity of the retroviral tags (versus use of an unknown number of integration site s) multinomial statistics can be used to confirm the significance of an observed result . This scaleable complexity offers an advantage over tags with an unknown complexit y. Another advantage of this approach is the sensitivity. Using PCR amplification a nd sequencing it was possible to reliably analyze DNA from as few as 100 cells. This is in comparison to Southern blots, which are typically used for integration site an alysis, that require DNA from 10000 to 100000 cells. This allows for analysis of a more defined population from a single mouse. Because pools of even 100 cells can have a variety of ta gs, the tags must be read as single copies. There are a number of approaches solve this dilemma. One approach would be to isolate single cells, either by FACS, micromanipul ation, or laser capt ure technology, and amplify the details of the re troviral tag sufficiently for se quence analysis. This method proved challenging due to the inefficiency of single-cell PCR. Instead, cloning and sampling of the bacterial clones was utilize d. Although more costly, this technique was more reliable.

PAGE 103

92 Identification of Hemangioblast Activity Retroviral tag analysis was performed from EPC and hematopoietic cells from bone marrow and peripheral blood of mice transp lanted with tagged whole bone marrow containing an estimated 5 HSCs. The results of the clonal anal ysis of this cohort of mice reveal an oligoclonal (tending to monoclonal) pool of stem cells that, in vivo, produce Flk1+ Sca-1+ EPCs and CD45+ Flk1CD11b/B220/Gr1+ hematopoietic cells (Figure 6-1). Figure 6-1. Overlap in oligoclonal pools de monstrating hemangioblast activity. This is an illustration of what was observed. Instead of observing hematopoietic and endothelial progenitor cells originati ng from distinct stem cells (bottom left), they originated from the same stem cells (bottom right). This observation suggests that they orig inated from tagged, transplanted hemangioblasts. The fact that the observed clonality wa s oligoclonal tending to monoclonal was likely due to two factors. Fi rst, other evidence suggests th at mice have an oligoclonal

PAGE 104

93 pool of active hematopoietic stem cells base d on their size (Lemischka et al. 1986; Nolta et al. 1996; Barquinero et al . 2000; Drize et al. 2001; Kurre et al. 2004). This is in contrast to a more polyclonal stem cell activ ity in higher primates and in human stem cells transplanted into mice (Guenechea et al . 2001; Ailles et al. 2002; Schmidt et al. 2002; Kiem et al. 2004). Second, a whole bone marrow population was intentionally transplanted with a limited numbe r of HSCs in order to adhere to the limits of complexity of the retroviral library while maximizing th e heterogeneity of the stem cell pool. An analysis of SKL enriched bone marrow HSCs from 2-3 month post-transplant animals, transplanted with 4-fold more bone marrow ce lls, did reveal that a more polyclonal pool of stem cells was tagged and was able to home to the bone marrow, despite not all activity producing mature blood a nd endothelial progenitor cells. Because there was a slight possibility th at long-lived (>3 month) progenitors may have been tagged, transplanted, and contam inated the clonal analysis, a homogenous population of stem cells was also transplant ed. Secondary transplants from the primary transplant bone marrow were performed. Tr ansplantation of 25000 S+K+ HSC enriched bone marrow cells proved successful. Although this number is larger than would be required for typical C57Bl/6 tran splantation, the requirement ma y reflect a lack of lineage selection and/or decreased potency due to 5-FU treatment and 4 days of in-vitro culture. Nevertheless, a clonal profile, similar to th e primary animals was observed. This was to be predicted based on the results from th e primary transplant. If progenitor cell contamination had been an issue, it would have manifested itself as a population of EPCs or mature blood cells with highly complex nonoverlapping clonal origins in the primary transplants. This had not been seen.

PAGE 105

94 General Conclusion Overall, this thesis presents a novel t echnique for tracking the clonal origins of stem cell progeny. Based on the design, it offers some advantages over traditional retroviral tracking techniques. Furthermore, this technique wa s used to demonstrate that blood is derived from an oligoc lonal (tending to monoclonal) p ool of stem cells and that those stem cells also generate endothelia l progenitor cells af ter whole bone marrow transplants. This finding suggests that the adult bone marrow hemangioblast is transplantable, mimics the he matopoietic stem cell phenotype , and its engraftment can be monitored using hematopoietic or endothelial cells. Furthermore, gene therapy of transplantable hemangioblasts may some day be used to provide a life-long treatment option for a myriad of vascular diseases by modulating the activity of the endothelial progenitor cells that it can generate. Future Studies I would like to begin the fu ture studies section by ac knowledging some limitations of the system that I used, si nce a consideration of these sh ould be the first steps in moving this work forward. First, the number of tags used in this work was 200. In retrospect, I now think that th e power of this technique would be much greater with a larger retroviral library, although, as is, it is approxima tely as good as retroviral integration site analysis (HIV has approximately 300 integration sites). When scaling up to a larger library size, I w ould expect that certain combin ations of random nucleotides would be toxic to various cells (bacteria, packaging cells, transduced cells). This may even have been observed in this work as a slight reduction in complexity when going from the NIH3T3 transduction to the bone marrow transduction. Nevertheless, this is

PAGE 106

95 something that one may need to consider in or der to accurately assess the complexity of the libray. Second, FACS sorting is not an exact te chnique. Populations sorted by FACS are merely enriched and not pure. Unfortunatel y, due to the low number of Sca-1+ Flk1+ EPCs in mice it is difficult to perform a seconda ry sort or use a more selective phenotype. One solution would be stimulation of EPC mobilization prior to harvest using SDF-1, G-CSF, or vascular trauma. A great follow-up study that woul d also address this issue would be to look at fully diffe rentiated endothelial cells that had originated from tagged EPCs. This was actually attempted using the retinal neovascularizataion model developed by our lab, but met with limited success. The di fficulty was likely not with the approach, but rather with technical factor s related to degree of injury. However, I still think that it would be possible to achieve GFP+ retrovira lly tagged neovascularization using this or another model, which could be then be inte rrogated using the clonal analysis techniques in this thesis. Since lasercapture microscopy followed by PCR has been used with individual cells, I believe this enhancement holds the signific ant promise. Furthermore, I have made progress in culturing mouse bone marrow derived EPCs and differentiating them into adherent endothelial cells, in culture. However, the optimal conditions for this process are largely unknown. Regardless, this technique would certainly be a great next step in studying the clon al origins of the EPC. With those limitations addressed, genetic ma rking of populations is one of the best ways of analyzing the lineage pot ential of progenitor and stem cells. In vitro culturing at clonal densities is extremely he lpful too, but very rarely mimi cs the complexity of all the interactions and regulation that takes place with the living organism. Single-cell transplantation is essentially a cell marking t echnique, but it is notorio usly inefficient and

PAGE 107

96 can be difficult to analyze. Therefore, I beli eve that genetic marking techniques could be one of the most valuable tool s for demonstrating the true pl asticity of stem cells. Indeed, a colleague and I have already had some success using the MIGRN tag library with neurospheres, a neural stem/progenitor cell. The main advantage that cell marking has over approaches such as transplantation of a transgenic GFP population is that it is possible to gain invaluable information about the clonal fate of potentially impure populations. Investigations in this thesis began with the possibility that there was another stem cell that, preferentially, c ontributes to the endothelia l cell population. The MAPC (multipotent adult progenitor cell) was a likely suspect. It is a cell derived from long-term culture of bone marrow. The phenotype of it is not the same as the HSC. The MAPC has been shown to have enormous plasticity in a very limited number of studies. The research in this thesis indicates that the same tr ansplantable cell that generates blood also generates endothelial progenito r cells. The only way to inco rporate the MAPC into the observations of this thesis is to propose that the MAPC is actually some type of reprogrammed stem cell, which may under ce rtain conditions generate the HSC-like hemangioblasts investigated in this wor k. Since whole bone marrow was used in the primary transplant, if there were a separa te, EPC-generating MAPC in the bone marrow that was unrelated to the HSC then it would have been observed in these studies. I feel that stem cell plasticity is likely to be much larger than many of the skeptics profess. In addition to endothe lial tissue, reports have been published that demonstrate that the HSC has the capacity to differentiate into hepatocytes, skeletal muscle, cardiac muscle, pancreatic cells, intestinal epithelial cells, and neurons. The difficulty is in determining how to unlock that plasticity. Due to often low levels of contribution in HSC

PAGE 108

97 plasticity studies, an understanding of the inte rnal genetic switches would be necessary to modulation of the amount and type of plastic repair. The ability to clone full living creatures from a nuclei of an adult cell is a testament to the potential plasticity of adult stem cells. Because embryonic stem cells seem to have a much easier time demonstrating plasticity, they will likely prove extremely us eful in learning how to, in a way, pick the lock. Genetic marking techniques, such as th e one presented in this thesis, along with a host of other powerful genetic techniques will play a critical role in these future explorations. With knowledge of how to direct the programming of stem cells, many new and powerful medical therapies will be available phys icians. Imagine being told, “The arteries in your hippocampus are showing signs of ag ing and this is accounting for your memory loss. So, I’m going to schedule you for a sess ion of endothelial re pair and collateral enhancement.” Based on the finding that tran splantable HSCs routinely generate EPCs and have hemangioblast activity, this ther apy might entail use of antibody-guided microparticles, which release chemokines in the hippocampus to promote a balanced enhancement of the bone marrow derived vascul arization. Or it might entail injection of cells from your local stem cell bank account , which had been genetically reprogrammed to home to the subtly ischemic region of th e hippocampus and promote balanced vascular regeneration. Stem cell-based tissue rege neration may seem like science fiction, particularly with all the nays ayers and technical obstacles. Ho wever, if you were living in 1900, who would have believed that mechanical devices the size of a notebook would be capable of displaying a universe of information, accessible at the touch of a finger. With the observations of those at clinical bedside, with the experiments of those at the research

PAGE 109

98 bench top, and with the clinical trials of those pushing medici ne forward, science fiction becomes science reality.

PAGE 110

99 LIST OF REFERENCES Adams, V., K. Lenk, A. Linke, D. Lenz, S. Erbs, M. Sandri, A. Tarnok, S. Gielen, F. Emmrich, G. Schuler and R. Hambrech t (2004). "Increase of circulating endothelial progenitor cells in patients with coronary ar tery disease after exercise-induced ischemia." Arterioscler Thromb Vasc Biol 24 (4): 684-90. Aicher, A., A. M. Zeiher and S. Dimmel er (2005). "Mobilizing endothelial progenitor cells." Hypertension 45 (3): 321-5. Ailles, L., M. Schmidt, F. R. Santoni de Sio, H. Glimm, S. Cavalieri, S. Bruno, W. Piacibello, C. Von Kalle and L. Naldini ( 2002). "Molecular evid ence of lentiviral vector-mediated gene transfer into hu man self-renewing, multi-potent, long-term NOD/SCID repopulating hematopoietic cells." Mol Ther 6 (5): 615-26. Alliot, F., I. Godin and B. Pessac (1999). "Mic roglia derive from progenitors, originating from the yolk sac, and which proliferate in the brain." Brai n Res Dev Brain Res 117 (2): 145-52. Amselgruber, W. M., M. Schafer and F. Sinowatz (1999). "Angiogenesis in the bovine corpus luteum: an immunocytochemical a nd ultrastructural st udy." Anat Histol Embryol 28 (3): 157-66. Apte, S. M., D. Fan, J. J. Killion and I. J. Fidler (2004). "Targeting the platelet-derived growth factor receptor in antivascular therapy for human ovarian carcinoma." Clin Cancer Res 10 (3): 897-908. Ara, T., K. Tokoyoda, T. Sugiyama, T. Egawa, K. Kawabata and T. Nagasawa (2003). "Long-term hematopoietic stem cells requi re stromal cell-derived factor-1 for colonizing bone marrow dur ing ontogeny." Immunity 19 (2): 257-67. Arai, F., A. Hirao, M. Ohmura, H. Sato, S. Matsuoka, K. Takubo, K. Ito, G. Y. Koh and T. Suda (2004). "Tie2/angiopoietin-1 signa ling regulates hematopoietic stem cell quiescence in the bone marrow niche." Cell 118 (2): 149-61. Arai, F., O. Ohneda, T. Miyamoto, X. Q. Zh ang and T. Suda (2002). "Mesenchymal stem cells in perichondrium expr ess activated leukocyte cell adhesion molecule and participate in bone marrow formation." J Exp Med 195 (12): 1549-63. Asahara, T., T. Murohara, A. Sullivan, M. Silver, R. van der Zee, T. Li, B. Witzenbichler, G. Schatteman and J. M. Isner (1997 ). "Isolation of putative progenitor endothelial cells for angiogenesis." Science 275 (5302): 964-7.

PAGE 111

100 Asahara, T., T. Takahashi, H. Masuda, C. Kalk a, D. Chen, H. Iwaguro, Y. Inai, M. Silver and J. M. Isner (1999). "VEGF contribut es to postnatal neovascularization by mobilizing bone marrow-derived endothelial progenitor cells." Embo J 18 (14): 3964-72. Bachelder, R. E., M. A. Wendt and A. M. Me rcurio (2002). "Vascula r endothelial growth factor promotes breast ca rcinoma invasion in an au tocrine manner by regulating the chemokine receptor CXCR4." Cancer Res 62 (24): 7203-6. Bahlmann, F. H., K. de Groot, O. Mueller, B. Hertel, H. Haller and D. Fliser (2005). "Stimulation of endothelial progenitor cells: a new putative therapeutic effect of angiotensin II receptor an tagonists." Hypertension 45 (4): 526-9. Bahlmann, F. H., K. De Groot, J. M. Spandau, A. L. Landry, B. Hertel, T. Duckert, S. M. Boehm, J. Menne, H. Haller and D. Flis er (2004). "Erythr opoietin regulates endothelial progenitor cells." Blood 103 (3): 921-6. Bahlmann, F. H., K. DeGroot, T. Duckert, E. Niemczyk, E. Bahlmann, S. M. Boehm, H. Haller and D. Fliser (2003). "Endothe lial progenitor cell proliferation and differentiation is regulated by erythropoietin." Kidney Int 64 (5): 1648-52. Bailey, A. S., S. Jiang, M. Afentoulis, C. I. Baumann, D. A. Schroede r, S. B. Olson, M. H. Wong and W. H. Fleming (2004). "Tra nsplanted adult hematopoietic stems cells differentiate into functional endothelial cells." Blood 103 (1): 13-9. Barker, J. E. (1968). "Development of the mouse hematopoietic system. I. Types of hemoglobin produced in embryonic yolk sac and liver." Dev Biol 18 (1): 14-29. Barleon, B., S. Sozzani, D. Zhou, H. A. Weich, A. Mantovani and D. Marme (1996). "Migration of human monocytes in response to vascular endothelial growth factor (VEGF) is mediated via the VEGF receptor flt-1." Blood 87 (8): 3336-43. Barquinero, J., J. C. Segovia, M. Ramirez, A. Limon, G. Guenechea, T. Puig, J. Briones, J. Garcia and J. A. Bueren (2000). "Effi cient transduction of human hematopoietic repopulating cells generating stable engraf tment of transgene-expressing cells in NOD/SCID mice." Blood 95 (10): 3085-93. Bernard, J. (1991). "The erythroblastic island: past and future." Blood Cells 17 (1): 5-10; discussion 10-4. Bertrand, J. Y., S. Giroux, R. Golub, M. Klaine , A. Jalil, L. Boucontet, I. Godin and A. Cumano (2005). "Characterization of purif ied intraembryonic hematopoietic stem cells as a tool to define their site of origin." Proc Natl Acad Sci U S A 102 (1): 134-9. Bethlenfalvay, N. C. and M. Block (1970) . "Fetal erythropoie sis. Maturation in megaloblastic (yolk sac) er ythropoiesis in the C 57 B1-6J mouse." Acta Haematol 44 (4): 240-5.

PAGE 112

101 Betsholtz, C., P. Lindblom and H. Gerhardt (2005). "Role of peri cytes in vascular morphogenesis." Exs (94): 115-25. Bossard, C., L. Van den Berghe, H. Laurell, C. Castano, M. Cerutti, A. C. Prats and H. Prats (2004). "Antiangiogenic properties of fibstatin, an ex tracellular FGF-2binding polypeptide." Cancer Res 64 (20): 7507-12. Brandt, J., S. Schrauth, A. M. Veith, A. Froschauer, T. Haneke, C. Schultheis, M. Gessler, C. Leimeister and J. N. Volff (2005). "Transposable elements as a source of genetic innovation: expression and evol ution of a family of retrotransposonderived neogenes in mammals." Gene 345 (1): 101-11. Brotherton, T. W., D. H. Chui, J. Gauldie and M. Patterson (1979). "Hemoglobin ontogeny during normal mouse fetal devel opment." Proc Natl Acad Sci U S A 76 (6): 2853-7. Brown, E. B., R. B. Campbell, Y. Tsuzuki, L. Xu, P. Carmeliet, D. Fukumura and R. K. Jain (2001). "In vivo measurement of gene expression, angiogenesis and physiological function in tumors usi ng multiphoton laser scanning microscopy." Nat Med 7 (7): 864-8. Burri, P. H., R. Hlushchuk and V. Djonov ( 2004). "Intussusceptive angiogenesis: its emergence, its characteristics, and its significance." Dev Dyn 231 (3): 474-88. Butler, J. M., S. M. Guthrie, M. Koc, A. Afzal, S. Caballero, H. L. Brooks, R. N. Mames, M. S. Segal, M. B. Grant and E. W. Sc ott (2005). "SDF-1 is both necessary and sufficient to promote prolifera tive retinopathy." J Clin Invest 115 (1): 86-93. Byrd, N. and L. Grabel (2004). "Hedgehog si gnaling in murine vasculogenesis and angiogenesis." Trends Cardiovasc Med 14 (8): 308-13. Calvi, L. M., G. B. Adams, K. W. Weibrecht, J. M. Weber, D. P. Olson, M. C. Knight, R. P. Martin, E. Schipani, P. Divieti, F. R. Bringhurst, L. A. Milner, H. M. Kronenberg and D. T. Scadden (2003). "Osteoblastic cells regulate the haematopoietic stem cell niche." Nature 425 (6960): 841-6. Caprioli, A., T. Jaffredo, R. Gautier, C. Dubourg and F. Dieterlen-Lievre (1998). "Blood-borne seeding by hematopoietic and endothelial precursors from the allantois." Proc Natl Acad Sci U S A 95 (4): 1641-6. Charbord, P., M. Tavian, L. Humeau and B. P eault (1996). "Early ontogeny of the human marrow from long bones: an immunohistoc hemical study of hematopoiesis and its microenvironment." Blood 87 (10): 4109-19. Chavakis, E., A. Aicher, C. Heeschen, K. Sasa ki, R. Kaiser, N. El Makhfi, C. Urbich, T. Peters, K. Scharffetter-Kochanek, A. M. Zeiher, T. Chavakis and S. Dimmeler (2005). "Role of beta2-integrins for homi ng and neovascularization capacity of endothelial progenitor cells." J Exp Med 201 (1): 63-72.

PAGE 113

102 Choi, K., M. Kennedy, A. Kazarov, J. C. Pa padimitriou and G. Keller (1998). "A common precursor for hematopoietic and endothelial cells." Development 125 (4): 725-32. Christensen, J. L., D. E. Wright, A. J. Wa gers and I. L. Weissman (2004). "Circulation and chemotaxis of fetal hematopoietic stem cells." PLoS Biol 2 (3): E75. Chung, Y. S., W. J. Zhang, E. Arentson, P. D. Kingsley, J. Palis and K. Choi (2002). "Lineage analysis of the hemangioblast as defined by FLK1 and SCL expression." Development 129 (23): 5511-20. Ciau-Uitz, A., M. Walmsley and R. Patient (2000). "Distinct origins of adult and embryonic blood in Xenopus." Cell 102 (6): 787-96. Ciulla, M. M., A. Giorgetti, L. Lazzari, M. Cortiana, I. Silvestris, G. Annoni, C. De Asmundis, A. V. Fiore, E. Montelatici, R. Paliotti, F. Magrini, P. Rebulla and A. Cortelezzi (2005). "High-Alti tude trekking in the Himala yas increases the activity of circulating endothelial cells." Am J Hematol 79 (1): 76-8. D'Amore, P. A. (1992). "Capillary growth : a two-cell system." Semin Cancer Biol 3 (2): 49-56. Darland, D. C. and P. A. D'Amore (1999) . "Blood vessel maturation: vascular development comes of age." J Clin Invest 103 (2): 157-8. de Bruijn, M. F., X. Ma, C. Robin, K. Otters bach, M. J. Sanchez and E. Dzierzak (2002). "Hematopoietic stem cells localize to the endothelial cell layer in the midgestation mouse aorta." Immunity 16 (5): 673-83. de Bruijn, M. F., M. C. Peeters, T. Luteij n, P. Visser, N. A. Speck and E. Dzierzak (2000a). "CFU-S(11) activity does not lo calize solely with the aorta in the aorta-gonad-mesonephros region." Blood 96 (8): 2902-4. de Bruijn, M. F., N. A. Speck, M. C. Peeters and E. Dzierz ak (2000b). "Definitive hematopoietic stem cells first develop w ithin the major arterial regions of the mouse embryo." Embo J 19 (11): 2465-74. Dick, J. E., G. Guenechea, O. I. Gan and C. Dorrell (2001). "In vivo dynamics of human stem cell repopulation in NOD/SC ID mice." Ann N Y Acad Sci 938 : 184-90. Dick, J. E., M. C. Magli, D. Huszar, R. A. Phillips and A. Bernstein (1985). "Introduction of a selectable gene into primitive stem cells capable of long-term reconstitution of the hemopoietic syst em of W/Wv mice." Cell 42 (1): 71-9. Dieterlen-Lievre, F., L. Pardanaud, K. Bolle rot and T. Jaffredo (2002). "Hemangioblasts and hemopoietic stem cells during ontogeny." C R Biol 325 (10): 1013-20.

PAGE 114

103 Dimmeler, S., A. Aicher, M. Vasa, C. Mil dner-Rihm, K. Adler, M. Tiemann, H. Rutten, S. Fichtlscherer, H. Martin and A. M. Zeiher (2001). "H MG-CoA reductase inhibitors (statins) increase endothelial progenitor cells via th e PI 3-kinase/Akt pathway." J Clin Invest 108 (3): 391-7. Dooley, K. A., A. J. Davidson and L. I. Zon (2005). "Zebrafish scl functions independently in hematopoietic and endothelial development." Dev Biol 277 (2): 522-36. Drake, C. J. (2003). "Embryonic and adult vasculogenesis." Birth Defects Res C Embryo Today 69 (1): 73-82. Drake, C. J. and P. A. Fleming (2000). "V asculogenesis in the day 6.5 to 9.5 mouse embryo." Blood 95 (5): 1671-9. Drize, N. J., Y. V. Olshanskaya, L. P. Gera simova, T. E. Manakova, N. L. Samoylina, T. V. Todria and J. L. Chertkov ( 2001). "Lifelong hematopoiesis in both reconstituted and sublethally irradiated mice is provided by multiple sequentially recruited stem cells." Exp Hematol 29 (6): 786-94. D'Souza, S. L., A. G. Elefanty and G. Keller (2005). "SCL/Tal-1 is essential for hematopoietic commitment of the hemangi oblast, but not for its development." Blood . Dzierzak, E., A. Medvinsky and M. de Bru ijn (1998). "Qualitative and quantitative aspects of haematopoietic cell development in the mammalian embryo." Immunol Today 19 (5): 228-36. Ema, M., P. Faloon, W. J. Zhang, M. Hirashim a, T. Reid, W. L. Stanford, S. Orkin, K. Choi and J. Rossant (2003). "Combinatori al effects of Flk1 and Tal1 on vascular and hematopoietic development in the mouse." Genes Dev 17 (3): 380-93. Emery, D. W., R. G. Andrews and T. Papayannopoulou (2000). "Differences among nonhuman primates in susceptibility to bone marrow progenitor transduction with retrovirus vectors." Gene Ther 7 (5): 359-67. Fadini, G. P., M. Miorin, M. Facco, S. Bonamico, I. Baesso, F. Grego, M. Menegolo, S. V. de Kreutzenberg, A. Tiengo, C. Agos tini and A. Avogaro (2005). "Circulating endothelial progenitor cells are reduced in peripheral vascular complications of type 2 diabetes mellitus." J Am Coll Cardiol 45 (9): 1449-57. Faloon, P., E. Arentson, A. Kazarov, C. X. De ng, C. Porcher, S. Orkin and K. Choi (2000). "Basic fibroblast growth factor positively regulates hematopoietic development." Development 127 (9): 1931-41. Fang, B., C. Zheng, L. Liao, Q. Han, Z. Sun, X. Jiang and R. C. Zhao (2005). "Identification of human chronic myel ogenous leukemia progenitor cells with hemangioblastic characteristics." Blood 105 (7): 2733-40.

PAGE 115

104 Fialkow, P. J., R. J. Jacobson and T. Papayannopoulou (1977). "Chronic myelocytic leukemia: clonal origin in a stem cell co mmon to the granulocyte, erythrocyte, platelet and monocyte/ macrophage." Am J Med 63 (1): 125-30. Fleischman, R. A., R. P. Custer and B. Mintz (1982). "Totipoten t hematopoietic stem cells: normal self-renewal and differentia tion after transplant ation between mouse fetuses." Cell 30 (2): 351-9. Fong, G. H., J. Rossant, M. Gertsenstein and M. L. Breitman (1995). "Role of the Flt-1 receptor tyrosine kinase in regulating the assembly of vascular endothelium." Nature 376 (6535): 66-70. Fong, G. H., L. Zhang, D. M. Bryce and J. Peng (1999). "Increased hemangioblast commitment, not vascular disorganization, is the primary defect in flt-1 knock-out mice." Development 126 (13): 3015-25. Garcia-Porrero, J. A., I. E. Godin and F. Dieterlen-Lievre (1995). "Potential intraembryonic hemogenic sites at pre-live r stages in the mouse." Anat Embryol (Berl) 192 (5): 425-35. Gekas, C., F. Dieterlen-Lievre, S. H. Orki n and H. K. Mikkola (2005). "The Placenta Is a Niche for Hematopoietic Stem Cells." Dev Cell 8 (3): 365-375. George, J., E. Goldstein, S. Abashidze, V. Deutsch, H. Shmilovich, A. Finkelstein, I. Herz, H. Miller and G. Keren (2004). "Cir culating endothelial progenitor cells in patients with unstable angina: association with systemic inflammation." Eur Heart J 25 (12): 1003-8. Gerhardt, H. and C. Betsholtz (2003). "Endothelial-pericyte in teractions in angiogenesis." Cell Tissue Res 314 (1): 15-23. Gering, M., Y. Yamada, T. H. Rabbitts and R. K. Patient (2003). "Lmo2 and Scl/Tal1 convert non-axial mesoderm into haemangioblasts which differentiate into endothelial cells in the abse nce of Gata1." Development 130 (25): 6187-99. Gill, M., S. Dias, K. Hattori, M. L. Rivera, D. Hicklin, L. Witte, L. Girardi, R. Yurt, H. Himel and S. Rafii (2001). "Vascular trauma induces rapid but transient mobilization of VEGFR2(+)AC133(+) e ndothelial precursor cells." Circ Res 88 (2): 167-74. Godin, I. and A. Cumano (2002). "The hare and the tortoise: an embryonic haematopoietic race." Nat Rev Immunol 2 (8): 593-604. Goodell, M. A., K. Brose, G. Paradis, A. S. Conner and R. C. Mulligan (1996). "Isolation and functional properties of murine hemat opoietic stem cells that are replicating in vivo." J Exp Med 183 (4): 1797-806.

PAGE 116

105 Gourdie, R. G., B. S. Harris, J. Bond, C. Ju stus, K. W. Hewett, T. X. O'Brien, R. P. Thompson and D. Sedmera (2003). "Development of the cardiac pacemaking and conduction system." Birth Defects Res C Embryo Today 69 (1): 46-57. Grant, M. B., W. S. May, S. Caballero, G. A. Brown, S. M. Guthrie, R. N. Mames, B. J. Byrne, T. Vaught, P. E. Spoerri, A. B. Peck and E. W. Scott (2002). "Adult hematopoietic stem cells provide functi onal hemangioblast activity during retinal neovascularization." Nat Med 8 (6): 607-12. Guenechea, G., O. I. Gan, C. Dorrell and J. E. Dick (2001). "Distinct classes of human stem cells that differ in proliferative and self-renewal pot ential." Nat Immunol 2 (1): 75-82. Guillemin, G. J. and B. J. Brew (2004). "Microglia, macrophages, perivascular macrophages, and pericytes: a review of function and identification." J Leukoc Biol 75 (3): 388-97. Gulati, R., D. Jevremovic, T. E. Peterson, S. Chatterjee, V. Shah, R. G. Vile and R. D. Simari (2003). "Diverse origin and func tion of cells with endothelial phenotype obtained from adult human blood." Circ Res 93 (11): 1023-5. Gunsilius, E. (2003). "Evidence from a leukemia model for maintenance of vascular endothelium by bone-marrow-derived e ndothelial cells." Adv Exp Med Biol 522 : 17-24. Guthrie, S. M., L. M. Curtis, R. N. Mames, G. G. Simon, M. B. Grant and E. W. Scott (2004). "The nitric oxide pathway modul ates hemangioblast activity of adult hematopoietic stem cells." Blood . Haar, J. L. and G. A. Ackerman (1971). "A phase and electron microscopic study of vasculogenesis and erythropoi esis in the yolk sac of the mouse." Anat Rec 170 (2): 199-223. Hamaguchi, I., X. L. Huang, N. Takakura, J. Tada, Y. Yamaguchi, H. Kodama and T. Suda (1999). "In vitro hematopoietic and endothelial cell devel opment from cells expressing TEK receptor in murine aorta-gonad-mesonephros region." Blood 93 (5): 1549-56. Hammes, H. P., J. Lin, P. Wagner, Y. Feng, F. Vom Hagen, T. Krzizok, O. Renner, G. Breier, M. Brownlee and U. Deutsch ( 2004). "Angiopoietin-2 causes pericyte dropout in the normal retina: evidence for involvement in diabetic retinopathy." Diabetes 53 (4): 1104-10. Harraz, M., C. Jiao, H. D. Hanlon, R. S. Hartley and G. C. Schatteman (2001). "CD34blood-derived human endothelial cell progenitors." Stem Cells 19 (4): 304-12.

PAGE 117

106 Harrison, D. E. and C. M. Astle (1997). "Shortand long-term multilineage repopulating hematopoietic stem cells in late feta l and newborn mice: models for human umbilical cord blood." Blood 90 (1): 174-81. Haruta, H., Y. Nagata and K. Todokoro (2001) . "Role of Flk-1 in mouse hematopoietic stem cells." FEBS Lett 507 (1): 45-8. Hattori, K., S. Dias, B. Heissig, N. R. Hack ett, D. Lyden, M. Tateno, D. J. Hicklin, Z. Zhu, L. Witte, R. G. Crystal, M. A. Moore and S. Rafii (2001). "Vascular endothelial growth factor and angiopoietin-1 stimulat e postnatal hematopoiesis by recruitment of vasculogenic and hema topoietic stem cells." J Exp Med 193 (9): 1005-14. Hattori, K., B. Heissig, Y. Wu, S. Dias, R. Te jada, B. Ferris, D. J. Hicklin, Z. Zhu, P. Bohlen, L. Witte, J. Hendrikx, N. R. Hackett, R. G. Crystal, M. A. Moore, Z. Werb, D. Lyden and S. Rafii (2002). "P lacental growth factor reconstitutes hematopoiesis by recruiting VEGFR1 (+) stem cells from bone-marrow microenvironment." Nat Med 8 (8): 841-9. Heeschen, C., A. Aicher, R. Lehmann, S. Fich tlscherer, M. Vasa, C. Urbich, C. MildnerRihm, H. Martin, A. M. Zeiher and S. Dimmeler (2003). "Erythropoietin is a potent physiologic stimulus for endothe lial progenitor cell mobilization." Blood 102 (4): 1340-6. Heissig, B., K. Hattori, S. Dias, M. Friedrich, B. Ferris, N. R. Hackett, R. G. Crystal, P. Besmer, D. Lyden, M. A. Moore, Z. Werb and S. Rafii (2002). "Recruitment of stem and progenitor cells from the bone marrow niche requires MMP-9 mediated release of kit-ligand." Cell 109 (5): 625-37. Hendrix, M. J., E. A. Seftor, P. S. Meltzer, L. M. Gardner, A. R. Hess, D. A. Kirschmann, G. C. Schatteman and R. E. Seft or (2001). "Expression and functional significance of VE-cadherin in aggressi ve human melanoma cells: role in vasculogenic mimicry." Proc Natl Acad Sci U S A 98 (14): 8018-23. Hendrix, M. J., R. E. Seftor, E. A. Seftor, L. M. Gruman, L. M. Lee, B. J. Nickoloff, L. Miele, D. D. Sheriff and G. C. Schatteman (2002). "Transendothelial function of human metastatic melanoma cells: role of the microenvironment in cell-fate determination." Cancer Res 62 (3): 665-8. Hicklin, D. J. and L. M. Ellis (2005). "Role of the vascular endothelial growth factor pathway in tumor growth and angiogenesis." J Clin Oncol 23 (5): 1011-27. Hill, J. M., G. Zalos, J. P. Halcox, W. H. Schenke, M. A. Waclawiw, A. A. Quyyumi and T. Finkel (2003). "Circulati ng endothelial progenitor cells , vascular function, and cardiovascular risk." N Engl J Med 348 (7): 593-600.

PAGE 118

107 Hiratsuka, S., Y. Kataoka, K. Nakao, K. Naka mura, S. Morikawa, S. Tanaka, M. Katsuki, Y. Maru and M. Shibuya (2005). "Vas cular endothelial growth factor A (VEGF-A) is involved in gui dance of VEGF receptor-positive cells to the anterior portion of early embryos." Mol Cell Biol 25 (1): 355-63. Hiratsuka, S., O. Minowa, J. Kuno, T. Noda and M. Shibuya (1998). "Flt-1 lacking the tyrosine kinase domain is sufficient for normal development and angiogenesis in mice." Proc Natl Acad Sci U S A 95 (16): 9349-54. Hristov, M. and C. Weber (2004). "Endothe lial progenitor cells : characterization, pathophysiology, and possible clinic al relevance." J Cell Mol Med 8 (4): 498-508. Hu, J., M. Takatoku, S. E. Sellers, B. A. Agri cola, M. E. Metzger, R. E. Donahue and C. E. Dunbar (2002). "Analysis of origin and optimization of expansion and transduction of circulatin g peripheral blood endothelial progenitor cells in the rhesus macaque model." Hum Gene Ther 13 (17): 2041-50. Hui, E. K., P. C. Wang and S. J. Lo (1998). "Strategies for cloning unknown cellular flanking DNA sequences from foreign integrants." Cell Mol Life Sci 54 (12): 1403-11. Hunting, C. B., W. A. Noort and J. J. Zwaginga (2005). "Circulating endothelial (progenitor) cells reflect the state of th e endothelium: vascular injury, repair and neovascularization." Vox Sang 88 (1): 1-9. Hur, J., C. H. Yoon, H. S. Kim, J. H. Choi , H. J. Kang, K. K. Hwang, B. H. Oh, M. M. Lee and Y. B. Park (2004). "Character ization of two types of endothelial progenitor cells and their different contributions to neovasculogenesis." Arterioscler Thromb Vasc Biol 24 (2): 288-93. Hurwitz, H., L. Fehrenbacher, W. Novotny, T. Cartwright, J. Hainsw orth, W. Heim, J. Berlin, A. Baron, S. Griffing, E. Holmgre n, N. Ferrara, G. Fyfe, B. Rogers, R. Ross and F. Kabbinavar (2004). "Bevacizu mab plus irinotecan, fluorouracil, and leucovorin for metastatic colo rectal cancer." N Engl J Med 350 (23): 2335-42. Iwakura, A., C. Luedemann, S. Shastry, A. Hanley, M. Kearney, R. Aikawa, J. M. Isner, T. Asahara and D. W. Losordo (2003). "Estrogen-mediated, endothelial nitric oxide synthase-dependent mobilizatio n of bone marrow-derived endothelial progenitor cells contributes to reendothelialization after arterial injury." Circulation 108 (25): 3115-21. Iwami, Y., H. Masuda and T. Asahara (2004). "Endothelial progenitor cells: past, state of the art, and future." J Cell Mol Med 8 (4): 488-97. Jaffredo, T., R. Gautier, V. Brajeul and F. Dieterlen-Lievre (2 000). "Tracing the progeny of the aortic hemangioblast in the avian embryo." Dev Biol 224 (2): 204-14.

PAGE 119

108 Jaffredo, T., R. Gautier, A. Eichmann and F. Dieterlen-Lievre (1998). "Intraaortic hemopoietic cells are derived from endothelial cells during ontogeny." Development 125 (22): 4575-83. Jain, R. K. (2005). "Normalization of tumo r vasculature: an emerging concept in antiangiogenic therapy." Science 307 (5706): 58-62. Jamieson, C. H., L. E. Ailles, S. J. Dylla, M. Muijtjens, C. Jones, J. L. Zehnder, J. Gotlib, K. Li, M. G. Manz, A. Keating, C. L. Sawyers and I. L. Weissman (2004). "Granulocyte-macrophage progenitors as candidate leukemic stem cells in blast-crisis CML." N Engl J Med 351 (7): 657-67. Jiang, S., L. Walker, M. Afentoulis, D. A. Anderson, L. Jauron-Mills, C. L. Corless and W. H. Fleming (2004). "Transplanted huma n bone marrow contributes to vascular endothelium." Proc Natl Acad Sci U S A 101 (48): 16891-6. Jodele, S., C. F. Chantrain, L. Blavier, C. Lutzko, G. M. Crooks, H. Shimada, L. M. Coussens and Y. A. Declerck (2005). "The contribution of bone marrow-derived cells to the tumor vasculature in neurobl astoma is matrix metalloproteinase-9 dependent." Cancer Res 65 (8): 3200-8. Jordan, C. T., C. M. Astle, J. Zawadzki, K. Mackarehtschian, I. R. Lemischka and D. E. Harrison (1995). "Long-term repopulating abil ities of enriched fetal liver stem cells measured by competitive repopulation." Exp Hematol 23 (9): 1011-5. Jordan, C. T. and I. R. Lemischka (1990). "C lonal and systemic analysis of long-term hematopoiesis in the mouse." Genes Dev 4 (2): 220-32. Kabrun, N., H. J. Buhring, K. Choi, A. Ullr ich, W. Risau and G. Keller (1997). "Flk-1 expression defines a population of early embryonic hematopoietic precursors." Development 124 (10): 2039-48. Kalka, C., H. Masuda, T. Takahashi, R. Gordon, O. Tepper, E. Gravereaux, A. Pieczek, H. Iwaguro, S. I. Hayashi, J. M. Isner and T. Asahara (2000a). "Vascular endothelial growth factor( 165) gene transfer augments circulating endothelial progenitor cells in human subjects." Circ Res 86 (12): 1198-202. Kalka, C., H. Tehrani, B. Laudenberg, P. R. Va le, J. M. Isner, T. Asahara and J. F. Symes (2000b). "VEGF gene transfer mobilizes endothelial progenitor cells in patients with inoperable coronary disease." Ann Thorac Surg 70 (3): 829-34. Kawamoto, A., T. Asahara and D. W. Loso rdo (2002). "Transplantation of endothelial progenitor cells for therapeutic neovasc ularization." Cardiovasc Radiat Med 3 (3-4): 221-5. Kazazian, H. H., Jr. (2004). "Mobile elements : drivers of genome evolution." Science 303 (5664): 1626-32.

PAGE 120

109 Keller, G., M. Kennedy, T. Papayannopoulou an d M. V. Wiles (1993). "Hematopoietic commitment during embryonic stem cell di fferentiation in culture." Mol Cell Biol 13 (1): 473-86. Keller, G., G. Lacaud and S. Robertson (1999). "Development of the hematopoietic system in the mouse." Exp Hematol 27 (5): 777-87. Kennedy, M., M. Firpo, K. Choi, C. Wall, S. Robertson, N. Kabrun and G. Keller (1997). "A common precursor for primitive erythropoiesis and definitive haematopoiesis." Nature 386 (6624): 488-93. Khakoo, A. Y. and T. Finkel (2005). "Endot helial progenitor cells." Annu Rev Med 56 : 79-101. Kiem, H. P., S. Sellers, B. Thomasson, J. C. Mo rris, J. F. Tisdale, P. A. Horn, P. Hematti, R. Adler, K. Kuramoto, B. Calmels, A. Bonifacino, J. Hu, C. von Kalle, M. Schmidt, B. Sorrentino, A. Nienhuis, C. A. Blau, R. G. Andrews, R. E. Donahue and C. E. Dunbar (2004). "Long-term clin ical and molecular follow-up of large animals receiving retrovirally tran sduced stem and progenitor cells: no progression to clonal hematopoi esis or leukemia." Mol Ther 9 (3): 389-95. Kimelman, D. and K. J. Griffin (2000) . "Vertebrate mesendoderm induction and patterning." Curr Opin Genet Dev 10 (4): 350-6. Klug, C. A., S. Cheshier and I. L. Weissman (2000). "Inactivation of a GFP retrovirus occurs at multiple levels in long-te rm repopulating stem cells and their differentiated progeny." Blood 96 (3): 894-901. Kluger, Y., Z. Lian, X. Zhang, P. E. Ne wburger and S. M. Weissman (2004). "A panorama of lineage-specific transcri ption in hematopoiesis." Bioessays 26 (12): 1276-87. Koide, N., A. Nishio, T. Sato, A. Sugiya ma and S. Miyagawa (2004). "Significance of macrophage chemoattractant protein-1 expr ession and macrophage infiltration in squamous cell carcinoma of the esophagus." Am J Gastroenterol 99 (9): 1667-74. Kong, D., L. G. Melo, M. Gnecchi, L. Zhang, G. Mostoslavsky, C. C. Liew, R. E. Pratt and V. J. Dzau (2004). "Cytokine-induced mobilization of circulating endothelial progenitor cells enhances repair of injured arteries." Circulation 110 (14): 203946. Kopp, H. G., S. T. Avecilla, A. T. Hooper, S. V. Shmelkov, C. A. Ramos, F. Zhang and S. Rafii (2005). "Tie2 activation contribu tes to hemangiogenic regeneration after myelosuppression." Blood 106 (2): 505-513.

PAGE 121

110 Krebs, L. T., Y. Xue, C. R. Norton, J. R. Shutter, M. Maguire, J. P. Sundberg, D. Gallahan, V. Closson, J. Kitajewski, R. Calla han, G. H. Smith, K. L. Stark and T. Gridley (2000). "Notch signaling is e ssential for vascular morphogenesis in mice." Genes Dev 14 (11): 1343-52. Krieg, M., R. Haas, H. Brauch, T. Acker, I. Flamme and K. H. Plate (2000). "Upregulation of hypoxia-inducible factors HIF-1alpha and HIF-2alpha under normoxic conditions in renal carcinoma cells by von Hippel-Lindau tumor suppressor gene loss of function." Oncogene 19 (48): 5435-43. Kryczek, I., A. Lange, P. Mottram, X. Alvar ez, P. Cheng, M. Hogan, L. Moons, S. Wei, L. Zou, V. Machelon, D. Emilie, M. Terrassa, A. Lackner, T. J. Curiel, P. Carmeliet and W. Zou (2005). "CXCL12 and vascular endothelial growth factor synergistically induce neoangiogenesis in human ovarian cancers." Cancer Res 65 (2): 465-72. Kurre, P., P. Anandakumar, M. A. Harkey, B. Thomasson and H. P. Kiem (2004). "Efficient marking of murine long-te rm repopulating stem cells targeting unseparated marrow cells at low lentiviral vector particle concentration." Mol Ther 9 (6): 914-22. Kurtz, A., A. Aigner, R. H. Cabal-Manzano, R. E. Butler, D. R. Hood, R. B. Sessions, F. Czubayko and A. Wellstein (2004). "Differen tial regulation of a fibroblast growth factor-binding protein during skin carc inogenesis and wound healing." Neoplasia 6 (5): 595-602. Kuwana, M., Y. Okazaki, H. Kodama, K. Izumi, H. Yasuoka, Y. Ogawa, Y. Kawakami and Y. Ikeda (2003). "Human circulating CD14+ monocytes as a source of progenitors that exhibit mesenchyma l cell differentiation." J Leukoc Biol 74 (5): 833-45. Kwabi-Addo, B., M. Ozen and M. Ittmann (2004). "The role of fibrobl ast growth factors and their receptors in prosta te cancer." Endocr Relat Cancer 11 (4): 709-24. Labastie, M. C., F. Cortes, P. H. Romeo, C. Dulac and B. Peault (1998). "Molecular identity of hematopoietic precursor cel ls emerging in the human embryo." Blood 92 (10): 3624-35. Lacaud, G., L. Gore, M. Kennedy, V. Kouskoff, P. Kingsley, C. Hogan, L. Carlsson, N. Speck, J. Palis and G. Keller (2002). "Runx1 is essential for hematopoietic commitment at the hemangioblast stage of development in vitro." Blood 100 (2): 458-66. Laufs, S., B. Gentner, K. Z. Nagy, A. Jauc h, A. Benner, S. Naundorf, K. Kuehlcke, B. Schiedlmeier, A. D. Ho, W. J. Zeller a nd S. Fruehauf (2003). "Retroviral vector integration occurs in preferred ge nomic targets of human bone marrowrepopulating cells." Blood 101 (6): 2191-8.

PAGE 122

111 Laufs, S., K. Z. Nagy, F. A. Giordano, A. Hotz -Wagenblatt, W. J. Zeller and S. Fruehauf (2004a). "Insertion of retroviral ve ctors in NOD/SCID repopulating human peripheral blood progenitor cells occurs preferentially in the vicinity of transcription start regions and in introns." Mol Ther 10 (5): 874-81. Laufs, U., N. Werner, A. Link, M. Endres, S. Wassmann, K. Jurgens, E. Miche, M. Bohm and G. Nickenig (2004b). "Physical tr aining increases endo thelial progenitor cells, inhibits neointima formation, a nd enhances angiogenesis." Circulation 109 (2): 220-6. Laukkanen, M. O., K. Kuramoto, B. Calmels, M. Takatoku, C. von Kalle, R. E. Donahue and C. E. Dunbar (2005). "Low-dose total body irradiation causes clonal fluctuation of primate hematopoietic stem and progenitor cells." Blood 105 (3): 1010-5. Lemischka, I. R., D. H. Raulet and R. C. Mulligan (1986). "Developmental potential and dynamic behavior of hemat opoietic stem cells." Cell 45 (6): 917-27. Li, W., M. J. Ferkowicz, S. A. Johnson, W. C. Shelley and M. C. Yoder (2005). "Endothelial cells in the early murine yolk sac give rise to CD41-expressing hematopoietic cells." Stem Cells Dev 14 (1): 44-54. Li, W., S. A. Johnson, W. C. Shelley and M. C. Yoder (2004). "Hematopoietic stem cell repopulating ability can be maintained in vitro by some primary endothelial cells." Exp Hematol 32 (12): 1226-37. Lim, F. T., H. H. Kanhai and J. H. Falk enburg (2005). "Character ization of the human CD34+ hematopoietic progenitor cell comp artment during the second trimester of pregnancy." Haematologica 90 (2): 173-9. Lindahl, P., B. R. Johansson, P. Leveen a nd C. Betsholtz (1997). "Pericyte loss and microaneurysm formation in PDGF-B-deficient mice." Science 277 (5323): 242-5. Liu, C., Z. Sun, X. Du, X. Chen, J. Feng and B. Jia (2005). "Implantation of Endothelial Progenitor Cells into Laser-Induced Channels in Rat Ischemia Hindlimb Augments Neovascularization." Ann Vasc Surg . Llevadot, J., S. Murasawa, Y. Kureishi, S. Uchida, H. Masuda, A. Kawamoto, K. Walsh, J. M. Isner and T. Asahara (2001). "H MG-CoA reductase inhibitor mobilizes bone marrow--derived endothelial progenitor cells." J Clin Invest 108 (3): 399405. Loges, S., B. Fehse, M. A. Brockmann, K. Lamszus, M. Butzal, M. Guckenbiehl, G. Schuch, S. Ergun, U. Fischer, A. R. Zander, D. K. Hossfeld, W. Fiedler and U. M. Gehling (2004). "Identifi cation of the adult human hemangioblast." Stem Cells Dev 13 (3): 229-42.

PAGE 123

112 Lyden, D., K. Hattori, S. Dias, C. Costa, P. Blaikie, L. Butros, A. Chadburn, B. Heissig, W. Marks, L. Witte, Y. Wu, D. Hicklin, Z. Zhu, N. R. Hackett, R. G. Crystal, M. A. Moore, K. A. Hajjar, K. Manova, R. Benezra and S. Rafii (2001). "Impaired recruitment of bone-marrow-derived endot helial and hematopoietic precursor cells blocks tumor angiogene sis and growth." Nat Med 7 (11): 1194-201. Maniotis, A. J., R. Folberg, A. Hess, E. A. Seftor, L. M. Gardner, J. Pe'er, J. M. Trent, P. S. Meltzer and M. J. Hendrix (1999). "Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry." Am J Pathol 155 (3): 739-52. Marx, J. (2003). "Cancer research. Mutant stem cells may seed cancer." Science 301 (5638): 1308-10. Massa, M., V. Rosti, M. Ferrario, R. Campanelli, I. Ramajoli, R. Rosso, G. M. De Ferrari, M. Ferlini, L. Goffredo, A. Bertol etti, C. Klersy, A. Pecci, R. Moratti and L. Tavazzi (2005). "Increased circul ating hematopoietic and endothelial progenitor cells in the early phase of acute myocardial infarction." Blood 105 (1): 199-206. Mole, D. R., P. H. Maxwell, C. W. Pugh a nd P. J. Ratcliffe (2001). "Regulation of HIF by the von Hippel-Lindau tumour suppre ssor: implications for cellular oxygen sensing." IUBMB Life 52 (1-2): 43-7. Montfort, M. J., C. R. Olivares, J. M. Mu lcahy and W. H. Fleming (2002). "Adult blood vessels restore host hematopoiesis follow ing lethal irradiation." Exp Hematol 30 (8): 950-6. Moore, M. A., K. Hattori, B. Heissig, J. H. Shieh, S. Dias, R. G. Crystal and S. Rafii (2001). "Mobilization of endot helial and hematopoietic stem and progenitor cells by adenovector-mediated elevation of serum levels of SDF-1, VEGF, and angiopoietin-1." Ann N Y Acad Sci 938 : 36-45; discussion 45-7. Moore, M. A. and D. Metcalf (1970). "Ont ogeny of the haemopoietic system: yolk sac origin of in vivo and in vitro colony forming cells in the developing mouse embryo." Br J Haematol 18 (3): 279-96. Motoike, T., D. W. Markham, J. Rossant and T. N. Sato (2003). "Evidence for novel fate of Flk1+ progenitor: contributi on to muscle lineage." Genesis 35 (3): 153-9. Muller, A. M., A. Medvinsky, J. Strouboulis, F. Grosveld and E. Dzierzak (1994). "Development of hematopoietic stem cell activity in the mouse embryo." Immunity 1 (4): 291-301. Murayama, T., O. M. Tepper, M. Silver, H. Ma, D. W. Losordo, J. M. Isner, T. Asahara and C. Kalka (2002). "Determination of bone marrow-derived endothelial progenitor cell significance in angi ogenic growth factor-induced neovascularization in vivo." Exp Hematol 30 (8): 967-72.

PAGE 124

113 Murray, P. D. F. (1932). "The development in vitro of the blood of the early chick embryo." Proc. Roy. Soc. London 11 : 497-521. Nakatani, K., S. Takeshita, H. Tsujimoto, Y. Kawamura, T. Tokutomi and I. Sekine (2003). "Circulating endothelial cells in Kawasaki disease." Clin Exp Immunol 131 (3): 536-40. Nakauchi, H., K. Sudo and H. Ema (2001). "Q uantitative assessment of the stem cell self-renewal capacity." Ann N Y Acad Sci 938 : 18-24; discussion 24-5. Nesbit, M., H. Schaider, T. H. Miller and M. Herlyn (2001). "Low-level monocyte chemoattractant protein-1 stimulation of monocytes leads to tumor formation in nontumorigenic melanoma cells." J Immunol 166 (11): 6483-90. Nishikawa, S. I., S. Fraser, T. Fujimoto, M. Endoh, S. Nishikawa and M. Ogawa (2000). "All B cells are progeny of endothelial cells: a new perspective." Immunol Rev 175 : 112-9. Nishikawa, S. I., S. Nishikawa, M. Hirash ima, N. Matsuyoshi and H. Kodama (1998a). "Progressive lineage analysis by ce ll sorting and culture identifies FLK1+VE-cadherin+ cells at a divergi ng point of endothelial and hemopoietic lineages." Development 125 (9): 1747-57. Nishikawa, S. I., S. Nishikawa, H. Kawamoto, H. Yoshida, M. Kizumoto, H. Kataoka and Y. Katsura (1998b). "In vitro genera tion of lymphohematopoietic cells from endothelial cells purified from murine embryos." Immunity 8 (6): 761-9. Nolta, J. A., M. A. Dao, S. Wells, E. M. Smogorzewska and D. B. Kohn (1996). "Transduction of pluripotent human hema topoietic stem cells demonstrated by clonal analysis after engraftment in imm une-deficient mice." Proc Natl Acad Sci U S A 93 (6): 2414-9. North, T. E., M. F. de Bruijn, T. Stacy, L. Talebian, E. Lind, C. Robin, M. Binder, E. Dzierzak and N. A. Speck (2002) . "Runx1 expression marks long-term repopulating hematopoietic stem cells in the midgestation mouse embryo." Immunity 16 (5): 661-72. North, T. E., T. Stacy, C. J. Matheny, N. A. Speck and M. F. de Bruijn (2004). "Runx1 is expressed in adult mouse hematopoietic stem cells and differentiating myeloid and lymphoid cells, but not in matu ring erythroid cells." Stem Cells 22 (2): 158-68. Oberlin, E., M. Tavian, I. Blazsek and B. Peault (2002). "Blood-fo rming potential of vascular endothelium in the human embryo." Development 129 (17): 4147-57.

PAGE 125

114 Ohneda, O., C. Fennie, Z. Zheng, C. Donahue, H. La, R. Villacorta, B. Cairns and L. A. Lasky (1998). "Hematopoietic stem ce ll maintenance and differentiation are supported by embryonic aorta-gonad-mesone phros region-derived endothelium." Blood 92 (3): 908-19. Ohta, M., Y. Kitadai, S. Tanaka, M. Yoshihar a, W. Yasui, N. Mukaida, K. Haruma and K. Chayama (2003). "Monocyte chemoattracta nt protein-1 expr ession correlates with macrophage infiltration and tumor vasc ularity in human ga stric carcinomas." Int J Oncol 22 (4): 773-8. Okuda, T., J. van Deursen, S. W. Hiebert, G. Grosveld and J. R. Downing (1996). "AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis." Cell 84 (2): 321-30. Ottersbach, K. and E. Dzierzak (2005). "T he Murine Placenta Contains Hematopoietic Stem Cells within the Vascular Labyrinth Region." Dev Cell 8 (3): 377-87. Ozerdem, U. and W. B. Stallcup (2003). "Ear ly contribution of pe ricytes to angiogenic sprouting and tube formation." Angiogenesis 6 (3): 241-9. Palis, J., S. Robertson, M. Kennedy, C. Wa ll and G. Keller (1999). "Development of erythroid and myeloid progenitors in the yolk sac and embryo proper of the mouse." Development 126 (22): 5073-84. Palis, J. and M. C. Yoder (2001). "Yolk-sac hematopoiesis: the first blood cells of mouse and man." Exp Hematol 29 (8): 927-36. Pardanaud, L. and F. Dieterlen-Li evre (1999). "Mani pulation of the angiopoietic/hemangiopoietic commitment in the avian embryo." Development 126 (4): 617-27. Pardanaud, L. and F. Dieterlen-Lievre (2000) . "Ontogeny of the endot helial system in the avian model." Adv Exp Med Biol 476 : 67-78. Pardanaud, L., D. Luton, M. Prigent, L. M. B ourcheix, M. Catala and F. Dieterlen-Lievre (1996). "Two distinct endothe lial lineages in ontogeny, one of them related to hemopoiesis." Development 122 (5): 1363-71. Pardanaud, L., F. Yassine and F. Dieter len-Lievre (1989). "Relationship between vasculogenesis, angiogenesis and haemopoiesis during avian ontogeny." Development 105 (3): 473-85. Pear, W. S., G. P. Nolan, M. L. Scott and D. Baltimore (1993). "Production of high-titer helper-free retroviruses by transient transfection." Proc Natl Acad Sci U S A 90 (18): 8392-6. Peault, B. and M. Tavian (2003). "Hematopoi etic stem cell emergence in the human embryo and fetus." Ann N Y Acad Sci 996 : 132-40.

PAGE 126

115 Peichev, M., A. J. Naiyer, D. Pereira, Z. Zhu, W. J. Lane, M. Williams, M. C. Oz, D. J. Hicklin, L. Witte, M. A. Moore and S. Rafii (2000). "Expression of VEGFR-2 and AC133 by circulating human CD34(+) cells identifies a population of functional endothelial precursors." Blood 95 (3): 952-8. Pelosi, E., M. Valtieri, S. Coppola, R. Botta , M. Gabbianelli, V. Lulli, G. Marziali, B. Masella, R. Muller, C. Sgadari, U. Te sta, G. Bonanno and C. Peschle (2002). "Identification of the hemangi oblast in postnatal life." Blood 100 (9): 3203-8. Peters, B. A., L. A. Diaz, K. Polyak, L. Mesz ler, K. Romans, E. C. Guinan, J. H. Antin, D. Myerson, S. R. Hamilton, B. Vogels tein, K. W. Kinzler and C. Lengauer (2005). "Contribution of bone marrow-deri ved endothelial cells to human tumor vasculature." Nat Med 11 (3): 261-262. Platten, M., A. Kretz, U. Naumann, S. Au lwurm, K. Egashira, S. Isenmann and M. Weller (2003). "Monocyte chemoattractan t protein-1 increases microglial infiltration and aggressivene ss of gliomas." Ann Neurol 54 (3): 388-92. Poole, T. J., E. B. Finkelstein and C. M. Cox (2001). "The role of FGF and VEGF in angioblast induction and migration duri ng vascular development." Dev Dyn 220 (1): 1-17. Powell, T. M., J. D. Paul, J. M. Hill, M. Thompson, M. Benjamin, M. Rodrigo, J. P. McCoy, E. J. Read, H. M. Khuu, S. F. Leitman, T. Finkel and R. O. Cannon, 3rd (2005). "Granulocyte colony-stimulating f actor mobilizes func tional endothelial progenitor cells in patients with coronary artery disease." Arterioscler Thromb Vasc Biol 25 (2): 296-301. Rafii, S., S. Avecilla, S. Shmelkov, K. Shido, R. Tejada, M. A. Moor e, B. Heissig and K. Hattori (2003). "Angiogenic factors reconstitute hemat opoiesis by recruiting stem cells from bone marrow microe nvironment." Ann N Y Acad Sci 996 : 49-60. Rafii, S. and D. Lyden (2003). "Therapeutic stem and progenitor cell transplantation for organ vascularization an d regeneration." Nat Med 9 (6): 702-12. Rehman, J., J. Li, L. Parvathaneni, G. Ka rlsson, V. R. Panchal, C. J. Temm, J. Mahenthiran and K. L. March (2004). "E xercise acutely incr eases circulating endothelial progenitor cells and mono cyte-/macrophage-derived angiogenic cells." J Am Coll Cardiol 43 (12): 2314-8. Reyes, M., A. Dudek, B. Jahagirdar, L. Koodie, P. H. Marker and C. M. Verfaillie (2002). "Origin of endothelial progenitors in human postnatal bone marrow." J Clin Invest 109 (3): 337-46. Rill, D. R., V. M. Santana, W. M. Roberts, T. Nilson, L. C. Bowman, R. A. Krance, H. E. Heslop, R. C. Moen, J. N. Ihle and M. K. Brenner (1994). "Direct demonstration that autologous bone marrow transplant ation for solid tumors can return a multiplicity of tumorigenic cells." Blood 84 (2): 380-3.

PAGE 127

116 Risau, W. (1997). "Mechanisms of angiogenesis." Nature 386 (6626): 671-4. Roberts, W. G., P. M. Whalen, E. Soderstrom, G. Moraski, J. P. Lyssikatos, H. F. Wang, B. Cooper, D. A. Baker, D. Savage, D. Dalvie, J. A. Atherton, S. Ralston, R. Szewc, J. C. Kath, J. Lin, C. Soderstrom, G. Tkalcevic, B. D. Cohen, V. Pollack, W. Barth, W. Hungerford and E. Ung (2005). "Antiangiogenic and antitumor activity of a selective PDGFR tyrosine ki nase inhibitor, CP-673,451." Cancer Res 65 (3): 957-66. Robertson, S., M. Kennedy and G. Keller (1999). "Hematopoietic commitment during embryogenesis." Ann N Y Acad Sci 872 : 9-15; discussion 15-6. Rumpold, H., D. Wolf, R. Koeck and E. G unsilius (2004). "Endothelial progenitor cells: a source for therapeutic vasculogenesis?" J Cell Mol Med 8 (4): 509-18. Russel, E. (1979). "Hereditary anemias of th e mouse: a review for geneticists." Adv Genet 20 : 357-459. Sabin, F. R. (1920). "Studies on the origin of blood-vessels and of red-blood corpuscles as seen in the living blas toderm of chicks during the second day of incubation." Contributions to Embryology 9 : 215-262. Salcedo, R., M. L. Ponce, H. A. Young, K. Wa sserman, J. M. Ward, H. K. Kleinman, J. J. Oppenheim and W. J. Murphy (2000). "H uman endothelial cells express CCR2 and respond to MCP-1: direct role of MCP-1 in angiogenesis and tumor progression." Blood 96 (1): 34-40. Salmaggi, A., M. Gelati, B. Pollo, S. Frigerio, M. Eoli, A. Silvani, G. Broggi, E. Ciusani, D. Croci, A. Boiardi and M. De Ro ssi (2004). "CXCL12 in malignant glial tumors: a possible role in angiogenesis and cross-talk betw een endothelial and tumoral cells." J Neurooncol 67 (3): 305-17. Sanchez, M. J., A. Holmes, C. Miles and E. Dzierzak (1996). "Cha racterization of the first definitive hematopoietic stem ce lls in the AGM and liver of the mouse embryo." Immunity 5 (6): 513-25. Sasaki, K. and H. Iwatsuki (1997). "Origi n and fate of the central macrophages of erythroblastic islands in the fetal a nd neonatal mouse liver ." Microsc Res Tech 39 (5): 398-405. Sato, A., A. Iwama, N. Takakura, H. Nish io, G. D. Yancopoulos and T. Suda (1998). "Characterization of TEK receptor tyrosine kinase and its ligands, Angiopoietins, in human hematopoietic progenitor cells." Int Immunol 10 (8): 1217-27. Sato, T. N., Y. Tozawa, U. Deutsch, K. Wolburg-Buchholz, Y. Fujiwara, M. GendronMaguire, T. Gridley, H. Wol burg, W. Risau and Y. Qin (1995). "Distinct roles of the receptor tyrosine kinases Tie-1 and Tie-2 in blood vessel formation." Nature 376 (6535): 70-4.

PAGE 128

117 Schmeisser, A., C. D. Garlichs, H. Zhang, S. Eskafi, C. Graffy, J. Ludwig, R. H. Strasser and W. G. Daniel (2001). "Monocyt es coexpress endothelial and macrophagocytic lineage markers and form cord-like structures in Matrigel under angiogenic conditions." Cardiovasc Res 49 (3): 671-80. Schmidt, M., G. Hoffmann, M. Wissler, N. Lemke, A. Mussig, H. Glimm, D. A. Williams, S. Ragg, C. U. Hesemann and C. von Kalle (2001). "Detection and direct genomic sequencing of multipl e rare unknown flanking DNA in highly complex samples." Hum Gene Ther 12 (7): 743-9. Schmidt, M., P. Zickler, G. Hoffmann, S. Haas , M. Wissler, A. Muessig, J. F. Tisdale, K. Kuramoto, R. G. Andrews, T. Wu, H. P. Kiem, C. E. Dunbar and C. von Kalle (2002). "Polyclonal long-ter m repopulating stem cell cl ones in a primate model." Blood 100 (8): 2737-43. Shalaby, F., J. Ho, W. L. Stanford, K. D. Fisc her, A. C. Schuh, L. Schwartz, A. Bernstein and J. Rossant (1997). "A requirement for Flk1 in primitive and definitive hematopoiesis and vasculogenesis." Cell 89 (6): 981-90. Shalaby, F., J. Rossant, T. P. Yamaguchi, M. Gertsenstein, X. F. Wu, M. L. Breitman and A. C. Schuh (1995). "Failure of blood-is land formation and vasculogenesis in Flk1-deficient mice." Nature 376 (6535): 62-6. Shaw, J. P., R. Basch and P. Shamamian (2004). "Hematopoietic stem cells and endothelial cell precursor s express Tie-2, CD31 and CD45." Blood Cells Mol Dis 32 (1): 168-75. Sherwood, R. I., J. L. Christensen, I. L. Weissman and A. J. Wagers (2004). "Determinants of skeletal muscle contri butions from circulating cells, bone marrow cells, and hematopoietic stem cells." Stem Cells 22 (7): 1292-304. Shi, Q., M. H. Wu, N. Hayashida, A. R. Wechezak, A. W. Clowes and L. R. Sauvage (1994). "Proof of fallout e ndothelialization of impervi ous Dacron grafts in the aorta and inferior vena cava of the dog." J Vasc Surg 20 (4): 546-56; discussion 556-7. Shintani, S., T. Murohara, H. Ikeda, T. Ueno, T. Honma, A. Katoh, K. Sasaki, T. Shimada, Y. Oike and T. Imaizumi (2001). "Mobilization of endothelial progenitor cells in patients with acut e myocardial infarction." Circulation 103 (23): 2776-9. Shivdasani, R. A., E. L. Mayer and S. H. Orkin (1995). "Absence of blood formation in mice lacking the T-cell leukaemia oncoprotein tal-1/SCL." Nature 373 (6513): 432-4. Shizuru, J. A., R. S. Negrin and I. L. Weissman (2005). "Hematopoietic stem and progenitor cells: clinical and preclinical regeneration of the hematolymphoid system." Annu Rev Med 56 : 509-38.

PAGE 129

118 Shmelkov, S. V., R. St Clair, D. Lyden and S. Rafii (2005). "AC133/CD133/Prominin-1." Int J Biochem Cell Biol 37 (4): 715-9. Sims, D. E. (2000). "Diversi ty within pericy tes." Clin Exp Pharmacol Physiol 27 (10): 842-6. Singer, R. S., W. O. Johnson, J. S. Jeffrey, R. P. Chin, T. E. Carpenter, E. R. Atwill and D. C. Hirsh (2000). "A statistical model for assessing sample size for bacterial colony selection: a case study of Escheric hia coli and avian cellulitis." J Vet Diagn Invest 12 (2): 118-25. Sitnicka, E., N. Lin, G. V. Priestley, N. Fox, V. C. Broudy, N. S. Wolf and K. Kaushansky (1996). "The effect of th rombopoietin on the proliferation and differentiation of murine hema topoietic stem cells." Blood 87 (12): 4998-5005. Smith, L. G., I. L. Weissman and S. Heimfe ld (1991). "Clonal analysis of hematopoietic stem-cell differentiation in vi vo." Proc Natl Acad Sci U S A 88 (7): 2788-92. Stiver, S. I. (2004). "Angiogenesis and its role in the behavior of astrocytic brain tumors." Front Biosci 9 : 3105-23. Stoll, B. R., C. Migliorini, A. Kadambi, L. L. Munn and R. K. Jain (2003). "A mathematical model of the contributi on of endothelial progenitor cells to angiogenesis in tumors: implications for antiangiogenic therapy." Blood 102 (7): 2555-61. Strehlow, K., N. Werner, J. Berweiler, A. Link, U. Dirnagl, J. Priller, K. Laufs, L. Ghaeni, M. Milosevic, M. Bohm and G. Nickenig (2003). "Estrogen increases bone marrow-derived endothelial proge nitor cell production and diminishes neointima formation." Circulation 107 (24): 3059-65. Suri, C., P. F. Jones, S. Patan, S. Bartunkova, P. C. Maisonpierre, S. Davis, T. N. Sato and G. D. Yancopoulos (1996). "Requisite role of angiopoietin -1, a ligand for the TIE2 receptor, during embryonic angiogenesis." Cell 87 (7): 1171-80. Suri, C., J. McClain, G. Thurston, D. M. McDonald, H. Zhou, E. H. Oldmixon, T. N. Sato and G. D. Yancopoulos (1998). "I ncreased vascularization in mice overexpressing angiopoietin-1." Science 282 (5388): 468-71. Swindle, C. S., H. G. Kim and C. A. Klug ( 2004). "Mutation of CpGs in the murine stem cell virus retroviral vector long terminal repeat represses silencing in embryonic stem cells." J Biol Chem 279 (1): 34-41. Takahashi, T., C. Kalka, H. Masuda, D. Che n, M. Silver, M. Kearne y, M. Magner, J. M. Isner and T. Asahara (1999). "Ischemiaand cytokine-induced mobilization of bone marrow-derived endothelial progenito r cells for neovascularization." Nat Med 5 (4): 434-8.

PAGE 130

119 Tavian, M., C. Robin, L. Coulombel and B. Peault (2001). "The human embryo, but not its yolk sac, generates lympho-myel oid stem cells: mapping multipotent hematopoietic cell fate in intraembryonic mesoderm." Immunity 15 (3): 487-95. Tepper, O. M., R. D. Galiano, J. M. Capla, C. Kalka, P. J. Gagne, G. R. Jacobowitz, J. P. Levine and G. C. Gurtner (2002). "Human endothelial progenitor cells from type II diabetics exhibit impair ed proliferation, adhesion, and incorporation into vascular structures." Circulation 106 (22): 2781-6. Toles, J. F., D. H. Chui, L. W. Belbeck, E. Starr and J. E. Barker (1989). "Hemopoietic stem cells in murine embryonic yolk s ac and peripheral blood." Proc Natl Acad Sci U S A 86 (19): 7456-9. Tong, Y., X. Zhang, F. Tian, Y. Yi, Q. Xu, L. Li, L. Tong, L. Lin and J. Ding (2005). "Philinopside A, a novel marine-der ived compound possessing dual antiangiogenic and anti-tumor effects." Int J Cancer 114 (6): 843-53. Ueno, T., M. Toi, H. Saji, M. Muta, H. Bando, K. Kuroi, M. Koike, H. Inadera and K. Matsushima (2000). "Significance of m acrophage chemoattractant protein-1 in macrophage recruitment, angiogenesis, and survival in human breast cancer." Clin Cancer Res 6 (8): 3282-9. Urbich, C., E. Dernbach, A. M. Zeiher a nd S. Dimmeler (2002). "Double-edged role of statins in angiogenesi s signaling." Circ Res 90 (6): 737-44. Urbich, C. and S. Dimmeler (2005) . "Risk factors for coronary artery disease, circulating endothelial progenitor cells, and the role of HMG-CoA reducta se inhibitors." Kidney Int 67 (5): 1672-6. Van Beusechem, V. W. and D. Valerio (1996). "Gene transfer into hematopoietic stem cells of nonhuman primates." Hum Gene Ther 7 (14): 1649-68. Van der Auwera, I., S. J. Van Laere, G. G. Van den Eynden, I. Benoy, P. van Dam, C. G. Colpaert, S. B. Fox, H. Turley, A. L. Harris, E. A. Van Marck, P. B. Vermeulen and L. Y. Dirix (2004). "Increased angiogenesis and lymphangiogenesis in inflammatory versus noninflammatory breast cancer by real-time reverse transcriptase-PCR gene expression quantification." Clin Cancer Res 10 (23): 7965-71. Vasa, M., S. Fichtlscherer, A. Aicher, K. Adle r, C. Urbich, H. Mar tin, A. M. Zeiher and S. Dimmeler (2001). "Number and migrator y activity of circul ating endothelial progenitor cells inversely correlate with ri sk factors for coronary artery disease." Circ Res 89 (1): E1-7.

PAGE 131

120 Verma, S., M. A. Kuliszewski, S. H. Li, P. E. Szmitko, L. Zucco, C. H. Wang, M. V. Badiwala, D. A. Mickle, R. D. Weisel, P. W. Fedak, D. J. Stewart and M. J. Kutryk (2004). "C-reactive protein attenuate s endothelial progenitor cell survival, differentiation, and function: further ev idence of a mechanistic link between C-reactive protein and cardiova scular disease." Circulation 109 (17): 2058-67. Vokes, S. A. and P. A. Krieg (2002). "Endoderm is required for vascular endothelial tube formation, but not for angioblas t specification." Development 129 (3): 775-85. Vokes, S. A., T. A. Yatskievych, R. L. Heimark, J. McMahon, A. P. McMahon, P. B. Antin and P. A. Krieg (2004). "Hedgehog signaling is essential for endothelial tube formation during vasculogenesis." Development 131 (17): 4371-80. Wadman, I. A., H. Osada, G. G. Grutz, A. D. Agulnick, H. Westpha l, A. Forster and T. H. Rabbitts (1997). "The LIM-only prot ein Lmo2 is a bridging molecule assembling an erythroid, DNA-binding co mplex which includes the TAL1, E47, GATA-1 and Ldb1/NLI proteins." Embo J 16 (11): 3145-57. Walter, D. H., K. Rittig, F. H. Bahlmann, R. Kirchmair, M. Silver, T. Murayama, H. Nishimura, D. W. Losordo, T. Asahara and J. M. Isner (2002). "Statin therapy accelerates reendothelialization: a nove l effect involving mobilization and incorporation of bone marro w-derived endothelial proge nitor cells." Circulation 105 (25): 3017-24. Wang, C. H., N. Ciliberti, S. H. Li, P. E. Szmitko, R. D. Weisel, P. W. Fedak, M. Al-Omran, W. J. Cherng, R. K. Li, W. L. Stanford and S. Verma (2004). "Rosiglitazone facilitate s angiogenic progenitor ce ll differentiation toward endothelial lineage: a new paradigm in glitazone pleiot ropy." Circulation 109 (11): 1392-400. Wang, H. U., Z. F. Chen and D. J. A nderson (1998). "Molecu lar distinction and angiogenic interaction between embryoni c arteries and veins revealed by ephrin-B2 and its receptor Eph-B4." Cell 93 (5): 741-53. Warren, A. J., W. H. Colledge, M. B. Carlt on, M. J. Evans, A. J. Smith and T. H. Rabbitts (1994). "The oncogenic cysteine-rich LIM domain protein rbtn2 is essential for erythroid development." Cell 78 (1): 45-57. Wei, J., S. Blum, M. Unger, G. Jarmy, M. La mparter, A. Geishauser , G. A. Vlastos, G. Chan, K. D. Fischer, D. Rattat, K. M. Debatin, A. K. Hatzopoulos and C. Beltinger (2004). "Embryonic endothelial pr ogenitor cells arme d with a suicide gene target hypoxic lung metastases af ter intravenous deliv ery." Cancer Cell 5 (5): 477-88. Weissman, I. L., V. E. Papaioannou and R. L. Gardner (1978). Fetal hematopoietic origins of the adult hematolymphoid system. Differentiation of normal and neoplastic hematopoietic cells . B. Clarkson, P. A. Marks and J. E. Till. Plainview, New York, Cold Spring s Harbor Laboratory : p.33-47.

PAGE 132

121 Williams, D. A., I. R. Lemischka, D. G. Nath an and R. C. Mulligan (1984). "Introduction of new genetic material into pluripotent haematopoietic stem cel ls of the mouse." Nature 310 (5977): 476-80. Winkler, F., S. V. Kozin, R. T. Tong, S. S. Chae, M. F. Booth, I. Garkavtsev, L. Xu, D. J. Hicklin, D. Fukumura, E. di Tomaso, L. L. Munn and R. K. Jain (2004). "Kinetics of vascular normalization by VEGFR2 bl ockade governs brain tumor response to radiation: role of oxygenation, angiopoiet in-1, and matrix metalloproteinases." Cancer Cell 6 (6): 553-63. Wolber, F. M., E. Leonard, S. Michael, C. M. Orschell-Traycoff, M. C. Yoder and E. F. Srour (2002). "Roles of spleen and liver in development of the murine hematopoietic system." Exp Hematol 30 (9): 1010-9. Wu, M. H., Q. Shi, A. R. Wechezak, A. W. Clowes, I. L. Gordon and L. R. Sauvage (1995). "Definitive proof of e ndothelialization of a Dacron arterial prosthesis in a human being." J Vasc Surg 21 (5): 862-7. Wu, X., Y. Li, B. Crise and S. M. Burgess (2003). "Transcription start regions in the human genome are favored targets for MLV integration." Science 300 (5626): 1749-51. Yamaguchi, J., K. F. Kusano, O. Masuo, A. Kawamoto, M. Silver, S. Murasawa, M. Bosch-Marce, H. Masuda, D. W. Losordo, J. M. Isner and T. Asahara (2003). "Stromal cell-derived factor-1 effects on ex vivo expanded endothelial progenitor cell recruitment for ischemic ne ovascularization." Circulation 107 (9): 1322-8. Yamashita, J., H. Itoh, M. Hirashima, M. Ogaw a, S. Nishikawa, T. Yurugi, M. Naito, K. Nakao and S. Nishikawa (2000). "Flk1-pos itive cells derived from embryonic stem cells serve as vascular progenitors." Nature 408 (6808): 92-6. Yoder, M. C. and K. Hiatt (1997). "Engraf tment of embryonic hematopoietic cells in conditioned newborn re cipients." Blood 89 (6): 2176-83. Yoder, M. C., K. Hiatt, P. Dutt, P. Mukherjee, D. M. Bodine and D. Orlic (1997). "Characterization of definitive lymphohe matopoietic stem cells in the day 9 murine yolk sac." Immunity 7 (3): 335-44. Zambidis, E. T., B. Peault, T. S. Park, F. Bunz and C. I. Civin (2005). "Hematopoietic differentiation of human embryonic stem cells progresses through sequential hemato-endothelial, primitive, and definitive stages resembling human yolk sac development." Blood . Zhang, C. C. and H. F. Lodish (2005). "Mur ine hematopoietic stem cells change their surface phenotype during ex vivo expansion." Blood .

PAGE 133

122 Zhang, J., C. Niu, L. Ye, H. Huang, X. He, W. G. Tong, J. Ross, J. Haug, T. Johnson, J. Q. Feng, S. Harris, L. M. Wiedem ann, Y. Mishina and L. Li (2003). "Identification of the haematopoietic st em cell niche and control of the niche size." Nature 425 (6960): 836-41. Zhang, X. Q., N. Takakura, Y. Oike, T. Inad a, N. W. Gale, G. D. Yancopoulos and T. Suda (2001). "Stromal cells expressi ng ephrin-B2 promote the growth and sprouting of ephrin-B2(+) endothelial cells." Blood 98 (4): 1028-37. Zhong, J. F., Y. Zhao, S. Sutton, A. Su, Y. Zhan, L. Zhu, C. Yan, T. Gallaher, P. B. Johnston, W. F. Anderson and M. P. C ooke (2005). "Gene expression profile of murine long-term reconstituting vs. shortterm reconstituting hematopoietic stem cells." Proc Natl Acad Sci U S A 102 (7): 2448-53. Ziegler, B. L., M. Valtieri, G. A. Porada, R. De Maria, R. Muller, B. Masella, M. Gabbianelli, I. Casella, E. Pelosi, T. Bo ck, E. D. Zanjani and C. Peschle (1999). "KDR receptor: a key marker defining hematopoietic stem cells." Science 285 (5433): 1553-8.

PAGE 134

123 BIOGRAPHICAL SKETCH Christopher Lawrence Bray was bor n on October 29, 1976 in Stamford, Connecticut. He graduated valedictorian from University High School in Orlando, Florida in 1994. He attended Florida State Univ ersity in Tallahassee, Florida where he received 4 state scholarships and 2 nati onal scholarships; and was inducted into 7 national honor societies. He graduated summa cum laude with honors in 1998 with majors in biology, biochemistry, computer science, and mathematics, and received the ranking of Top Male Senior in Academics. He then entered the Medical Scientist Training Program, a dual-degree program lead ing to the MD and PhD degrees, at the University of Florida. During hi s medical training, he performe d a clinical trial related to stimulant abuse in sleep-deprived students that resulted in pu blication. During his graduate training, he has been published for hi s contribution to eluc idating the role of PU.1 in eyrthropoiesis. Publica tions that are currently in submission include his review on hemangioblasts and his manuscript detaili ng the results of his work on clonally identifying the hemangioblastic origins of th e endothelial progenitor cell. He has also presented this work at two internationa l conferences: Keystone Symposia 2003 Stem Cells, and 2005 Molecular Re gulation of Stem Cells.