Developmental Biology of Hematopoietic Stem Cells

Introduction

The hematopoietic system is made up of all of the blood cells, including red blood cells, platelets, myeloid cells such as granulocytes, monocyte-macrophages, and dendritic cells (DC), and lymphoid cells such as T, B, and natural killer (NK) cells. Together, these cells function to provide oxygen-carrying capacity, hemostasis, innate and adaptive immune function, and tissue regeneration and repair. Hematopoiesis is the process by which these lineage-committed blood cells are produced from hematopoietic stem and progenitor cells (HSPCs). In adults, hematopoietic stem cells (HSCs) sit at the top of the hematopoietic hierarchy and are functionally defined as cells capable of both self-renewal and differentiation into all the lineages of the blood system, giving rise to approximately 10 ¹² cells daily. Adult hematopoiesis takes place in the bone marrow in specific microenvironmental areas called niches . During development, transient waves of hematopoiesis supply the embryo with blood cells tailored to the needs of the developing organism, with the production of these cells occurring in distinct anatomic locations. This chapter will focus on the emergence of the hematopoietic system during embryogenesis, the establishment of “definitive,” or adult-like, hematopoiesis, and developmental regulation of HSC function during fetal and neonatal life. Basic concepts in stem cell biology will be introduced, including the recent development of induced pluripotent stem cells (iPSCs) and their use in further investigating developmental hematopoiesis.

Methodologies to Study Hematopoiesis

Our understanding of human developmental hematopoiesis has been shaped by the availability (or lack thereof) of human embryonic and fetal specimens that are accessible for investigation and, just as important, by the methodologies that have been available to study them. Phenotypic definition of HSCs that would allow for their prospective identification, by microscopy or flow cytometry, is imperfect. Our best definition of HSCs still only identifies a heterogeneous population of cells, only 1 in 10 of which possesses HSC activity (see below). HSCs remain defined functionally, on the basis of both their multi-lineage potential (multipotency) and their long-term reconstitution (self-renewal). As such, our ability to identify HSCs has only been as good as the assays available at the time that allow determination of these features of HSC function. Early experiments used in vitro colony-forming assays as surrogates for HSC activity, with the formation of cobblestone areas under stromal support cells (cobblestone area forming cells, CAFCs) or colony formation in semisolid methylcellulose medium (colony-forming unit cell, CFU-C, posited to be generated by long-term culture-initiating cells, LTC-ICs) being representative of HSC function. It is important that these measures of HSC activity are by definition retrospective, in that progenitor activity can only be inferred after differentiation has occurred, and heterogeneous populations of cells are identified. ^, In vitro colony formation can also be used to assay multipotency. In methylcellulose, HSCs and their derivatives are categorized on the basis of the characteristics of the colonies that they form. A single hematopoietic colony consists of more than 50 hematopoietic cells and may include hundreds or even thousands of cells. The more immature (early) progenitor cells are highly proliferative and form colonies consisting of more than two kinds of cell types; thus a colony containing cells for several lineages is considered to be derived from a multipotent progenitor (MPP) cell, such as a CFU-GEMM (granulocyte, erythrocyte, monocyte, megakaryocyte) or CFU-mix. A progenitor cell that forms a colony consisting of granulocyte and macrophages is called a CFU-GM . Later committed progenitor cells that form a single lineage are called CFU-G (for granulocyte ), or CFU-M (for macrophage ). In the erythroid lineage, early erythroid progenitors are proliferative and form burst colonies and are called erythroid burst-forming units (BFU-E), whereas late erythroid progenitors form relatively small colonies and are called CFU-E.

While methylcellulose has been the standard for years, it has some important limitations. It does not support the development of colonies from cells isolated from early embryonic tissues, thus precluding an analysis of the hematopoietic potential of these cells. To overcome this limitation, explant cultures were developed, in which undissociated regions of early human embryos are cultured in toto on a mouse stromal cell line (MS-5) that supports the further differentiation of hematopoietic precursors. Subsequent to this short undissociated culture period, cells are dissociated to yield a single-cell suspension and then seeded in methylcellulose for standard colony formation assessment. Such an approach allowed for the identification of hematopoietic potential at the earliest embryonic timepoints (see below). Another limitation is that methylcellulose only supports the development of megakaryocytic, erythrocytic, and myeloid cells. Identification of mouse stromal cell lines (including MS-5) that also support the differentiation of human hematopoietic differentiation has allowed for the refinement of in vitro assays to reveal megakaryocytic, erythroid, myeloid, and lymphoid potential from a single sorted progenitor cell. ^,

Perhaps the most significant advance in assessing human HSC function arose with the advent of immune-deficient mice that allow for xenotransplantation of human cells. These efforts were facilitated by the discovery of severe combined immune-deficient (Scid) mice that lack B and T cells, allowing for partial engraftment of human cells. Scid mice were found to allow infused human peripheral blood cells to produce antibodies specific to tetanus toxin ( Scid -PBL model), ^, and also could sustain long-term production of human B and T cells when surgically grafted with human fetal tissue and human fetal liver cells ( Scid-hu model). ^, Myeloid reconstitution was achieved by transplanting human bone marrow cells into Scid mice that also received myeloid-promoting cytokines. ^, The functional unit responsible for Scid mouse repopulation was thus referred to as the “ Scid -repopulating cell” or SRC, and to this day reconstitution in immune-deficient mice remains a gold standard for proof of “stemness.” Improvements to the Scid model came when Scid mice were crossed to nonobese diabetic mice (NOD- Scid mice) that supported higher levels of human cell engraftment due to a mutation in the Sirpa gene that promotes host macrophage tolerance of transplanted human cells. ^, NOD- Scid mice with a mutation in the IL2R common γ chain (NSG mice) ^, are completely devoid of B, T, and NK cells and support fivefold higher engraftment of human CD34 ⁺ cells than NOD- Scid mice. Finally, NSG mice engineered to express human cytokines are allowing for an even more robust evaluation of human hematopoiesis. ^, The most stringent definition of self-renewal activity is reserved for cells that when transplanted into an immunodeficient recipient give rise not only to multiple hematopoietic lineages over the long term (generally longer than 12 to 20 weeks is considered long term) but can also reconstitute a secondary immunodeficient recipient upon harvest from the primary recipient bone marrow, so-called serial transplantation . Although transplantation into immunodeficient mice remains the gold standard for “stemness” in the field, it should be remembered that this represents the activity of a stem cell removed from its native microenvironment and may thus not be representative of unperturbed hematopoiesis.

Developmental Hematopoiesis in the Human Embryo and Fetus

Hematopoiesis starts in the extraembryonic yolk sac (YS) of the human embryo. The human embryonic period encompasses the time from the moment of fertilization to the end of the eighth week of embryonic age (tenth postmenstrual week). Human hematopoiesis during the embryonic period comprises a period of primitive hematopoiesis in the YS, HSC emergence in the aorta-gonad-mesonephros (AGM) region from the hemogenic endothelium, and the seeding of HSCs from the AGM into the liver. During the fetal period (8 weeks until birth), HSCs and their progenitors exist in the placenta and circulation, expand in the fetal liver, and subsequently move into the bone marrow during the second trimester, where hematopoiesis resides throughout adult life. This is summarized in Fig. 107.1 .

Yolk Sac Hematopoiesis

The YS forms outside of the embryo in a balloon-like structure 7 to 8 days after implantation. Within the YS, mesodermal cells aggregate in initially homogeneous solid clusters. The cells at the periphery of these clusters acquire the characteristics of endothelial cells while the internal cells regress to form vessel lumens. Groups of mesodermal cells remain associated to newly formed endothelial cells, forming “blood islands” by embryonic day 16. Within this early YS vascular network, prior to the onset of circulation at embryonic day 21, are found primarily large primitive nucleated erythrocytes (megaloblasts), with scarce interspersed macrophages and megakaryocytes.

Early functional studies of human YS hematopoiesis from 4.5 week embryos has demonstrated the presence of clonogenic progenitors including early (BFU-E) and late (CFU-E) erythroid, granulo-macrocytic (CFU-GM), and mixed (CFU-GEMM) progenitors. A subsequent study confirmed the presence of progenitors with both erythroid and myeloid clonogenic potential as early as 25 days. The frequency of these progenitors drops sharply after about the sixth week of development, a time at which hematopoiesis is leaving the YS and transitioning to the embryo proper, primarily taking up residence in the liver rudiment (see below). This period is also marked by the switch, in the liver, from embryonic to fetal hemoglobin (ξ-globin → α-globin and ε- globin → γ-globin) and the transition from primitive (nucleated) megaloblasts to definitive (enucleated) macrocytes. These studies were initially taken as supporting a monoclonal model of human hematopoiesis, in which cells arising in the YS give rise first to primitive cells of the erythroid and myeloid (primarily macrophage) lineage, then migrate to the liver where definitive hematopoiesis (including the appearance of definitive HSCs) is established. However, these studies were done in specimens obtained after the onset of circulation (day 21). It is now generally accepted that definitive hematopoiesis originates in the embryo proper (within specialized regions of the aorta, see below), and that any long-term reconstituting HSCs found in the YS have migrated from intra-embryonic tissues where they originate. The only study to have examined the hematopoietic potential of the YS and intra-embryonic tissues prior to the onset of circulation (2 specimens at 19 days) demonstrated that YS explants generate few CD34 ⁺ cells that decrease rapidly in numbers as ex vivo culture progress, and only give rise to myeloid and NK cells in vitro. In contrast, intra-embryonic tissues from the same specimen give rise to larger numbers of CD34 ⁺ cells that survive longer in culture and have myeloid, NK, B and T cell potential, suggestive of discrete waves of hematopoiesis with a restriction in the hematopoietic potential of YS progenitors to primitive cells. In agreement with the findings of this study, a spatiotemporal analysis of HSC activity from early human embryonic tissues found that cells capable of long-term hematopoietic reconstitution in NSG mice are present in the AGM region as early as 32 to 33 days, 5 days prior to detection of this activity in the YS, supporting the hypothesis that any cell with long-term reconstituting activity in the YS has arrived there via the circulation but originated in the embryo proper.

Definitive Hematopoiesis Emerges From Endothelial Cells Within the Aorta-Gonad-Mesonephros

The cells that will ultimately give rise to self-renewing adult HSCs arise within the ventral wall of the dorsal aorta between embryonic days 27 to 40. ^, This area of the vessel wall is thus referred to as hemogenic endothelium. The paired dorsal aortae appear on day 21 and start fusing from day 25 onward. In the caudal region of the embryo, the aorta does not fuse completely, giving rise to the left and right umbilical arteries. On day 30, the vitelline artery appears and connects the embryonic blood vessels with those of the YS. Endothelial cells express CD34 (which will subsequently mark adult HSCs) as early as embryonic day 19. ^, Clusters of cells expressing both CD34 and CD45 (a pan-hematopoietic cell surface marker) are found in the periumbilical region of the aorta as early as day 27, as well as in the vitelline artery between days 30 and 36.

The intra-aortic hematopoietic clusters (IAHCs) increase in size to reach several hundred by day 36 and decrease in size gradually by day 40 ( Fig. 107.2 ). ^, IAHCs express known early hematopoietic transcription factors such as T-cell acute lymphocytic leukemia 1/stem cell leukemia (Tal/SCL), myeloblastosis oncogene (c-Myb), and GATA-binding protein 2 (GATA2) (shown by in situ hybridization). The mesenchymal cells underlying the ventral wall of the aorta express GATA3 or bone morphogenetic protein 4 (BMP4) , consistent with a hypothesis that HSCs emerge from the subaortic mesenchymal region and migrate into the blood vessel lumen. Whether these GATA3 ⁺ or BMP4 ⁺ cells are precursors or inducers of human HSCs remains to be resolved. ^, IAHCs also express angiotensin-converting enzyme (ACE; CD143), which is expressed by HSCs from fetal liver, umbilical cord blood, and adult hematopoietic tissues. Additionally, ACE marks rare CD34 ⁻ CD45 ⁻ cells in the hemogenic region of the embryo at embryonic days 23 to 26 that have 40 times the hematopoietic colony generating capacity of ACE-negative cells, suggesting that ACE may be a marker for the mesodermal precursors that give rise to IAHCs and definitive hematopoiesis. In fact, sections of intra-embryonic splanchnopleura from embryonic day 19 specimens, before the aorta has even begun to form, also contains hematopoietic potential, demonstrating that mesodermal cells fated for hematopoietic differentiation are already present within the region of the embryo that will subsequently give rise to the aorta. Explant culture of defined portions of the aorta between embryonic days 24 and 58 confirmed that only the median region of the aorta, encompassing the periumbilical region and origin of the vitelline artery, gives rise to hematopoietic colonies, and furthermore that cells contained within this region are multipotent, being able to give rise to myeloid as well as lymphoid cells in vitro and all lineages (erythroid, myeloid, lymphoid) in vivo. Cells in this region can also self-renew, as demonstrated by the ability of AGM-derived bone marrow cells in recipient NSG mice to reconstitute the hematopoietic system of secondary NSG recipient mice. Based on the numbers of cells transplanted, this allowed for the calculation that one HSC from the isolated AGM region was capable of generating at least 300 daughter HSCs on transplantation. Further refinement in the location and marker expression has shown that cells with the phenotype CD45 ⁺ cKit ⁺ Thy1 ⁺ Endoglin ⁺ Runx1 ⁺ CD38 ^−/lo CD45RA ⁻ localized within the ventral wall of the dorsal aorta are enriched up to 1000-fold for hematopoietic activity compared to the total AGM region, a surface marker phenotype that fits well with our current understanding of the human hematopoietic hierarchy (see below).