by Jos Domen*, Amy Wagers** and Irving L. Weissman***

Blood and the system that forms it, known as the hematopoietic system, consist of many cell types with specialized functions (see Figure 2.1). Red blood cells (erythrocytes) carry oxygen to the tissues. Platelets (derived from megakaryocytes) help prevent bleeding. Granulocytes (neutrophils, basophils and eosinophils) and macrophages (collectively known as myeloid cells) fight infections from bacteria, fungi, and other parasites such as nematodes (ubiquitous small worms). Some of these cells are also involved in tissue and bone remodeling and removal of dead cells. B-lymphocytes produce antibodies, while T-lymphocytes can directly kill or isolate by inflammation cells recognized as foreign to the body, including many virus-infected cells and cancer cells. Many blood cells are short-lived and need to be replenished continuously; the average human requires approximately one hundred billion new hematopoietic cells each day. The continued production of these cells depends directly on the presence of Hematopoietic Stem Cells (HSCs), the ultimate, and only, source of all these cells.

Figure 2.1. Hematopoietic and stromal cell differentiation.

2001 Terese Winslow (assisted by Lydia Kibiuk)

The search for stem cells began in the aftermath of the bombings in Hiroshima and Nagasaki in 1945. Those who died over a prolonged period from lower doses of radiation had compromised hematopoietic systems that could not regenerate either sufficient white blood cells to protect against otherwise nonpathogenic infections or enough platelets to clot their blood. Higher doses of radiation also killed the stem cells of the intestinal tract, resulting in more rapid death. Later, it was demonstrated that mice that were given doses of whole body X-irradiation developed the same radiation syndromes; at the minimal lethal dose, the mice died from hematopoietic failure approximately two weeks after radiation exposure.1 Significantly, however, shielding a single bone or the spleen from radiation prevented this irradiation syndrome. Soon thereafter, using inbred strains of mice, scientists showed that whole-body-irradiated mice could be rescued from otherwise fatal hematopoietic failure by injection of suspensions of cells from blood-forming organs such as the bone marrow.2 In 1956, three laboratories demonstrated that the injected bone marrow cells directly regenerated the blood-forming system, rather than releasing factors that caused the recipients' cells to repair irradiation damage.35 To date, the only known treatment for hematopoietic failure following whole body irradiation is transplantation of bone marrow cells or HSCs to regenerate the blood-forming system in the host organisms.6,7

The hematopoietic system is not only destroyed by the lowest doses of lethal X-irradiation (it is the most sensitive of the affected vital organs), but also by chemotherapeutic agents that kill dividing cells. By the 1960s, physicians who sought to treat cancer that had spread (metastasized) beyond the primary cancer site attempted to take advantage of the fact that a large fraction of cancer cells are undergoing cell division at any given point in time. They began using agents (e.g., chemical and X-irradiation) that kill dividing cells to attempt to kill the cancer cells. This required the development of a quantitative assessment of damage to the cancer cells compared that inflicted on normal cells. Till and McCulloch began to assess quantitatively the radiation sensitivity of one normal cell type, the bone marrow cells used in transplantation, as it exists in the body. They found that, at sub-radioprotective doses of bone marrow cells, mice that died 1015 days after irradiation developed colonies of myeloid and erythroid cells (see Figure 2.1 for an example) in their spleens. These colonies correlated directly in number with the number of bone marrow cells originally injected (approximately 1 colony per 7,000 bone marrow cells injected).8 To test whether these colonies of blood cells derived from single precursor cells, they pre-irradiated the bone marrow donors with low doses of irradiation that would induce unique chromosome breaks in most hematopoietic cells but allow some cells to survive. Surviving cells displayed radiation-induced and repaired chromosomal breaks that marked each clonogenic (colony-initiating) hematopoietic cell.9 The researchers discovered that all dividing cells within a single spleen colony, which contained different types of blood cells, contained the same unique chromosomal marker. Each colony displayed its own unique chromosomal marker, seen in its dividing cells.9 Furthermore, when cells from a single spleen colony were re-injected into a second set of lethally-irradiated mice, donor-derived spleen colonies that contained the same unique chromosomal marker were often observed, indicating that these colonies had been regenerated from the same, single cell that had generated the first colony. Rarely, these colonies contained sufficient numbers of regenerative cells both to radioprotect secondary recipients (e.g., to prevent their deaths from radiation-induced blood cell loss) and to give rise to lymphocytes and myeloerythroid cells that bore markers of the donor-injected cells.10,11 These genetic marking experiments established the fact that cells that can both self-renew and generate most (if not all) of the cell populations in the blood must exist in bone marrow. At the time, such cells were called pluripotent HSCs, a term later modified to multipotent HSCs.12,13 However, identifying stem cells in retrospect by analysis of randomly chromosome-marked cells is not the same as being able to isolate pure populations of HSCs for study or clinical use.

Achieving this goal requires markers that uniquely define HSCs. Interestingly, the development of these markers, discussed below, has revealed that most of the early spleen colonies visible 8 to 10 days after injection, as well as many of the later colonies, visible at least 12 days after injection, are actually derived from progenitors rather than from HSCs. Spleen colonies formed by HSCs are relatively rare and tend to be present among the later colonies.14,15 However, these findings do not detract from Till and McCulloch's seminal experiments to identify HSCs and define these unique cells by their capacities for self-renewal and multilineage differentiation.

While much of the original work was, and continues to be, performed in murine model systems, strides have been made to develop assays to study human HSCs. The development of Fluorescence Activated Cell Sorting (FACS) has been crucial for this field (see Figure 2.2). This technique enables the recognition and quantification of small numbers of cells in large mixed populations. More importantly, FACS-based cell sorting allows these rare cells (1 in 2000 to less than 1 in 10,000) to be purified, resulting in preparations of near 100% purity. This capability enables the testing of these cells in various assays.

Figure 2.2. Enrichment and purification methods for hematopoietic stem cells. Upper panels illustrate column-based magnetic enrichment. In this method, the cells of interest are labeled with very small iron particles (A). These particles are bound to antibodies that only recognize specific cells. The cell suspension is then passed over a column through a strong magnetic field which retains the cells with the iron particles (B). Other cells flow through and are collected as the depleted negative fraction. The magnet is removed, and the retained cells are collected in a separate tube as the positive or enriched fraction (C). Magnetic enrichment devices exist both as small research instruments and large closed-system clinical instruments.

Lower panels illustrate Fluorescence Activated Cell Sorting (FACS). In this setting, the cell mixture is labeled with fluorescent markers that emit light of different colors after being activated by light from a laser. Each of these fluorescent markers is attached to a different monoclonal antibody that recognizes specific sets of cells (D). The cells are then passed one by one in a very tight stream through a laser beam (blue in the figure) in front of detectors (E) that determine which colors fluoresce in response to the laser. The results can be displayed in a FACS-plot (F). FACS-plots (see figures 3 and 4 for examples) typically show fluorescence levels per cell as dots or probability fields. In the example, four groups can be distinguished: Unstained, red-only, green-only, and red-green double labeling. Each of these groups, e.g., green fluorescence-only, can be sorted to very high purity. The actual sorting happens by breaking the stream shown in (E) into tiny droplets, each containing 1 cell, that then can be sorted using electric charges to move the drops. Modern FACS machines use three different lasers (that can activate different set of fluorochromes), to distinguish up to 8 to 12 different fluorescence colors and sort 4 separate populations, all simultaneously.

Magnetic enrichment can process very large samples (billions of cells) in one run, but the resulting cell preparation is enriched for only one parameter (e.g., CD34) and is not pure. Significant levels of contaminants (such as T-cells or tumor cells) remain present. FACS results in very pure cell populations that can be selected for several parameters simultaneously (e.g., Linneg, CD34pos, CD90pos), but it is more time consuming (10,000 to 50,000 cells can be sorted per second) and requires expensive instrumentation.

2001 Terese Winslow (assisted by Lydia Kibiuk)

Assays have been developed to characterize hematopoietic stem and progenitor cells in vitro and in vivo (Figure 2.3).16,17In vivo assays that are used to study HSCs include Till and McCulloch's classical spleen colony forming (CFU-S) assay,8 which measures the ability of HSC (as well as blood-forming progenitor cells) to form large colonies in the spleens of lethally irradiated mice. Its main advantage (and limitation) is the short-term nature of the assay (now typically 12 days). However, the assays that truly define HSCs are reconstitution assays.16,18 Mice that have been quot;preconditionedquot; by lethal irradiation to accept new HSCs are injected with purified HSCs or mixed populations containing HSCs, which will repopulate the hematopoietic systems of the host mice for the life of the animal. These assays typically use different types of markers to distinguish host and donor-derived cells.

For example, allelic assays distinguish different versions of a particular gene, either by direct analysis of dna or of the proteins expressed by these alleles. These proteins may be cell-surface proteins that are recognized by specific monoclonal antibodies that can distinguish between the variants (e.g., CD45 in Figure 2.3) or cellular proteins that may be recognized through methods such as gel-based analysis. Other assays take advantage of the fact that male cells can be detected in a female host by detecting the male-cell-specific Y-chromosome by molecular assays (e.g., polymerase chain reaction, or PCR).

Figure 2.3. Assays used to detect hematopoietic stem cells. The tissue culture assays, which are used frequently to test human cells, include the ability of the cells to be tested to grow as quot;cobblestonesquot; (the dark cells in the picture) for 5 to 7 weeks in culture. The Long Term Culture-Initiating Cell assay measures whether hematopoietic progenitor cells (capable of forming colonies in secondary assays, as shown in the picture) are still present after 5 to 7 weeks of culture.

In vivo assays in mice include the CFU-S assay, the original stem cell assay discussed in the introduction. The most stringent hematopoietic stem cell assay involves looking for the long-term presence of donor-derived cells in a reconstituted host. The example shows host-donor recognition by antibodies that recognize two different mouse alleles of CD45, a marker present on nearly all blood cells. CD45 is also a good marker for distinguishing human blood cells from mouse blood cells when testing human cells in immunocompromised mice such as NOD/SCID. Other methods such as pcr-markers, chromosomal markers, and enzyme markers can also be used to distinguish host and donor cells.

Small numbers of HSCs (as few as one cell in mouse experiments) can be assayed using competitive reconstitutions, in which a small amount of host-type bone marrow cells (enough to radioprotect the host and thus ensure survival) is mixed in with the donor-HSC population. To establish long-term reconstitutions in mouse models, the mice are followed for at least 4 months after receiving the HSCs. Serial reconstitution, in which the bone marrow from a previously-irradiated and reconstituted mouse becomes the HSC source for a second irradiated mouse, extends the potential of this assay to test lifespan and expansion limits of HSCs. Unfortunately, the serial transfer assay measures both the lifespan and the transplantability of the stem cells. The transplantability may be altered under various conditions, so this assay is not the sine qua non of HSC function. Testing the in vivo activity of human cells is obviously more problematic.

Several experimental models have been developed that allow the testing of human cells in mice. These assays employ immunologically-incompetent mice (mutant mice that cannot mount an immune response against foreign cells) such as SCID1921 or NOD-SCID mice.22,23 Reconstitution can be performed in either the presence or absence of human fetal bone or thymus implants to provide a more natural environment in which the human cells can grow in the mice. Recently NOD/SCID/c-/- mice have been used as improved recipients for human HSCs, capable of complete reconstitution with human lymphocytes, even in the absence of additional human tissues.24 Even more promising has been the use of newborn mice with an impaired immune system (Rag-2-/-C-/-), which results in reproducible production of human B- and T-lymphoid and myeloerythroid cells.25 These assays are clearly more stringent, and thus more informative, but also more difficult than the in vitro HSC assays discussed below. However, they can only assay a fraction of the lifespan under which the cells would usually have to function. Information on the long-term functioning of cells can only be derived from clinical HSC transplantations.

A number of assays have been developed to recognize HSCs in vitro (e.g., in tissue culture). These are especially important when assaying human cells. Since transplantation assays for human cells are limited, cell culture assays often represent the only viable option. In vitro assays for HSCs include Long-Term Culture-Initializing Cell (LTC-IC) assays2628 and Cobble-stone Area Forming Cell (CAFC) assays.29 LTC-IC assays are based on the ability of HSCs, but not more mature progenitor cells, to maintain progenitor cells with clonogenic potential over at least a five-week culture period. CAFC assays measure the ability of HSCs to maintain a specific and easily recognizable way of growing under stromal cells for five to seven weeks after the initial plating. Progenitor cells can only grow in culture in this manner for shorter periods of time.

While initial experiments studied HSC activity in mixed populations, much progress has been made in specifically describing the cells that have HSC activity. A variety of markers have been discovered to help recognize and isolate HSCs. Initial marker efforts focused on cell size, density, and recognition by lectins (carbohydrate-binding proteins derived largely from plants),30 but more recent efforts have focused mainly on cell surface protein markers, as defined by monoclonal antibodies. For mouse HSCs, these markers include panels of 8 to 14 different monoclonal antibodies that recognize cell surface proteins present on differentiated hematopoietic lineages, such as the red blood cell and macrophage lineages (thus, these markers are collectively referred to as quot;Linquot;),13,31 as well as the proteins Sca-1,13,31 CD27,32 CD34,33 CD38,34 CD43,35 CD90.1(Thy-1.1),13,31 CD117(c-Kit),36 AA4.1,37 and MHC class I,30 and CD150.38 Human HSCs have been defined with respect to staining for Lin,39 CD34,40 CD38,41 CD43,35 CD45RO,42 CD45RA,42 CD59,43 CD90,39 CD109,44 CD117,45 CD133,46,47CD166,48 and HLA DR(human).49,50 In addition, metabolic markers/dyes such as rhodamine123 (which stains mitochondria),51 Hoechst33342 (which identifies MDR-type drug efflux activity),52 Pyronin-Y (which stains RNA),53 and BAAA (indicative of aldehyde dehydrogenase enzyme activity)54 have been described. While none of these markers recognizes functional stem cell activity, combinations (typically with 3 to 5 different markers, see examples below) allow for the purification of near-homogenous populations of HSCs. The ability to obtain pure preparations of HSCs, albeit in limited numbers, has greatly facilitated the functional and biochemical characterization of these important cells. However, to date there has been limited impact of these discoveries on clinical practice, as highly purified HSCs have only rarely been used to treat patients (discussed below). The undeniable advantages of using purified cells (e.g., the absence of contaminating tumor cells in autologous transplantations) have been offset by practical difficulties and increased purification costs.

Figure 2.4. Examples of Hematopoietic Stem Cell staining patterns in mouse bone marrow (top) and human mobilized peripheral blood (bottom). The plots on the right show only the cells present in the left blue box. The cells in the right blue box represent HSCs. Stem cells form a rare fraction of the cells present in both cases.

HSC assays, when combined with the ability to purify HSCs, have provided increasingly detailed insight into the cells and the early steps involved in the differentiation process. Several marker combinations have been developed that describe murine HSCs, including [CD117high, CD90.1low, Linneg/low, Sca-1pos],15 [CD90.1low, Linneg, Sca-1pos Rhodamine123low],55 [CD34neg/low, CD117pos, Sca-1pos, Linneg],33 [CD150 pos, CD48neg, CD244neg],38 and quot;side-populationquot; cells using Hoechst-dye.52 Each of these combinations allows purification of HSCs to near-homogeneity. Figure 2.4 shows an example of an antibody combination that can recognize mouse HSCs. Similar strategies have been developed to purify human HSCs, employing markers such as CD34, CD38, Lin, CD90, CD133 and fluorescent substrates for the enzyme, aldehyde dehydrogenase. The use of highly purified human HSCs has been mainly experimental, and clinical use typically employs enrichment for one marker, usually CD34. CD34 enrichment yields a population of cells enriched for HSC and blood progenitor cells but still contains many other cell types. However, limited trials in which highly FACS-purified CD34pos CD90pos HSCs (see Figure 2.4) were used as a source of reconstituting cells have demonstrated that rapid reconstitution of the blood system can reliably be obtained using only HSCs.5658

The purification strategies described above recognize a rare subset of cells. Exact numbers depend on the assay used as well as on the genetic background studied.16 In mouse bone marrow, 1 in 10,000 cells is a hematopoietic stem cell with the ability to support long-term hematopoiesis following transplantation into a suitable host. When short-term stem cells, which have a limited self-renewal capacity, are included in the estimation, the frequency of stem cells in bone marrow increases to 1 in 1,000 to 1 in 2,000 cells in humans and mice. The numbers present in normal blood are at least ten-fold lower than in marrow.

None of the HSC markers currently used is directly linked to an essential HSC function, and consequently, even within a species, markers can differ depending on genetic alleles,59 mouse strains,60 developmental stages,61 and cell activation stages.62,63 Despite this, there is a clear correlation in HSC markers between divergent species such as humans and mice. However, unless the ongoing attempts at defining the complete HSC gene expression patterns will yield usable markers that are linked to essential functions for maintaining the quot;stemnessquot; of the cells,64,65 functional assays will remain necessary to identify HSCs unequivocally.16

More recently, efforts at defining hematopoietic populations by cell surface or other FACS-based markers have been extended to several of the progenitor populations that are derived from HSCs (see Figure 2.5). Progenitors differ from stem cells in that they have a reduced differentiation capacity (they can generate only a subset of the possible lineages) but even more importantly, progenitors lack the ability to self-renew. Thus, they have to be constantly regenerated from the HSC population. However, progenitors do have extensive proliferative potential and can typically generate large numbers of mature cells. Among the progenitors defined in mice and humans are the Common Lymphoid Progenitor (CLP),66,67 which in adults has the potential to generate all of the lymphoid but not myeloerythroid cells, and a Common Myeloid Progenitor (CMP), which has the potential to generate all of the mature myeloerythroid, but not lymphoid, cells.68,69 While beyond the scope of this overview, hematopoietic progenitors have clinical potential and will likely see clinical use.70,71

Figure 2.5. Relationship between several of the characterized hematopoietic stem cells and early progenitor cells. Differentiation is indicated by colors; the more intense the color, the more mature the cells. Surface marker distinctions are subtle between these early cell populations, yet they have clearly distinct potentials. Stem cells can choose between self-renewal and differentiation. Progenitors can expand temporarily but always continue to differentiate (other than in certain leukemias). The mature lymphoid (T-cells, B-cells, and Natural Killer cells) and myeloerythroid cells (granulocytes, macrophages, red blood cells, and platelets) that are produced by these stem and progenitor cells are shown in more detail in Figure 2.1.

HSCs have a number of unique properties, the combination of which defines them as such.16 Among the core properties are the ability to choose between self-renewal (remain a stem cell after cell division) or differentiation (start the path towards becoming a mature hematopoietic cell). In addition, HSCs migrate in regulated fashion and are subject to regulation by apoptosis (programmed cell death). The balance between these activities determines the number of stem cells that are present in the body.

One essential feature of HSCs is the ability to self-renew, that is, to make copies with the same or very similar potential. This is an essential property because more differentiated cells, such as hematopoietic progenitors, cannot do this, even though most progenitors can expand significantly during a limited period of time after being generated. However, for continued production of the many (and often short-lived) mature blood cells, the continued presence of stem cells is essential. While it has not been established that adult HSCs can self-renew indefinitely (this would be difficult to prove experimentally), it is clear from serial transplantation experiments that they can produce enough cells to last several (at least four to five) lifetimes in mice. It is still unclear which key signals allow self-renewal. One link that has been noted is telomerase, the enzyme necessary for maintaining telomeres, the DNA regions at the end of chromosomes that protect them from accumulating damage due to DNA replication. Expression of telomerase is associated with self-renewal activity.72 However, while absence of telomerase reduces the self-renewal capacity of mouse HSCs, forced expression is not sufficient to enable HSCs to be transplanted indefinitely; other barriers must exist.73,74

It has proven surprisingly difficult to grow HSCs in culture despite their ability to self-renew. Expansion in culture is routine with many other cells, including neural stem cells and ES cells. The lack of this capacity for HSCs severely limits their application, because the number of HSCs that can be isolated from mobilized blood, umbilical cord blood, or bone marrow restricts the full application of HSC transplantation in man (whether in the treatment of nuclear radiation exposure or transplantation in the treatment of blood cell cancers or genetic diseases of the blood or blood-forming system). Engraftment periods of 50 days or more were standard when limited numbers of bone marrow or umbilical cord blood cells were used in a transplant setting, reflecting the low level of HSCs found in these native tissues. Attempts to expand HSCs in tissue culture with known stem-cell stimulators, such as the cytokines stem cell factor/steel factor (KitL), thrombopoietin (TPO), interleukins 1, 3, 6, 11, plus or minus the myeloerythroid cytokines GM-CSF, G-CSF, M-CSF, and erythropoietin have never resulted in a significant expansion of HSCs.16,75 Rather, these compounds induce many HSCs into cell divisions that are always accompanied by cellular differentiation.76 Yet many experiments demonstrate that the transplantation of a single or a few HSCs into an animal results in a 100,000-fold or greater expansion in the number of HSCs at the steady state while simultaneously generating daughter cells that permitted the regeneration of the full blood-forming system.7780 Thus, we do not know the factors necessary to regenerate HSCs by self-renewing cell divisions. By investigating genes transcribed in purified mouse LT-HSCs, investigators have found that these cells contain expressed elements of the Wnt/fzd/beta-catenin signaling pathway, which enables mouse HSCs to undergo self-renewing cell divisions.81,82 Overexpression of several other proteins, including HoxB48386 and HoxA987 has also been reported to achieve this. Other signaling pathways that are under investigation include Notch and Sonic hedgehog.75 Among the intracellular proteins thought to be essential for maintaining the quot;stem cellquot; state are Polycomb group genes, including Bmi-1.88 Other genes, such as c-Myc and JunB have also been shown to play a role in this process.89,90Much remains to be discovered, including the identity of the stimuli that govern self-renewal in vivo, as well as the composition of the environment (the stem cell quot;nichequot;) that provides these stimuli.91 The recent identification of osteoblasts, a cell type known to be involved in bone formation, as a critical component of this environment92,93 will help to focus this search. For instance, signaling by Angiopoietin-1 on osteoblasts to Tie-2 receptors on HSCs has recently been suggested to regulate stem cell quiescence (the lack of cell division).94 It is critical to discover which pathways operate in the expansion of human HSCs to take advantage of these pathways to improve hematopoietic transplantation.

Differentiation into progenitors and mature cells that fulfill the functions performed by the hematopoietic system is not a unique HSC property, but, together with the option to self-renew, defines the core function of HSCs. Differentiation is driven and guided by an intricate network of growth factors and cytokines. As discussed earlier, differentiation, rather than self-renewal, seems to be the default outcome for HSCs when stimulated by many of the factors to which they have been shown to respond. It appears that, once they commit to differentiation, HSCs cannot revert to a self-renewing state. Thus, specific signals, provided by specific factors, seem to be needed to maintain HSCs. This strict regulation may reflect the proliferative potential present in HSCs, deregulation of which could easily result in malignant diseases such as leukemia or lymphoma.

Migration of HSCs occurs at specific times during development (i.e., seeding of fetal liver, spleen and eventually, bone marrow) and under certain conditions (e.g., cytokine-induced mobilization) later in life. The latter has proven clinically useful as a strategy to enhance normal HSC proliferation and migration, and the optimal mobilization regimen for HSCs currently used in the clinic is to treat the stem cell donor with a drug such as cytoxan, which kills most of his or her dividing cells. Normally, only about 8% of LT-HSCs enter the cell cycle per day,95,96 so HSCs are not significantly affected by a short treatment with cytoxan. However, most of the downstream blood progenitors are actively dividing,66,68 and their numbers are therefore greatly depleted by this dose, creating a demand for a regenerated blood-forming system. Empirically, cytokines or growth factors such as G-CSF and KitL can increase the number of HSCs in the blood, especially if administered for several days following a cytoxan pulse. The optimized protocol of cytoxan plus G-CSF results in several self-renewing cell divisions for each resident LT-HSC in mouse bone marrow, expanding the number of HSCs 12- to 15-fold within two to three days.97 Then, up to one-half of the daughter cells of self-renewing dividing LT-HSCs (estimated to be up to 105 per mouse per day98) leave the bone marrow, enter the blood, and within minutes engraft other hematopoietic sites, including bone marrow, spleen, and liver.98 These migrating cells can and do enter empty hematopoietic niches elsewhere in the bone marrow and provide sustained hematopoietic stem cell self-renewal and hematopoiesis.98,99 It is assumed that this property of mobilization of HSCs is highly conserved in evolution (it has been shown in mouse, dog and humans) and presumably results from contact with natural cell-killing agents in the environment, after which regeneration of hematopoiesis requires restoring empty HSC niches. This means that functional, transplantable HSCs course through every tissue of the body in large numbers every day in normal individuals.

Apoptosis, or programmed cell death, is a mechanism that results in cells actively self-destructing without causing inflammation. Apoptosis is an essential feature in multicellular organisms, necessary during development and normal maintenance of tissues. Apoptosis can be triggered by specific signals, by cells failing to receive the required signals to avoid apoptosis, and by exposure to infectious agents such as viruses. HSCs are not exempt; apoptosis is one mechanism to regulate their numbers. This was demonstrated in transgenic mouse experiments in which HSC numbers doubled when the apoptosis threshold was increased.76 This study also showed that HSCs are particularly sensitive and require two signals to avoid undergoing apoptosis.

The best-known location for HSCs is bone marrow, and bone marrow transplantation has become synonymous with hematopoietic cell transplantation, even though bone marrow itself is increasingly infrequently used as a source due to an invasive harvesting procedure that requires general anesthesia. In adults, under steady-state conditions, the majority of HSCs reside in bone marrow. However, cytokine mobilization can result in the release of large numbers of HSCs into the blood. As a clinical source of HSCs, mobilized peripheral blood (MPB) is now replacing bone marrow, as harvesting peripheral blood is easier for the donors than harvesting bone marrow. As with bone marrow, mobilized peripheral blood contains a mixture of hematopoietic stem and progenitor cells. MPB is normally passed through a device that enriches cells that express CD34, a marker on both stem and progenitor cells. Consequently, the resulting cell preparation that is infused back into patients is not a pure HSC preparation, but a mixture of HSCs, hematopoietic progenitors (the major component), and various contaminants, including T cells and, in the case of autologous grafts from cancer patients, quite possibly tumor cells. It is important to distinguish these kinds of grafts, which are the grafts routinely given, from highly purified HSC preparations, which essentially lack other cell types.

In the late 1980s, umbilical cord blood (UCB) was recognized as an important clinical source of HSCs.100,101 Blood from the placenta and umbilical cord is a rich source of hematopoietic stem cells, and these cells are typically discarded with the afterbirth. Increasingly, UCB is harvested, frozen, and stored in cord blood banks, as an individual resource (donor-specific source) or as a general resource, directly available when needed. Cord blood has been used successfully to transplant children and (far less frequently) adults. Specific limitations of UCB include the limited number of cells that can be harvested and the delayed immune reconstitution observed following UCB transplant, which leaves patients vulnerable to infections for a longer period of time. Advantages of cord blood include its availability, ease of harvest, and the reduced risk of graft-versus-host-disease (GVHD). In addition, cord blood HSCs have been noted to have a greater proliferative capacity than adult HSCs. Several approaches have been tested to overcome the cell dose issue, including, with some success, pooling of cord blood samples.101,102 Ex vivo expansion in tissue culture, to which cord blood cells are more amenable than adult cells, is another approach under active investigation.103

The use of cord blood has opened a controversial treatment strategyembryo selection to create a related UCB donor.104 In this procedure, embryos are conceived by in vitro fertilization. The embryos are tested by pre-implantation genetic diagnosis, and embryos with transplantation antigens matching those of the affected sibling are implanted. Cord blood from the resulting newborn is then used to treat this sibling. This approach, successfully pioneered at the University of Minnesota, can in principle be applied to a wide variety of hematopoietic disorders. However, the ethical questions involved argue for clear regulatory guidelines.105

Embryonic stem (ES) cells form a potential future source of HSCs. Both mouse and human ES cells have yielded hematopoietic cells in tissue culture, and they do so relatively readily.106 However, recognizing the actual HSCs in these cultures has proven problematic, which may reflect the variability in HSC markers or the altered reconstitution behavior of these HSCs, which are expected to mimic fetal HSC. This, combined with the potential risks of including undifferentiated cells in an ES-cell-derived graft means that, based on the current science, clinical use of ES cell-derived HSCs remains only a theoretical possibility for now.

An ongoing set of investigations has led to claims that HSCs, as well as other stem cells, have the capacity to differentiate into a much wider range of tissues than previously thought possible. It has been claimed that, following reconstitution, bone marrow cells can differentiate not only into blood cells but also muscle cells (both skeletal myocytes and cardiomyocytes),107111 brain cells,112,113 liver cells,114,115 skin cells, lung cells, kidney cells, intestinal cells,116 and pancreatic cells.117 Bone marrow is a complex mixture that contains numerous cell types. In addition to HSCs, at least one other type of stem cell, the mesenchymal stem cell (MSC), is present in bone marrow. MSCs, which have become the subject of increasingly intense investigation, seem to retain a wide range of differentiation capabilities in vitro that is not restricted to mesodermal tissues, but includes tissues normally derived from other embryonic germ layers (e.g., neurons).118120MSCs are discussed in detail in Dr. Catherine Verfaillie's testimony to the President's Council on Bioethics at this website: refer to Appendix J (page 295) and will not be discussed further here. However, similar claims of differentiation into multiple diverse cell types, including muscle,111 liver,114 and different types of epithelium116 have been made in experiments that assayed partially- or fully-purified HSCs. These experiments have spawned the idea that HSCs may not be entirely or irreversibly committed to forming the blood, but under the proper circumstances, HSCs may also function in the regeneration or repair of non-blood tissues. This concept has in turn given rise to the hypothesis that the fate of stem cells is quot;plastic,quot; or changeable, allowing these cells to adopt alternate fates if needed in response to tissue-derived regenerative signals (a phenomenon sometimes referred to as quot;transdifferentiationquot;). This in turn seems to bolster the argument that the full clinical potential of stem cells can be realized by studying only adult stem cells, foregoing research into defining the conditions necessary for the clinical use of the extensive differentiation potential of embryonic stem cells. However, as discussed below, such quot;transdifferentiationquot; claims for specialized adult stem cells are controversial, and alternative explanations for these observations remain possible, and, in several cases, have been documented directly.

While a full discussion of this issue is beyond the scope of this overview, several investigators have formulated criteria that must be fulfilled to demonstrate stem cell plasticity.121,122 These include (i) clonal analysis, which requires the transfer and analysis of single, highly-purified cells or individually marked cells and the subsequent demonstration of both quot;normalquot; and quot;plasticquot; differentiation outcomes, (ii) robust levels of quot;plasticquot; differentiation outcome, as extremely rare events are difficult to analyze and may be induced by artefact, and (iii) demonstration of tissue-specific function of the quot;transdifferentiatedquot; cell type. Few of the current reports fulfill these criteria, and careful analysis of individually transplanted KTLS HSCs has failed to show significant levels of non-hematopoietic engraftment.123,124In addition, several reported trans-differentiation events that employed highly purified HSCs, and in some cases a very strong selection pressure for trans-differentiation, now have been shown to result from fusion of a blood cell with a non-blood cell, rather than from a change in fate of blood stem cells.125127 Finally, in the vast majority of cases, reported contributions of adult stem cells to cell types outside their tissue of origin are exceedingly rare, far too rare to be considered therapeutically useful. These findings have raised significant doubts about the biological importance and immediate clinical utility of adult hematopoietic stem cell plasticity. Instead, these results suggest that normal tissue regeneration relies predominantly on the function of cell type-specific stem or progenitor cells, and that the identification, isolation, and characterization of these cells may be more useful in designing novel approaches to regenerative medicine. Nonetheless, it is possible that a rigorous and concerted effort to identify, purify, and potentially expand the appropriate cell populations responsible for apparent quot;plasticityquot; events, characterize the tissue-specific and injury-related signals that recruit, stimulate, or regulate plasticity, and determine the mechanism(s) underlying cell fusion or transdifferentiation, may eventually enhance tissue regeneration via this mechanism to clinically useful levels.

Recent progress in genomic sequencing and genome-wide expression analysis at the RNA and protein levels has greatly increased our ability to study cells such as HSCs as quot;systems,quot; that is, as combinations of defined components with defined interactions. This goal has yet to be realized fully, as computational biology and system-wide protein biochemistry and proteomics still must catch up with the wealth of data currently generated at the genomic and transcriptional levels. Recent landmark events have included the sequencing of the human and mouse genomes and the development of techniques such as array-based analysis. Several research groups have combined cDNA cloning and sequencing with array-based analysis to begin to define the full transcriptional profile of HSCs from different species and developmental stages and compare these to other stem cells.64,65,128131 Many of the data are available in online databases, such as the NIH/NIDDK Stem Cell Genome Anatomy Projects. While transcriptional profiling is clearly a work in progress, comparisons among various types of stem cells may eventually identify sets of genes that are involved in defining the general quot;stemnessquot; of a cell, as well as sets of genes that define their exit from the stem cell pool (e.g., the beginning of their path toward becoming mature differentiated cells, also referred to as commitment). In addition, these datasets will reveal sets of genes that are associated with specific stem cell populations, such as HSCs and MSCs, and thus define their unique properties. Assembly of these datasets into pathways will greatly help to understand and to predict the responses of HSCs (and other stem cells) to various stimuli.

The clinical use of stem cells holds great promise, although the application of most classes of adult stem cells is either currently untested or is in the earliest phases of clinical testing.132,133 The only exception is HSCs, which have been used clinically since 1959 and are used increasingly routinely for transplantations, albeit almost exclusively in a non-pure form. By 1995, more than 40,000 transplants were performed annually world-wide.134,135 Currently the main indications for bone marrow transplantation are either hematopoietic cancers (leukemias and lymphomas), or the use of high-dose chemotherapy for non-hematopoietic malignancies (cancers in other organs). Other indications include diseases that involve genetic or acquired bone marrow failure, such as aplastic anemia, thalassemia sickle cell anemia, and increasingly, autoimmune diseases.

Transplantation of bone marrow and HSCs are carried out in two rather different settings, autologous and allogeneic. Autologous transplantations employ a patient's own bone marrow tissue and thus present no tissue incompatibility between the donor and the host. Allogeneic transplantations occur between two individuals who are not genetically identical (with the rare exceptions of transplantations between identical twins, often referred to as syngeneic transplantations). Non-identical individuals differ in their human leukocyte antigens (HLAs), proteins that are expressed by their white blood cells. The immune system uses these HLAs to distinguish between quot;selfquot; and quot;nonself.quot; For successful transplantation, allogeneic grafts must match most, if not all, of the six to ten major HLA antigens between host and donor. Even if they do, however, enough differences remain in mostly uncharacterized minor antigens to enable immune cells from the donor and the host to recognize the other as quot;nonself.quot; This is an important issue, as virtually all HSC transplants are carried out with either non-purified, mixed cell populations (mobilized peripheral blood, cord blood, or bone marrow) or cell populations that have been enriched for HSCs (e.g., by column selection for CD34+ cells) but have not been fully purified. These mixed population grafts contain sufficient lymphoid cells to mount an immune response against host cells if they are recognized as quot;non-self.quot; The clinical syndrome that results from this quot;non-selfquot; response is known as graft-versus-host disease (GVHD).136

In contrast, autologous grafts use cells harvested from the patient and offer the advantage of not causing GVHD. The main disadvantage of an autologous graft in the treatment of cancer is the absence of a graft-versusleukemia (GVL) or graft-versus-tumor (GVT) response, the specific immunological recognition of host tumor cells by donor-immune effector cells present in the transplant. Moreover, the possibility exists for contamination with cancerous or pre-cancerous cells.

Allogeneic grafts also have disadvantages. They are limited by the availability of immunologically-matched donors and the possibility of developing potentially lethal GVHD. The main advantage of allogeneic grafts is the potential for a GVL response, which can be an important contribution to achieving and maintaining complete remission.137,138

Today, most grafts used in the treatment of patients consist of either whole or CD34+-enriched bone marrow or, more likely, mobilized peripheral blood. The use of highly purified hematopoietic stem cells as grafts is rare.5658 However, the latter have the advantage of containing no detectable contaminating tumor cells in the case of autologous grafts, therefore not inducing GVHD, or presumably GVL,139141in allogeneic grafts. While they do so less efficiently than lymphocyte-containing cell mixtures, HSCs alone can engraft across full allogeneic barriers (i.e., when transplanted from a donor who is a complete mismatch for both major and minor transplantation antigens).139141The use of donor lymphocyte infusions (DLI) in the context of HSC transplantation allows for the controlled addition of lymphocytes, if necessary, to obtain or maintain high levels of donor cells and/or to induce a potentially curative GVL-response.142,143 The main problems associated with clinical use of highly purified HSCs are the additional labor and costs144 involved in obtaining highly purified cells in sufficient quantities.

While the possibilities of GVL and other immune responses to malignancies remain the focus of intense interest, it is also clear that in many cases, less-directed approaches such as chemotherapy or irradiation offer promise. However, while high-dose chemotherapy combined with autologous bone marrow transplantation has been reported to improve outcome (usually measured as the increase in time to progression, or increase in survival time),145154 this has not been observed by other researchers and remains controversial.155161 The tumor cells present in autologous grafts may be an important limitation in achieving long-term disease-free survival. Only further purification/ purging of the grafts, with rigorous separation of HSCs from cancer cells, can overcome this limitation. Initial small scale trials with HSCs purified by flow cytometry suggest that this is both possible and beneficial to the clinical outcome.56 In summary, purification of HSCs from cancer/lymphoma/leukemia patients offers the only possibility of using these cells post-chemotherapy to regenerate the host with cancer-free grafts. Purification of HSCs in allotransplantation allows transplantation with cells that regenerate the blood-forming system but cannot induce GVHD.

An important recent advance in the clinical use of HSCs is the development of non-myeloablative preconditioning regimens, sometimes referred to as quot;mini transplants.quot;162164 Traditionally, bone marrow or stem cell transplantation has been preceded by a preconditioning regimen consisting of chemotherapeutic agents, often combined with irradiation, that completely destroys host blood and bone marrow tissues (a process called myeloablation). This creates quot;spacequot; for the incoming cells by freeing stem cell niches and prevents an undesired immune response of the host cells against the graft cells, which could result in graft failure. However, myeloablation immunocompromises the patient severely and necessitates a prolonged hospital stay under sterile conditions. Many protocols have been developed that use a more limited and targeted approach to preconditioning. These nonmyeloablative preconditioning protocols, which combine excellent engraftment results with the ability to perform hematopoietic cell transplantation on an outpatient basis, have greatly changed the clinical practice of bone marrow transplantation.

FACS purification of HSCs in mouse and man completely eliminates contaminating T cells, and thus GVHD (which is caused by T-lymphocytes) in allogeneic transplants. Many HSC transplants have been carried out in different combinations of mouse strains. Some of these were matched at the major transplantation antigens but otherwise different (Matched Unrelated Donors or MUD); in others, no match at the major or minor transplantation antigens was expected. To achieve rapid and sustained engraftment, higher doses of HSCs were required in these mismatched allogeneic transplants than in syngeneic transplants.139141,165167 In these experiments, hosts whose immune and blood-forming systems were generated from genetically distinct donors were permanently capable of accepting organ transplants (such as the heart) from either donor or host, but not from mice unrelated to the donor or host. This phenomenon is known as transplant-induced tolerance and was observed whether the organ transplants were given the same day as the HSCs or up to one year later.139,166Hematopoietic cell transplant-related complications have limited the clinical application of such tolerance induction for solid organ grafts, but the use of non-myeloablative regimens to prepare the host, as discussed above, should significantly reduce the risk associated with combined HSC and organ transplants. Translation of these findings to human patients should enable a switch from chronic immunosuppression to prevent rejection to protocols wherein a single conditioning dose allows permanent engraftment of both the transplanted blood system and solid organ(s) or other tissue stem cells from the same donor. This should eliminate both GVHD and chronic host transplant immunosuppression, which lead to many complications, including life-threatening opportunistic infections and the development of malignant neoplasms.

We now know that several autoimmune diseasesdiseases in which immune cells attack normal body tissuesinvolve the inheritance of high risk-factor genes.168 Many of these genes are expressed only in blood cells. Researchers have recently tested whether HSCs could be used in mice with autoimmune disease (e.g., type 1 diabetes) to replace an autoimmune blood system with one that lacks the autoimmune risk genes. The HSC transplants cured mice that were in the process of disease development when nonmyeloablative conditioning was used for transplant.169 It has been observed that transplant-induced tolerance allows co-transplantation of pancreatic islet cells to replace destroyed islets.170 If these results using nonmyeloablative conditioning can be translated to humans, type 1 diabetes and several other autoimmune diseases may be treatable with pure HSC grafts. However, the reader should be cautioned that the translation of treatments from mice to humans is often complicated and time-consuming.

Banking is currently a routine procedure for UCB samples. If expansion of fully functional HSCs in tissue culture becomes a reality, HSC transplants may be possible by starting with small collections of HSCs rather than massive numbers acquired through mobilization and apheresis. With such a capability, collections of HSCs from volunteer donors or umbilical cords could be theoretically converted into storable, expandable stem cell banks useful on demand for clinical transplantation and/or for protection against radiation accidents. In mice, successful HSC transplants that regenerate fully normal immune and blood-forming systems can be accomplished when there is only a partial transplantation antigen match. Thus, the establishment of useful human HSC banks may require a match between as few as three out of six transplantation antigens (HLA). This might be accomplished with stem cell banks of as few as 4,00010,000 independent samples.

Leukemias are proliferative diseases of the hematopoietic system that fail to obey normal regulatory signals. They derive from stem cells or progenitors of the hematopoietic system and almost certainly include several stages of progression. During this progression, genetic and/or epigenetic changes occur, either in the DNA sequence itself (genetic) or other heritable modifications that affect the genome (epigenetic). These (epi)genetic changes alter cells from the normal hematopoietic system into cells capable of robust leukemic growth. There are a variety of leukemias, usually classified by the predominant pathologic cell types and/or the clinical course of the disease. It has been proposed that these are diseases in which self-renewing but poorly regulated cells, so-called "leukemia stem cells" (LSCs), are the populations that harbor all the genetic and epigenetic changes that allow leukemic progression.171176 While their progeny may be the characteristic cells observed with the leukemia, these progeny cells are not the self-renewing "malignant" cells of the disease. In this view, the events contributing to tumorigenic transformation, such as interrupted or decreased expression of "tumor suppressor" genes, loss of programmed death pathways, evasion of immune cells and macrophage surveillance mechanisms, retention of telomeres, and activation or amplification of self-renewal pathways, occur as single, rare events in the clonal progression to blast-crisis leukemia. As LT HSCs are the only selfrenewing cells in the myeloid pathway, it has been proposed that most, if not all, progression events occur at this level of differentiation, creating clonal cohorts of HSCs with increasing malignancy (see Figure 2.6). In this disease model, the final event, explosive selfrenewal, could occur at the level of HSC or at any of the known progenitors (see Figures 2.5 and 2.6). Activation of the -catenin/lef-tcf signal transduction and transcription pathway has been implicated in leukemic stem cell self-renewal in mouse AML and human CML.177 In both cases, the granulocyte-macrophage progenitors, not the HSCs or progeny blast cells, are the malignant self-renewing entities. In other models, such as the JunB-deficient tumors in mice and in chronic-phase CML in humans, the leukemic stem cell is the HSC itself.90,177 However, these HSCs still respond to regulatory signals, thus representing steps in the clonal progression toward blast crisis (see Figure 2.6).

Figure 2.6. Leukemic progression at the hematopoietic stem cell level. Self-renewing HSCs are the cells present long enough to accumulate the many activating events necessary for full transformation into tumorigenic cells. Under normal conditions, half of the offspring of HSC cell divisions would be expected to undergo differentiation, leaving the HSC pool stable in size. (A) (Pre) leukemic progression results in cohorts of HSCs with increasing malignant potential. The cells with the additional event (two events are illustrated, although more would be expected to occur) can outcompete less-transformed cells in the HSC pool if they divide faster (as suggested in the figure) or are more resistant to differentiation or apoptosis (cell death), two major exit routes from the HSC pool. (B) Normal HSCs differentiate into progenitors and mature cells; this is linked with limited proliferation (left). Partially transformed HSCs can still differentiate into progenitors and mature cells, but more cells are produced. Also, the types of mature cells that are produced may be skewed from the normal ratio. Fully transformed cells may be completely blocked in terminal differentiation, and large numbers of primitive blast cells, representing either HSCs or self-renewing, transformed progenitor cells, can be produced. While this sequence of events is true for some leukemias (e.g., AML), not all of the events occur in every leukemia. As with non-transformed cells, most leukemia cells (other than the leukemia stem cells) can retain the potential for (limited) differentiation.

Many methods have revealed contributing protooncogenes and lost tumor suppressors in myeloid leukemias. Now that LSCs can be isolated, researchers should eventually be able to assess the full sequence of events in HSC clones undergoing leukemic transformation. For example, early events, such as the AML/ETO translocation in AML or the BCR/ABL translocation in CML can remain present in normal HSCs in patients who are in remission (e.g., without detectable cancer).177,178 The isolation of LSCs should enable a much more focused attack on these cells, drawing on their known gene expression patterns, the mutant genes they possess, and the proteomic analysis of the pathways altered by the proto-oncogenic events.173,176,179 Thus, immune therapies for leukemia would become more realistic, and approaches to classify and isolate LSCs in blood could be applied to search for cancer stem cells in other tissues.180

After more than 50 years of research and clinical use, hematopoietic stem cells have become the best-studied stem cells and, more importantly, hematopoietic stem cells have seen widespread clinical use. Yet the study of HSCs remains active and continues to advance very rapidly. Fueled by new basic research and clinical discoveries, HSCs hold promise for such indications as treating autoimmunity, generating tolerance for solid organ transplants, and directing cancer therapy. However, many challenges remain. The availability of (matched) HSCs for all of the potential applications continues to be a major hurdle. Efficient expansion of HSCs in culture remains one of the major research goals. Future developments in genomics and proteomics, as well as in gene therapy, have the potential to widen the horizon for clinical application of hematopoietic stem cells even further.

Notes:

* Cellerant Therapeutics, 1531 Industrial Road, San Carlos, CA 94070. Current address: Department of Surgery, Arizona Health Sciences Center, 1501 N. Campbell Avenue, P.O. Box 245071, Tucson, AZ 857245071,e-mail: jdomen@surgery.arizona.edu.

** Section on Developmental and Stem Cell Biology, Joslin Diabetes Center, One Joslin Place, Boston, MA 02215, E-mail: Amy_Wagers@harvard.edu

*** Director, Institute for Cancer/Stem Cell Biology and Medicine, Professor of Pathology and Developmental Biology, Stanford University School of Medicine, Stanford, CA 94305, Irv@stanford.edu.

Chapter1|Table of Contents|Chapter3

Read more:
Bone Marrow (Hematopoietic) Stem Cells | stemcells.nih.gov

Comments are closed.